Unnamed: 0.1
int64 0
113k
| Unnamed: 0
float64 0
113k
⌀ | title
stringlengths 7
246
| abstract
stringlengths 6
3.31k
|
|---|---|---|---|
106,800 | 106,800 |
Joint Coreset Construction and Quantization for Distributed Machine
Learning
|
Coresets are small, weighted summaries of larger datasets, aiming at
providing provable error bounds for machine learning (ML) tasks while
significantly reducing the communication and computation costs. To achieve a
better trade-off between ML error bounds and costs, we propose the first
framework to incorporate quantization techniques into the process of coreset
construction. Specifically, we theoretically analyze the ML error bounds caused
by a combination of coreset construction and quantization. Based on that, we
formulate an optimization problem to minimize the ML error under a fixed budget
of communication cost. To improve the scalability for large datasets, we
identify two proxies of the original objective function, for which efficient
algorithms are developed. For the case of data on multiple nodes, we further
design a novel algorithm to allocate the communication budget to the nodes
while minimizing the overall ML error. Through extensive experiments on
multiple real-world datasets, we demonstrate the effectiveness and efficiency
of our proposed algorithms for a variety of ML tasks. In particular, our
algorithms have achieved more than 90% data reduction with less than 10%
degradation in ML performance in most cases.
|
106,801 | 106,801 |
Sketching Algorithms and Lower Bounds for Ridge Regression
|
We give a sketching-based iterative algorithm that computes $1+\varepsilon$
approximate solutions for the ridge regression problem $\min_x \|{Ax-b}\|_2^2
+\lambda\|{x}\|_2^2$ where $A \in \mathbb{R}^{n \times d}$ with $d \ge n$. Our
algorithm, for a constant number of iterations (requiring a constant number of
passes over the input), improves upon earlier work of Chowdhury et al., by
requiring that the sketching matrix only has a weaker Approximate Matrix
Multiplication (AMM) guarantee that depends on $\epsilon$, along with a
constant subspace embedding guarantee. The earlier work instead requires that
the sketching matrix have a subspace embedding guarantee that depends on
$\epsilon$. For example, to produce a $1+\varepsilon$ approximate solution in
$1$ iteration, which requires $2$ passes over the input, our algorithm requires
the OSNAP embedding to have $m= O(n\sigma^2/\lambda\varepsilon)$ rows with a
sparsity parameter $s = O(\log(n))$, whereas the earlier algorithm of Chowdhury
et al., with the same number of rows of OSNAP requires a sparsity $s =
O(\sqrt{\sigma^2/\lambda\varepsilon} \cdot \log(n))$, where $\sigma =
\|{A}\|_2$ is the spectral norm of the matrix $A$. We also show that this
algorithm can be used to give faster algorithms for kernel ridge regression.
Finally, we show that the sketch size required for our algorithm is essentially
optimal for a natural framework of algorithms for ridge regression by proving
lower bounds on oblivious sketching matrices for AMM. The sketch size lower
bounds for AMM may be of independent interest.
|
106,802 | 106,802 |
Second Order Regret Bounds Against Generalized Expert Sequences under
Partial Bandit Feedback
|
We study the problem of expert advice under partial bandit feedback setting
and create a sequential minimax optimal algorithm. Our algorithm works with a
more general partial monitoring setting, where, in contrast to the classical
bandit feedback, the losses can be revealed in an adversarial manner. Our
algorithm adopts a universal prediction perspective, whose performance is
analyzed with regret against a general expert selection sequence. The regret we
study is against a general competition class that covers many settings (such as
the switching or contextual experts settings) and the expert selection
sequences in the competition class are determined by the application at hand.
Our regret bounds are second order bounds in terms of the sum of squared losses
and the normalized regret of our algorithm is invariant under arbitrary affine
transforms of the loss sequence. Our algorithm is truly online and does not use
any preliminary information about the loss sequences.
|
106,803 | 106,803 |
Achieving Representative Data via Convex Hull Feasibility Sampling
Algorithms
|
Sampling biases in training data are a major source of algorithmic biases in
machine learning systems. Although there are many methods that attempt to
mitigate such algorithmic biases during training, the most direct and obvious
way is simply collecting more representative training data. In this paper, we
consider the task of assembling a training dataset in which minority groups are
adequately represented from a given set of data sources. In essence, this is an
adaptive sampling problem to determine if a given point lies in the convex hull
of the means from a set of unknown distributions. We present adaptive sampling
methods to determine, with high confidence, whether it is possible to assemble
a representative dataset from the given data sources. We also demonstrate the
efficacy of our policies in simulations in the Bernoulli and a multinomial
setting.
|
106,804 | 106,804 |
A deep learning algorithm for reducing false positives in screening
mammography
|
Screening mammography improves breast cancer outcomes by enabling early
detection and treatment. However, false positive callbacks for additional
imaging from screening exams cause unnecessary procedures, patient anxiety, and
financial burden. This work demonstrates an AI algorithm that reduces false
positives by identifying mammograms not suspicious for breast cancer. We
trained the algorithm to determine the absence of cancer using 123,248 2D
digital mammograms (6,161 cancers) and performed a retrospective study on
14,831 screening exams (1,026 cancers) from 15 US and 3 UK sites. Retrospective
evaluation of the algorithm on the largest of the US sites (11,592 mammograms,
101 cancers) a) left the cancer detection rate unaffected (p=0.02,
non-inferiority margin 0.25 cancers per 1000 exams), b) reduced callbacks for
diagnostic exams by 31.1% compared to standard clinical readings, c) reduced
benign needle biopsies by 7.4%, and d) reduced screening exams requiring
radiologist interpretation by 41.6% in the simulated clinical workflow. This
work lays the foundation for semi-autonomous breast cancer screening systems
that could benefit patients and healthcare systems by reducing false positives,
unnecessary procedures, patient anxiety, and expenses.
|
106,805 | 106,805 |
GM-TOuNN: Graded Multiscale Topology Optimization using Neural Networks
|
Multiscale topology optimization (M-TO) entails generating an optimal global
topology, and an optimal set of microstructures at a smaller scale, for a
physics-constrained problem. With the advent of additive manufacturing, M-TO
has gained significant prominence. However, generating optimal microstructures
at various locations can be computationally very expensive. As an alternate,
graded multiscale topology optimization (GM-TO) has been proposed where one or
more pre-selected and graded (parameterized) microstructural topologies are
used to fill the domain optimally. This leads to a significant reduction in
computation while retaining many of the benefits of M-TO.
A successful GM-TO framework must: (1) be capable of efficiently handling
numerous pre-selected microstructures, (2) be able to continuously switch
between these microstructures during optimization, (3) ensure that the
partition of unity is satisfied, and (4) discourage microstructure mixing at
termination.
In this paper, we propose to meet these requirements by exploiting the unique
classification capacity of neural networks. Specifically, we propose a graded
multiscale topology optimization using neural-network (GM-TOuNN) framework with
the following features: (1) the number of design variables is only weakly
dependent on the number of pre-selected microstructures, (2) it guarantees
partition of unity while discouraging microstructure mixing, and (3) it
supports automatic differentiation, thereby eliminating manual sensitivity
analysis. The proposed framework is illustrated through several examples.
|
106,806 | 106,806 |
Multifidelity deep neural operators for efficient learning of partial
differential equations with application to fast inverse design of nanoscale
heat transport
|
Deep neural operators can learn operators mapping between
infinite-dimensional function spaces via deep neural networks and have become
an emerging paradigm of scientific machine learning. However, training neural
operators usually requires a large amount of high-fidelity data, which is often
difficult to obtain in real engineering problems. Here, we address this
challenge by using multifidelity learning, i.e., learning from multifidelity
datasets. We develop a multifidelity neural operator based on a deep operator
network (DeepONet). A multifidelity DeepONet includes two standard DeepONets
coupled by residual learning and input augmentation. Multifidelity DeepONet
significantly reduces the required amount of high-fidelity data and achieves
one order of magnitude smaller error when using the same amount of
high-fidelity data. We apply a multifidelity DeepONet to learn the phonon
Boltzmann transport equation (BTE), a framework to compute nanoscale heat
transport. By combining a trained multifidelity DeepONet with genetic algorithm
or topology optimization, we demonstrate a fast solver for the inverse design
of BTE problems.
|
106,807 | 106,807 |
Time Series of Non-Additive Metrics: Identification and Interpretation
of Contributing Factors of Variance by Linear Decomposition
|
The research paper addresses linear decomposition of time series of
non-additive metrics that allows for the identification and interpretation of
contributing factors (input features) of variance. Non-additive metrics, such
as ratios, are widely used in a variety of domains. It commonly requires
preceding aggregations of underlying variables that are used to calculate the
metric of interest. The latest poses a dimensionality challenge when the input
features and underlying variables are formed as two-dimensional arrays along
elements, such as account or customer identifications, and time points. It
rules out direct modeling of the time series of a non-additive metric as a
function of input features. The article discusses a five-step approach: (1)
segmentations of input features and the underlying variables of the metric that
are supported by unsupervised autoencoders, (2) univariate or joint fittings of
the metric by the aggregated input features on the segmented domains, (3)
transformations of pre-screened input features according to the fitted models,
(4) aggregation of the transformed features as time series, and (5) modelling
of the metric time series as a sum of constrained linear effects of the
aggregated features. Alternatively, approximation by numerical differentiation
has been considered to linearize the metric. It allows for element level
univariate or joint modeling of step (2). The process of these analytical steps
allows for a backward-looking explanatory decomposition of the metric as a sum
of time series of the survived input features. The paper includes a synthetic
example that studies loss-to-balance monthly rates of a hypothetical retail
credit portfolio. To validate that no latent factors other than the survived
input features have significant impacts on the metric, Statistical Process
Control has been introduced for the residual time series.
|
106,808 | 106,808 |
HASA: Hybrid Architecture Search with Aggregation Strategy for
Echinococcosis Classification and Ovary Segmentation in Ultrasound Images
|
Different from handcrafted features, deep neural networks can automatically
learn task-specific features from data. Due to this data-driven nature, they
have achieved remarkable success in various areas. However, manual design and
selection of suitable network architectures are time-consuming and require
substantial effort of human experts. To address this problem, researchers have
proposed neural architecture search (NAS) algorithms which can automatically
generate network architectures but suffer from heavy computational cost and
instability if searching from scratch. In this paper, we propose a hybrid NAS
framework for ultrasound (US) image classification and segmentation. The hybrid
framework consists of a pre-trained backbone and several searched cells (i.e.,
network building blocks), which takes advantage of the strengths of both NAS
and the expert knowledge from existing convolutional neural networks.
Specifically, two effective and lightweight operations, a mixed depth-wise
convolution operator and a squeeze-and-excitation block, are introduced into
the candidate operations to enhance the variety and capacity of the searched
cells. These two operations not only decrease model parameters but also boost
network performance. Moreover, we propose a re-aggregation strategy for the
searched cells, aiming to further improve the performance for different vision
tasks. We tested our method on two large US image datasets, including a 9-class
echinococcosis dataset containing 9566 images for classification and an ovary
dataset containing 3204 images for segmentation. Ablation experiments and
comparison with other handcrafted or automatically searched architectures
demonstrate that our method can generate more powerful and lightweight models
for the above US image classification and segmentation tasks.
|
106,809 | 106,809 |
Leveraging convergence behavior to balance conflicting tasks in
multi-task learning
|
Multi-Task Learning is a learning paradigm that uses correlated tasks to
improve performance generalization. A common way to learn multiple tasks is
through the hard parameter sharing approach, in which a single architecture is
used to share the same subset of parameters, creating an inductive bias between
them during the training process. Due to its simplicity, potential to improve
generalization, and reduce computational cost, it has gained the attention of
the scientific and industrial communities. However, tasks often conflict with
each other, which makes it challenging to define how the gradients of multiple
tasks should be combined to allow simultaneous learning. To address this
problem, we use the idea of multi-objective optimization to propose a method
that takes into account temporal behaviour of the gradients to create a dynamic
bias that adjust the importance of each task during the backpropagation. The
result of this method is to give more attention to the tasks that are diverging
or that are not being benefited during the last iterations, allowing to ensure
that the simultaneous learning is heading to the performance maximization of
all tasks. As a result, we empirically show that the proposed method
outperforms the state-of-art approaches on learning conflicting tasks. Unlike
the adopted baselines, our method ensures that all tasks reach good
generalization performances.
|
106,810 | 106,810 |
SNP2Vec: Scalable Self-Supervised Pre-Training for Genome-Wide
Association Study
|
Self-supervised pre-training methods have brought remarkable breakthroughs in
the understanding of text, image, and speech. Recent developments in genomics
has also adopted these pre-training methods for genome understanding. However,
they focus only on understanding haploid sequences, which hinders their
applicability towards understanding genetic variations, also known as single
nucleotide polymorphisms (SNPs), which is crucial for genome-wide association
study. In this paper, we introduce SNP2Vec, a scalable self-supervised
pre-training approach for understanding SNP. We apply SNP2Vec to perform
long-sequence genomics modeling, and we evaluate the effectiveness of our
approach on predicting Alzheimer's disease risk in a Chinese cohort. Our
approach significantly outperforms existing polygenic risk score methods and
all other baselines, including the model that is trained entirely with haploid
sequences. We release our code and dataset on
https://github.com/HLTCHKUST/snp2vec.
|
106,811 | 106,811 |
LSTM-Autoencoder based Anomaly Detection for Indoor Air Quality Time
Series Data
|
Anomaly detection for indoor air quality (IAQ) data has become an important
area of research as the quality of air is closely related to human health and
well-being. However, traditional statistics and shallow machine learning-based
approaches in anomaly detection in the IAQ area could not detect anomalies
involving the observation of correlations across several data points (i.e.,
often referred to as long-term dependences). We propose a hybrid deep learning
model that combines LSTM with Autoencoder for anomaly detection tasks in IAQ to
address this issue. In our approach, the LSTM network is comprised of multiple
LSTM cells that work with each other to learn the long-term dependences of the
data in a time-series sequence. Autoencoder identifies the optimal threshold
based on the reconstruction loss rates evaluated on every data across all
time-series sequences. Our experimental results, based on the Dunedin CO2
time-series dataset obtained through a real-world deployment of the schools in
New Zealand, demonstrate a very high and robust accuracy rate (99.50%) that
outperforms other similar models.
|
106,812 | 106,812 |
Control-oriented meta-learning
|
Real-time adaptation is imperative to the control of robots operating in
complex, dynamic environments. Adaptive control laws can endow even nonlinear
systems with good trajectory tracking performance, provided that any uncertain
dynamics terms are linearly parameterizable with known nonlinear features.
However, it is often difficult to specify such features a priori, such as for
aerodynamic disturbances on rotorcraft or interaction forces between a
manipulator arm and various objects. In this paper, we turn to data-driven
modeling with neural networks to learn, offline from past data, an adaptive
controller with an internal parametric model of these nonlinear features. Our
key insight is that we can better prepare the controller for deployment with
control-oriented meta-learning of features in closed-loop simulation, rather
than regression-oriented meta-learning of features to fit input-output data.
Specifically, we meta-learn the adaptive controller with closed-loop tracking
simulation as the base-learner and the average tracking error as the
meta-objective. With both fully-actuated and underactuated nonlinear planar
rotorcraft subject to wind, we demonstrate that our adaptive controller
outperforms other controllers trained with regression-oriented meta-learning
when deployed in closed-loop for trajectory tracking control.
|
106,813 | 106,813 |
Learning Convolutional Neural Networks in the Frequency Domain
|
Convolutional neural network (CNN) has achieved impressive success in
computer vision during the past few decades. The image convolution operation
helps CNNs to get good performance on image-related tasks. However, the image
convolution has high computation complexity and hard to be implemented. This
paper proposes the CEMNet, which can be trained in the frequency domain. The
most important motivation of this research is that we can use the
straightforward element-wise multiplication operation to replace the image
convolution in the frequency domain based on the Cross-Correlation Theorem,
which obviously reduces the computation complexity. We further introduce a
Weight Fixation mechanism to alleviate the problem of over-fitting, and analyze
the working behavior of Batch Normalization, Leaky ReLU, and Dropout in the
frequency domain to design their counterparts for CEMNet. Also, to deal with
complex inputs brought by Discrete Fourier Transform, we design a two-branches
network structure for CEMNet. Experimental results imply that CEMNet achieves
good performance on MNIST and CIFAR-10 databases.
|
106,814 | 106,814 |
Improving Top-K Decoding for Non-Autoregressive Semantic Parsing via
Intent Conditioning
|
Semantic parsing (SP) is a core component of modern virtual assistants like
Google Assistant and Amazon Alexa. While sequence-to-sequence-based
auto-regressive (AR) approaches are common for conversational semantic parsing,
recent studies employ non-autoregressive (NAR) decoders and reduce inference
latency while maintaining competitive parsing quality. However, a major
drawback of NAR decoders is the difficulty of generating top-k (i.e., k-best)
outputs with approaches such as beam search. To address this challenge, we
propose a novel NAR semantic parser that introduces intent conditioning on the
decoder. Inspired by the traditional intent and slot tagging parsers, we
decouple the top-level intent prediction from the rest of a parse. As the
top-level intent largely governs the syntax and semantics of a parse, the
intent conditioning allows the model to better control beam search and improves
the quality and diversity of top-k outputs. We introduce a hybrid
teacher-forcing approach to avoid training and inference mismatch. We evaluate
the proposed NAR on conversational SP datasets, TOP & TOPv2. Like the existing
NAR models, we maintain the O(1) decoding time complexity while generating more
diverse outputs and improving the top-3 exact match (EM) by 2.4 points. In
comparison with AR models, our model speeds up beam search inference by 6.7
times on CPU with competitive top-k EM.
|
106,815 | 106,815 |
HCFL: A High Compression Approach for Communication-Efficient Federated
Learning in Very Large Scale IoT Networks
|
Federated learning (FL) is a new artificial intelligence concept that enables
Internet-of-Things (IoT) devices to learn a collaborative model without sending
the raw data to centralized nodes for processing. Despite numerous advantages,
low computing resources at IoT devices and high communication costs for
exchanging model parameters make applications of FL in massive IoT networks
very limited. In this work, we develop a novel compression scheme for FL,
called high-compression federated learning (HCFL), for very large scale IoT
networks. HCFL can reduce the data load for FL processes without changing their
structure and hyperparameters. In this way, we not only can significantly
reduce communication costs, but also make intensive learning processes more
adaptable on low-computing resource IoT devices. Furthermore, we investigate a
relationship between the number of IoT devices and the convergence level of the
FL model and thereby better assess the quality of the FL process. We
demonstrate our HCFL scheme in both simulations and mathematical analyses. Our
proposed theoretical research can be used as a minimum level of satisfaction,
proving that the FL process can achieve good performance when a determined
configuration is met. Therefore, we show that HCFL is applicable in any
FL-integrated networks with numerous IoT devices.
|
106,816 | 106,816 |
Supplementation of deep neural networks with simplified physics-based
features to increase model prediction accuracy
|
To improve predictive models for STEM applications, supplemental
physics-based features computed from input parameters are introduced into
single and multiple layers of a deep neural network (DNN). While many studies
focus on informing DNNs with physics through differential equations or
numerical simulation, much may be gained through integration of simplified
relationships. To evaluate this hypothesis, a number of thin rectangular plates
simply-supported on all edges are simulated for five materials. With plate
dimensions and material properties as input features and fundamental natural
frequency as the sole output, predictive performance of a purely data-driven
DNN-based model is compared with models using additional inputs computed from
simplified physical relationships among baseline parameters, namely plate
weight, modulus of rigidity, and shear modulus. To better understand the
benefit to model accuracy, these additional features are injected into various
single and multiple DNN layers, and trained with four different dataset sizes.
When these physics-enhanced models are evaluated against independent data of
the same materials and similar dimensions to the training sets, supplementation
with simplified physics-based parameters provides little reduction in
prediction error over the baseline for models trained with dataset sizes of 60
and greater, although small improvement from 19.3% to 16.1% occurs when trained
with a sparse size of 30. Conversely, notable accuracy gains occur when the
independent test data is of material and dimensions not conforming to the
training set. Specifically, when physics-enhanced data is injected into
multiple DNN layers, reductions in error from 33.2% to 19.6%, 34.9% to 19.9%,
35.8% to 22.4%, and 43.0% to 28.4% are achieved for training dataset sizes of
261, 117, 60, and 30, respectively, demonstrating attainment of a degree of
generalizability.
|
106,817 | 106,817 |
Multimodal spatiotemporal graph neural networks for improved prediction
of 30-day all-cause hospital readmission
|
Measures to predict 30-day readmission are considered an important quality
factor for hospitals as accurate predictions can reduce the overall cost of
care by identifying high risk patients before they are discharged. While recent
deep learning-based studies have shown promising empirical results on
readmission prediction, several limitations exist that may hinder widespread
clinical utility, such as (a) only patients with certain conditions are
considered, (b) existing approaches do not leverage data temporality, (c)
individual admissions are assumed independent of each other, which is
unrealistic, (d) prior studies are usually limited to single source of data and
single center data. To address these limitations, we propose a multimodal,
modality-agnostic spatiotemporal graph neural network (MM-STGNN) for prediction
of 30-day all-cause hospital readmission that fuses multimodal in-patient
longitudinal data. By training and evaluating our methods using longitudinal
chest radiographs and electronic health records from two independent centers,
we demonstrate that MM-STGNN achieves AUROC of 0.79 on both primary and
external datasets. Furthermore, MM-STGNN significantly outperforms the current
clinical reference standard, LACE+ score (AUROC=0.61), on the primary dataset.
For subset populations of patients with heart and vascular disease, our model
also outperforms baselines on predicting 30-day readmission (e.g., 3.7 point
improvement in AUROC in patients with heart disease). Lastly, qualitative model
interpretability analysis indicates that while patients' primary diagnoses were
not explicitly used to train the model, node features crucial for model
prediction directly reflect patients' primary diagnoses. Importantly, our
MM-STGNN is agnostic to node feature modalities and could be utilized to
integrate multimodal data for triaging patients in various downstream resource
allocation tasks.
|
106,818 | 106,818 |
Learning Task-Aware Energy Disaggregation: a Federated Approach
|
We consider the problem of learning the energy disaggregation signals for
residential load data. Such task is referred as non-intrusive load monitoring
(NILM), and in order to find individual devices' power consumption profiles
based on aggregated meter measurements, a machine learning model is usually
trained based on large amount of training data coming from a number of
residential homes. Yet collecting such residential load datasets require both
huge efforts and customers' approval on sharing metering data, while load data
coming from different regions or electricity users may exhibit heterogeneous
usage patterns. Both practical concerns make training a single, centralized
NILM model challenging. In this paper, we propose a decentralized and
task-adaptive learning scheme for NILM tasks, where nested meta learning and
federated learning steps are designed for learning task-specific models
collectively. Simulation results on benchmark dataset validate proposed
algorithm's performance on efficiently inferring appliance-level consumption
for a variety of homes and appliances.
|
106,819 | 106,819 |
Sign Bit is Enough: A Learning Synchronization Framework for Multi-hop
All-reduce with Ultimate Compression
|
Traditional one-bit compressed stochastic gradient descent can not be
directly employed in multi-hop all-reduce, a widely adopted distributed
training paradigm in network-intensive high-performance computing systems such
as public clouds. According to our theoretical findings, due to the cascading
compression, the training process has considerable deterioration on the
convergence performance. To overcome this limitation, we implement a sign-bit
compression-based learning synchronization framework, Marsit. It prevents
cascading compression via an elaborate bit-wise operation for unbiased sign
aggregation and its specific global compensation mechanism for mitigating
compression deviation. The proposed framework retains the same theoretical
convergence rate as non-compression mechanisms. Experimental results
demonstrate that Marsit reduces up to 35% training time while preserving the
same accuracy as training without compression.
|
106,820 | 106,820 |
YOLO-Pose: Enhancing YOLO for Multi Person Pose Estimation Using Object
Keypoint Similarity Loss
|
We introduce YOLO-pose, a novel heatmap-free approach for joint detection,
and 2D multi-person pose estimation in an image based on the popular YOLO
object detection framework. Existing heatmap based two-stage approaches are
sub-optimal as they are not end-to-end trainable and training relies on a
surrogate L1 loss that is not equivalent to maximizing the evaluation metric,
i.e. Object Keypoint Similarity (OKS). Our framework allows us to train the
model end-to-end and optimize the OKS metric itself. The proposed model learns
to jointly detect bounding boxes for multiple persons and their corresponding
2D poses in a single forward pass and thus bringing in the best of both
top-down and bottom-up approaches. Proposed approach doesn't require the
postprocessing of bottom-up approaches to group detected keypoints into a
skeleton as each bounding box has an associated pose, resulting in an inherent
grouping of the keypoints. Unlike top-down approaches, multiple forward passes
are done away with since all persons are localized along with their pose in a
single inference. YOLO-pose achieves new state-of-the-art results on COCO
validation (90.2% AP50) and test-dev set (90.3% AP50), surpassing all existing
bottom-up approaches in a single forward pass without flip test, multi-scale
testing, or any other test time augmentation. All experiments and results
reported in this paper are without any test time augmentation, unlike
traditional approaches that use flip-test and multi-scale testing to boost
performance. Our training codes will be made publicly available at
https://github.com/TexasInstruments/edgeai-yolov5 and
https://github.com/TexasInstruments/edgeai-yolox
|
106,821 | 106,821 |
deep-significance - Easy and Meaningful Statistical Significance Testing
in the Age of Neural Networks
|
A lot of Machine Learning (ML) and Deep Learning (DL) research is of an
empirical nature. Nevertheless, statistical significance testing (SST) is still
not widely used. This endangers true progress, as seeming improvements over a
baseline might be statistical flukes, leading follow-up research astray while
wasting human and computational resources. Here, we provide an easy-to-use
package containing different significance tests and utility functions
specifically tailored towards research needs and usability.
|
106,822 | 106,822 |
Stream-based Active Learning with Verification Latency in Non-stationary
Environments
|
Data stream classification is an important problem in the field of machine
learning. Due to the non-stationary nature of the data where the underlying
distribution changes over time (concept drift), the model needs to continuously
adapt to new data statistics. Stream-based Active Learning (AL) approaches
address this problem by interactively querying a human expert to provide new
data labels for the most recent samples, within a limited budget. Existing AL
strategies assume that labels are immediately available, while in a real-world
scenario the expert requires time to provide a queried label (verification
latency), and by the time the requested labels arrive they may not be relevant
anymore. In this article, we investigate the influence of finite,
time-variable, and unknown verification delay, in the presence of concept drift
on AL approaches. We propose PRopagate (PR), a latency independent utility
estimator which also predicts the requested, but not yet known, labels.
Furthermore, we propose a drift-dependent dynamic budget strategy, which uses a
variable distribution of the labelling budget over time, after a detected
drift. Thorough experimental evaluation, with both synthetic and real-world
non-stationary datasets, and different settings of verification latency and
budget are conducted and analyzed. We empirically show that the proposed method
consistently outperforms the state-of-the-art. Additionally, we demonstrate
that with variable budget allocation in time, it is possible to boost the
performance of AL strategies, without increasing the overall labeling budget.
|
106,823 | 106,823 |
The Vision of Self-Evolving Computing Systems
|
Computing systems are omnipresent; their sustainability has become crucial
for our society. A key aspect of this sustainability is the ability of
computing systems to cope with the continuous change they face, ranging from
dynamic operating conditions, to changing goals, and technological progress.
While we are able to engineer smart computing systems that autonomously deal
with various types of changes, handling unanticipated changes requires system
evolution, which remains in essence a human-centered process. This will
eventually become unmanageable. To break through the status quo, we put forward
an arguable opinion for the vision of self-evolving computing systems that are
equipped with an evolutionary engine enabling them to evolve autonomously.
Specifically, when a self-evolving computing system detects conditions outside
its operational domain, such as an anomaly or a new goal, it activates an
evolutionary engine that runs online experiments to determine how the system
needs to evolve to deal with the changes, thereby evolving its architecture.
During this process the engine can integrate new computing elements that are
provided by computing warehouses. These computing elements provide
specifications and procedures enabling their automatic integration. We motivate
the need for self-evolving computing systems in light of the state of the art,
outline a conceptual architecture of self-evolving computing systems, and
illustrate the architecture for a future smart city mobility system that needs
to evolve continuously with changing conditions. To conclude, we highlight key
research challenges to realize the vision of self-evolving computing systems.
|
106,824 | 106,824 |
MARF: Multiscale Adaptive-switch Random Forest for Leg Detection with 2D
Laser Scanners
|
For the 2D laser-based tasks, e.g., people detection and people tracking, leg
detection is usually the first step. Thus, it carries great weight in
determining the performance of people detection and people tracking. However,
many leg detectors ignore the inevitable noise and the multiscale
characteristics of the laser scan, which makes them sensitive to the unreliable
features of point cloud and further degrades the performance of the leg
detector. In this paper, we propose a multiscale adaptive-switch Random Forest
(MARF) to overcome these two challenges. Firstly, the adaptive-switch decision
tree is designed to use noisesensitive features to conduct weighted
classification and noiseinvariant features to conduct binary classification,
which makes our detector perform more robust to noise. Secondly, considering
the multiscale property that the sparsity of the 2D point cloud is proportional
to the length of laser beams, we design a multiscale random forest structure to
detect legs at different distances. Moreover, the proposed approach allows us
to discover a sparser human leg from point clouds than others. Consequently,
our method shows an improved performance compared to other state-of-the-art leg
detectors on the challenging Moving Legs dataset and retains the whole pipeline
at a speed of 60+ FPS on lowcomputational laptops. Moreover, we further apply
the proposed MARF to the people detection and tracking system, achieving a
considerable gain in all metrics.
|
106,825 | 106,825 |
Surface Similarity Parameter: A New Machine Learning Loss Metric for
Oscillatory Spatio-Temporal Data
|
Supervised machine learning approaches require the formulation of a loss
functional to be minimized in the training phase. Sequential data are
ubiquitous across many fields of research, and are often treated with Euclidean
distance-based loss functions that were designed for tabular data. For smooth
oscillatory data, those conventional approaches lack the ability to penalize
amplitude, frequency and phase prediction errors at the same time, and tend to
be biased towards amplitude errors. We introduce the surface similarity
parameter (SSP) as a novel loss function that is especially useful for training
machine learning models on smooth oscillatory sequences. Our extensive
experiments on chaotic spatio-temporal dynamical systems indicate that the SSP
is beneficial for shaping gradients, thereby accelerating the training process,
reducing the final prediction error, and implementing a stronger regularization
effect compared to using classical loss functions. The results indicate the
potential of the novel loss metric particularly for highly complex and chaotic
data, such as data stemming from the nonlinear two-dimensional
Kuramoto-Sivashinsky equation and the linear propagation of dispersive surface
gravity waves in fluids.
|
106,826 | 106,826 |
ULF: Unsupervised Labeling Function Correction using Cross-Validation
for Weak Supervision
|
A way to overcome expensive and time-consuming manual data labeling is weak
supervision - automatic annotation of data samples via a predefined set of
labeling functions (LFs), rule-based mechanisms that generate potentially
erroneous labels. In this work, we investigate noise reduction techniques for
weak supervision based on the principle of k-fold cross-validation. In
particular, we extend two frameworks for detecting the erroneous samples in
manually annotated data to the weakly supervised setting. Our methods profit
from leveraging the information about matching LFs and detect noisy samples
more accurately. We also introduce a new algorithm for denoising the weakly
annotated data called ULF, that refines the allocation of LFs to classes by
estimating the reliable LFs-to-classes joint matrix. Evaluation on several
datasets shows that ULF successfully improves weakly supervised learning
without using any manually labeled data.
|
106,827 | 106,827 |
Program Analysis of Probabilistic Programs
|
Probabilistic programming is a growing area that strives to make statistical
analysis more accessible, by separating probabilistic modelling from
probabilistic inference. In practice this decoupling is difficult. No single
inference algorithm can be used as a probabilistic programming back-end that is
simultaneously reliable, efficient, black-box, and general. Probabilistic
programming languages often choose a single algorithm to apply to a given
problem, thus inheriting its limitations. While substantial work has been done
both to formalise probabilistic programming and to improve efficiency of
inference, there has been little work that makes use of the available program
structure, by formally analysing it, to better utilise the underlying inference
algorithm.
This dissertation presents three novel techniques (both static and dynamic),
which aim to improve probabilistic programming using program analysis. The
techniques analyse a probabilistic program and adapt it to make inference more
efficient, sometimes in a way that would have been tedious or impossible to do
by hand.
|
106,828 | 106,828 |
Shedding New Light on the Language of the Dark Web
|
The hidden nature and the limited accessibility of the Dark Web, combined
with the lack of public datasets in this domain, make it difficult to study its
inherent characteristics such as linguistic properties. Previous works on text
classification of Dark Web domain have suggested that the use of deep neural
models may be ineffective, potentially due to the linguistic differences
between the Dark and Surface Webs. However, not much work has been done to
uncover the linguistic characteristics of the Dark Web. This paper introduces
CoDA, a publicly available Dark Web dataset consisting of 10000 web documents
tailored towards text-based Dark Web analysis. By leveraging CoDA, we conduct a
thorough linguistic analysis of the Dark Web and examine the textual
differences between the Dark Web and the Surface Web. We also assess the
performance of various methods of Dark Web page classification. Finally, we
compare CoDA with an existing public Dark Web dataset and evaluate their
suitability for various use cases.
|
106,829 | 106,829 |
Gradient boosting for convex cone predict and optimize problems
|
Many problems in engineering and statistics involve both predictive
forecasting and decision-based optimization. Traditionally, predictive models
are optimized independently from the final decision-based optimization problem.
In contrast, a `smart, predict then optimize' (SPO) framework optimizes
prediction models to explicitly minimize the final downstream decision loss. In
this paper we present dboost, a gradient boosting algorithm for training
prediction model ensembles to minimize decision regret. The dboost framework
supports any convex optimization program that can be cast as convex quadratic
cone program and gradient boosting is performed by implicit differentiation of
a custom fixed-point mapping. To our knowledge, the dboost framework is the
first general purpose implementation of gradient boosting to predict and
optimize problems. Experimental results comparing with state-of-the-art SPO
methods show that dboost can further reduce out-of-sample decision regret.
|
106,830 | 106,830 |
Efficient and practical quantum compiler towards multi-qubit systems
with deep reinforcement learning
|
Efficient quantum compiling tactics greatly enhance the capability of quantum
computers to execute complicated quantum algorithms. Due to its fundamental
importance, a plethora of quantum compilers has been designed in past years.
However, there are several caveats to current protocols, which are low
optimality, high inference time, limited scalability, and lack of universality.
To compensate for these defects, here we devise an efficient and practical
quantum compiler assisted by advanced deep reinforcement learning (RL)
techniques, i.e., data generation, deep Q-learning, and AQ* search. In this
way, our protocol is compatible with various quantum machines and can be used
to compile multi-qubit operators. We systematically evaluate the performance of
our proposal in compiling quantum operators with both inverse-closed and
inverse-free universal basis sets. In the task of single-qubit operator
compiling, our proposal outperforms other RL-based quantum compilers in the
measure of compiling sequence length and inference time. Meanwhile, the output
solution is near-optimal, guaranteed by the Solovay-Kitaev theorem. Notably,
for the inverse-free universal basis set, the achieved sequence length
complexity is comparable with the inverse-based setting and dramatically
advances previous methods. These empirical results contribute to improving the
inverse-free Solovay-Kitaev theorem. In addition, for the first time, we
demonstrate how to leverage RL-based quantum compilers to accomplish two-qubit
operator compiling. The achieved results open an avenue for integrating RL with
quantum compiling to unify efficiency and practicality and thus facilitate the
exploration of quantum advantages.
|
106,831 | 106,831 |
Measurement-based Admission Control in Sliced Networks: A Best Arm
Identification Approach
|
In sliced networks, the shared tenancy of slices requires adaptive admission
control of data flows, based on measurements of network resources. In this
paper, we investigate the design of measurement-based admission control
schemes, deciding whether a new data flow can be admitted and in this case, on
which slice. The objective is to devise a joint measurement and decision
strategy that returns a correct decision (e.g., the least loaded slice) with a
certain level of confidence while minimizing the measurement cost (the number
of measurements made before committing to the decision). We study the design of
such strategies for several natural admission criteria specifying what a
correct decision is. For each of these criteria, using tools from best arm
identification in bandits, we first derive an explicit information-theoretical
lower bound on the cost of any algorithm returning the correct decision with
fixed confidence. We then devise a joint measurement and decision strategy
achieving this theoretical limit. We compare empirically the measurement costs
of these strategies, and compare them both to the lower bounds as well as a
naive measurement scheme. We find that our algorithm significantly outperforms
the naive scheme (by a factor $2-8$).
|
106,832 | 106,832 |
Global Counterfactual Explanations: Investigations, Implementations and
Improvements
|
Counterfactual explanations have been widely studied in explainability, with
a range of application dependent methods emerging in fairness, recourse and
model understanding. However, the major shortcoming associated with these
methods is their inability to provide explanations beyond the local or
instance-level. While some works touch upon the notion of a global explanation,
typically suggesting to aggregate masses of local explanations in the hope of
ascertaining global properties, few provide frameworks that are either reliable
or computationally tractable. Meanwhile, practitioners are requesting more
efficient and interactive explainability tools. We take this opportunity to
investigate existing global methods, with a focus on implementing and improving
Actionable Recourse Summaries (AReS), the only known global counterfactual
explanation framework for recourse.
|
106,833 | 106,833 |
Sketch guided and progressive growing GAN for realistic and editable
ultrasound image synthesis
|
Ultrasound (US) imaging is widely used for anatomical structure inspection in
clinical diagnosis. The training of new sonographers and deep learning based
algorithms for US image analysis usually requires a large amount of data.
However, obtaining and labeling large-scale US imaging data are not easy tasks,
especially for diseases with low incidence. Realistic US image synthesis can
alleviate this problem to a great extent. In this paper, we propose a
generative adversarial network (GAN) based image synthesis framework. Our main
contributions include: 1) we present the first work that can synthesize
realistic B-mode US images with high-resolution and customized texture editing
features; 2) to enhance structural details of generated images, we propose to
introduce auxiliary sketch guidance into a conditional GAN. We superpose the
edge sketch onto the object mask and use the composite mask as the network
input; 3) to generate high-resolution US images, we adopt a progressive
training strategy to gradually generate high-resolution images from
low-resolution images. In addition, a feature loss is proposed to minimize the
difference of high-level features between the generated and real images, which
further improves the quality of generated images; 4) the proposed US image
synthesis method is quite universal and can also be generalized to the US
images of other anatomical structures besides the three ones tested in our
study (lung, hip joint, and ovary); 5) extensive experiments on three large US
image datasets are conducted to validate our method. Ablation studies,
customized texture editing, user studies, and segmentation tests demonstrate
promising results of our method in synthesizing realistic US images.
|
106,834 | 106,834 |
Geometric Deep Learning to Identify the Critical 3D Structural Features
of the Optic Nerve Head for Glaucoma Diagnosis
|
Purpose: The optic nerve head (ONH) undergoes complex and deep 3D
morphological changes during the development and progression of glaucoma.
Optical coherence tomography (OCT) is the current gold standard to visualize
and quantify these changes, however the resulting 3D deep-tissue information
has not yet been fully exploited for the diagnosis and prognosis of glaucoma.
To this end, we aimed: (1) To compare the performance of two relatively recent
geometric deep learning techniques in diagnosing glaucoma from a single OCT
scan of the ONH; and (2) To identify the 3D structural features of the ONH that
are critical for the diagnosis of glaucoma.
Methods: In this study, we included a total of 2,247 non-glaucoma and 2,259
glaucoma scans from 1,725 subjects. All subjects had their ONHs imaged in 3D
with Spectralis OCT. All OCT scans were automatically segmented using deep
learning to identify major neural and connective tissues. Each ONH was then
represented as a 3D point cloud. We used PointNet and dynamic graph
convolutional neural network (DGCNN) to diagnose glaucoma from such 3D ONH
point clouds and to identify the critical 3D structural features of the ONH for
glaucoma diagnosis.
Results: Both the DGCNN (AUC: 0.97$\pm$0.01) and PointNet (AUC:
0.95$\pm$0.02) were able to accurately detect glaucoma from 3D ONH point
clouds. The critical points formed an hourglass pattern with most of them
located in the inferior and superior quadrant of the ONH.
Discussion: The diagnostic accuracy of both geometric deep learning
approaches was excellent. Moreover, we were able to identify the critical 3D
structural features of the ONH for glaucoma diagnosis that tremendously
improved the transparency and interpretability of our method. Consequently, our
approach may have strong potential to be used in clinical applications for the
diagnosis and prognosis of a wide range of ophthalmic disorders.
|
106,835 | 106,835 |
Concentration of Random Feature Matrices in High-Dimensions
|
The spectra of random feature matrices provide essential information on the
conditioning of the linear system used in random feature regression problems
and are thus connected to the consistency and generalization of random feature
models. Random feature matrices are asymmetric rectangular nonlinear matrices
depending on two input variables, the data and the weights, which can make
their characterization challenging. We consider two settings for the two input
variables, either both are random variables or one is a random variable and the
other is well-separated, i.e. there is a minimum distance between points. With
conditions on the dimension, the complexity ratio, and the sampling variance,
we show that the singular values of these matrices concentrate near their full
expectation and near one with high-probability. In particular, since the
dimension depends only on the logarithm of the number of random weights or the
number of data points, our complexity bounds can be achieved even in moderate
dimensions for many practical setting. The theoretical results are verified
with numerical experiments.
|
106,836 | 106,836 |
EEG-ITNet: An Explainable Inception Temporal Convolutional Network for
Motor Imagery Classification
|
In recent years, neural networks and especially deep architectures have
received substantial attention for EEG signal analysis in the field of
brain-computer interfaces (BCIs). In this ongoing research area, the end-to-end
models are more favoured than traditional approaches requiring signal
transformation pre-classification. They can eliminate the need for prior
information from experts and the extraction of handcrafted features. However,
although several deep learning algorithms have been already proposed in the
literature, achieving high accuracies for classifying motor movements or mental
tasks, they often face a lack of interpretability and therefore are not quite
favoured by the neuroscience community. The reasons behind this issue can be
the high number of parameters and the sensitivity of deep neural networks to
capture tiny yet unrelated discriminative features. We propose an end-to-end
deep learning architecture called EEG-ITNet and a more comprehensible method to
visualise the network learned patterns. Using inception modules and causal
convolutions with dilation, our model can extract rich spectral, spatial, and
temporal information from multi-channel EEG signals with less complexity (in
terms of the number of trainable parameters) than other existing end-to-end
architectures, such as EEG-Inception and EEG-TCNet. By an exhaustive evaluation
on dataset 2a from BCI competition IV and OpenBMI motor imagery dataset,
EEG-ITNet shows up to 5.9\% improvement in the classification accuracy in
different scenarios with statistical significance compared to its competitors.
We also comprehensively explain and support the validity of network
illustration from a neuroscientific perspective. We have also made our code
open at https://github.com/AbbasSalami/EEG-ITNet
|
106,837 | 106,837 |
LEFM-Nets: Learnable Explicit Feature Map Deep Networks for Segmentation
of Histopathological Images of Frozen Sections
|
Accurate segmentation of medical images is essential for diagnosis and
treatment of diseases. These problems are solved by highly complex models, such
as deep networks (DN), requiring a large amount of labeled data for training.
Thereby, many DNs possess task- or imaging modality specific architectures with
a decision-making process that is often hard to explain and interpret. Here, we
propose a framework that embeds existing DNs into a low-dimensional subspace
induced by the learnable explicit feature map (LEFM) layer. Compared to the
existing DN, the framework adds one hyperparameter and only modestly increase
the number of learnable parameters. The method is aimed at, but not limited to,
segmentation of low-dimensional medical images, such as color histopathological
images of stained frozen sections. Since features in the LEFM layer are
polynomial functions of the original features, proposed LEFM-Nets contribute to
the interpretability of network decisions. In this work, we combined LEFM with
the known networks: DeepLabv3+, UNet, UNet++ and MA-net. New LEFM-Nets are
applied to the segmentation of adenocarcinoma of a colon in a liver from images
of hematoxylin and eosin (H&E) stained frozen sections. LEFM-Nets are also
tested on nuclei segmentation from images of H&E stained frozen sections of ten
human organs. On the first problem, LEFM-Nets achieved statistically
significant performance improvement in terms of micro balanced accuracy and
$F_1$ score than original networks. LEFM-Nets achieved only better performance
in comparison with the original networks on the second problem. The source code
is available at https://github.com/dsitnik/lefm.
|
106,838 | 106,838 |
Finding MNEMON: Reviving Memories of Node Embeddings
|
Previous security research efforts orbiting around graphs have been
exclusively focusing on either (de-)anonymizing the graphs or understanding the
security and privacy issues of graph neural networks. Little attention has been
paid to understand the privacy risks of integrating the output from graph
embedding models (e.g., node embeddings) with complex downstream machine
learning pipelines. In this paper, we fill this gap and propose a novel
model-agnostic graph recovery attack that exploits the implicit graph
structural information preserved in the embeddings of graph nodes. We show that
an adversary can recover edges with decent accuracy by only gaining access to
the node embedding matrix of the original graph without interactions with the
node embedding models. We demonstrate the effectiveness and applicability of
our graph recovery attack through extensive experiments.
|
106,839 | 106,839 |
Latent Aspect Detection from Online Unsolicited Customer Reviews
|
Within the context of review analytics, aspects are the features of products
and services at which customers target their opinions and sentiments. Aspect
detection helps product owners and service providers to identify shortcomings
and prioritize customers' needs, and hence, maintain revenues and mitigate
customer churn. Existing methods focus on detecting the surface form of an
aspect by training supervised learning methods that fall short when aspects are
latent in reviews. In this paper, we propose an unsupervised method to extract
latent occurrences of aspects. Specifically, we assume that a customer
undergoes a two-stage hypothetical generative process when writing a review:
(1) deciding on an aspect amongst the set of aspects available for the product
or service, and (2) writing the opinion words that are more interrelated to the
chosen aspect from the set of all words available in a language. We employ
latent Dirichlet allocation to learn the latent aspects distributions for
generating the reviews. Experimental results on benchmark datasets show that
our proposed method is able to improve the state of the art when the aspects
are latent with no surface form in reviews.
|
106,840 | 106,840 |
The multi-modal universe of fast-fashion: the Visuelle 2.0 benchmark
|
We present Visuelle 2.0, the first dataset useful for facing diverse
prediction problems that a fast-fashion company has to manage routinely.
Furthermore, we demonstrate how the use of computer vision is substantial in
this scenario. Visuelle 2.0 contains data for 6 seasons / 5355 clothing
products of Nuna Lie, a famous Italian company with hundreds of shops located
in different areas within the country. In particular, we focus on a specific
prediction problem, namely short-observation new product sale forecasting
(SO-fore). SO-fore assumes that the season has started and a set of new
products is on the shelves of the different stores. The goal is to forecast the
sales for a particular horizon, given a short, available past (few weeks),
since no earlier statistics are available. To be successful, SO-fore approaches
should capture this short past and exploit other modalities or exogenous data.
To these aims, Visuelle 2.0 is equipped with disaggregated data at the
item-shop level and multi-modal information for each clothing item, allowing
computer vision approaches to come into play. The main message that we deliver
is that the use of image data with deep networks boosts performances obtained
when using the time series in long-term forecasting scenarios, ameliorating the
WAPE by 8.2% and the MAE by 7.7%. The dataset is available at:
https://humaticslab.github.io/forecasting/visuelle.
|
106,841 | 106,841 |
Planting Undetectable Backdoors in Machine Learning Models
|
Given the computational cost and technical expertise required to train
machine learning models, users may delegate the task of learning to a service
provider. We show how a malicious learner can plant an undetectable backdoor
into a classifier. On the surface, such a backdoored classifier behaves
normally, but in reality, the learner maintains a mechanism for changing the
classification of any input, with only a slight perturbation. Importantly,
without the appropriate "backdoor key", the mechanism is hidden and cannot be
detected by any computationally-bounded observer. We demonstrate two frameworks
for planting undetectable backdoors, with incomparable guarantees.
First, we show how to plant a backdoor in any model, using digital signature
schemes. The construction guarantees that given black-box access to the
original model and the backdoored version, it is computationally infeasible to
find even a single input where they differ. This property implies that the
backdoored model has generalization error comparable with the original model.
Second, we demonstrate how to insert undetectable backdoors in models trained
using the Random Fourier Features (RFF) learning paradigm or in Random ReLU
networks. In this construction, undetectability holds against powerful
white-box distinguishers: given a complete description of the network and the
training data, no efficient distinguisher can guess whether the model is
"clean" or contains a backdoor.
Our construction of undetectable backdoors also sheds light on the related
issue of robustness to adversarial examples. In particular, our construction
can produce a classifier that is indistinguishable from an "adversarially
robust" classifier, but where every input has an adversarial example! In
summary, the existence of undetectable backdoors represent a significant
theoretical roadblock to certifying adversarial robustness.
|
106,842 | 106,842 |
Solving AC Power Flow with Graph Neural Networks under Realistic
Constraints
|
In this paper we propose a graph neural network architecture solving the AC
power flow problem under realistic constraints. While the energy transition is
changing the energy industry to a digitalized and decentralized energy system,
the challenges are increasingly shifting to the distribution grid level to
integrate new loads and generation technologies. To ensure a save and resilient
operation of distribution grids, AC power flow calculations are the means of
choice to determine grid operating limits or analyze grid asset utilization in
planning procedures. In our approach we demonstrate the development of a
framework which makes use of graph neural networks to learn the physical
constraints of the power flow. We present our model architecture on which we
perform unsupervised training to learn a general solution of the AC power flow
formulation that is independent of the specific topologies and supply tasks
used for training. Finally, we demonstrate, validate and discuss our results on
medium voltage benchmark grids.
|
106,843 | 106,843 |
Medical Application of Geometric Deep Learning for the Diagnosis of
Glaucoma
|
Purpose: (1) To assess the performance of geometric deep learning (PointNet)
in diagnosing glaucoma from a single optical coherence tomography (OCT) 3D scan
of the optic nerve head (ONH); (2) To compare its performance to that obtained
with a standard 3D convolutional neural network (CNN), and with a gold-standard
glaucoma parameter, i.e. retinal nerve fiber layer (RNFL) thickness.
Methods: 3D raster scans of the ONH were acquired with Spectralis OCT for 477
glaucoma and 2,296 non-glaucoma subjects at the Singapore National Eye Centre.
All volumes were automatically segmented using deep learning to identify 7
major neural and connective tissues including the RNFL, the prelamina, and the
lamina cribrosa (LC). Each ONH was then represented as a 3D point cloud with
1,000 points chosen randomly from all tissue boundaries. To simplify the
problem, all ONH point clouds were aligned with respect to the plane and center
of Bruch's membrane opening. Geometric deep learning (PointNet) was then used
to provide a glaucoma diagnosis from a single OCT point cloud. The performance
of our approach was compared to that obtained with a 3D CNN, and with RNFL
thickness.
Results: PointNet was able to provide a robust glaucoma diagnosis solely from
the ONH represented as a 3D point cloud (AUC=95%). The performance of PointNet
was superior to that obtained with a standard 3D CNN (AUC=87%) and with that
obtained from RNFL thickness alone (AUC=80%).
Discussion: We provide a proof-of-principle for the application of geometric
deep learning in the field of glaucoma. Our technique requires significantly
less information as input to perform better than a 3D CNN, and with an AUC
superior to that obtained from RNFL thickness alone. Geometric deep learning
may have wide applicability in the field of Ophthalmology.
|
106,844 | 106,844 |
Interpretability of Machine Learning Methods Applied to Neuroimaging
|
Deep learning methods have become very popular for the processing of natural
images, and were then successfully adapted to the neuroimaging field. As these
methods are non-transparent, interpretability methods are needed to validate
them and ensure their reliability. Indeed, it has been shown that deep learning
models may obtain high performance even when using irrelevant features, by
exploiting biases in the training set. Such undesirable situations can
potentially be detected by using interpretability methods. Recently, many
methods have been proposed to interpret neural networks. However, this domain
is not mature yet. Machine learning users face two major issues when aiming to
interpret their models: which method to choose, and how to assess its
reliability? Here, we aim at providing answers to these questions by presenting
the most common interpretability methods and metrics developed to assess their
reliability, as well as their applications and benchmarks in the neuroimaging
context. Note that this is not an exhaustive survey: we aimed to focus on the
studies which we found to be the most representative and relevant.
|
106,845 | 106,845 |
Learning Invariances with Generalised Input-Convex Neural Networks
|
Considering smooth mappings from input vectors to continuous targets, our
goal is to characterise subspaces of the input domain, which are invariant
under such mappings. Thus, we want to characterise manifolds implicitly defined
by level sets. Specifically, this characterisation should be of a global
parametric form, which is especially useful for different informed data
exploration tasks, such as building grid-based approximations, sampling points
along the level curves, or finding trajectories on the manifold. However,
global parameterisations can only exist if the level sets are connected. For
this purpose, we introduce a novel and flexible class of neural networks that
generalise input-convex networks. These networks represent functions that are
guaranteed to have connected level sets forming smooth manifolds on the input
space. We further show that global parameterisations of these level sets can be
always found efficiently. Lastly, we demonstrate that our novel technique for
characterising invariances is a powerful generative data exploration tool in
real-world applications, such as computational chemistry.
|
106,846 | 106,846 |
From Environmental Sound Representation to Robustness of 2D CNN Models
Against Adversarial Attacks
|
This paper investigates the impact of different standard environmental sound
representations (spectrograms) on the recognition performance and adversarial
attack robustness of a victim residual convolutional neural network, namely
ResNet-18. Our main motivation for focusing on such a front-end classifier
rather than other complex architectures is balancing recognition accuracy and
the total number of training parameters. Herein, we measure the impact of
different settings required for generating more informative Mel-frequency
cepstral coefficient (MFCC), short-time Fourier transform (STFT), and discrete
wavelet transform (DWT) representations on our front-end model. This
measurement involves comparing the classification performance over the
adversarial robustness. We demonstrate an inverse relationship between
recognition accuracy and model robustness against six benchmarking attack
algorithms on the balance of average budgets allocated by the adversary and the
attack cost. Moreover, our experimental results have shown that while the
ResNet-18 model trained on DWT spectrograms achieves a high recognition
accuracy, attacking this model is relatively more costly for the adversary than
other 2D representations. We also report some results on different
convolutional neural network architectures such as ResNet-34, ResNet-56,
AlexNet, and GoogLeNet, SB-CNN, and LSTM-based.
|
106,847 | 106,847 |
Q-TART: Quickly Training for Adversarial Robustness and
in-Transferability
|
Raw deep neural network (DNN) performance is not enough; in real-world
settings, computational load, training efficiency and adversarial security are
just as or even more important. We propose to simultaneously tackle
Performance, Efficiency, and Robustness, using our proposed algorithm Q-TART,
Quickly Train for Adversarial Robustness and in-Transferability. Q-TART follows
the intuition that samples highly susceptible to noise strongly affect the
decision boundaries learned by DNNs, which in turn degrades their performance
and adversarial susceptibility. By identifying and removing such samples, we
demonstrate improved performance and adversarial robustness while using only a
subset of the training data. Through our experiments we highlight Q-TART's high
performance across multiple Dataset-DNN combinations, including ImageNet, and
provide insights into the complementary behavior of Q-TART alongside existing
adversarial training approaches to increase robustness by over 1.3% while using
up to 17.9% less training time.
|
106,848 | 106,848 |
Exploring the Distributed Knowledge Congruence in Proxy-data-free
Federated Distillation
|
Federated learning (FL) is a distributed machine learning paradigm in which
the server periodically aggregates local model parameters from clients without
assembling their private data.
Constrained communication and personalization requirements pose severe
challenges to FL. Federated distillation (FD) is proposed to simultaneously
address the above two problems, which exchanges knowledge between the server
and clients, supporting heterogeneous local models while significantly reducing
communication overhead. However, most existing FD methods require a proxy
dataset, which is often unavailable in reality.
A few recent proxy-data-free FD approaches can eliminate the need for
additional public data, but suffer from remarkable discrepancy among local
knowledge due to model heterogeneity, leading to ambiguous representation on
the server and inevitable accuracy degradation.
To tackle this issue, we propose a proxy-data-free FD algorithm based on
distributed knowledge congruence (FedDKC). FedDKC leverages well-designed
refinement strategies to narrow local knowledge differences into an acceptable
upper bound, so as to mitigate the negative effects of knowledge incongruence.
Specifically, from perspectives of peak probability and Shannon entropy of
local knowledge, we design kernel-based knowledge refinement (KKR) and
searching-based knowledge refinement (SKR) respectively, and theoretically
guarantee that the refined-local knowledge can satisfy an approximately-similar
distribution and be regarded as congruent.
Extensive experiments conducted on three common datasets demonstrate that our
proposed FedDKC significantly outperforms the state-of-the-art (accuracy boosts
in 93.33% comparisons,
Top-1 accuracy boosts by up to 4.38%, and Top-5 accuracy boosts by up to
10.31%) on various heterogeneous settings while evidently improving the
convergence speed.
|
106,849 | 106,849 |
Activation Regression for Continuous Domain Generalization with
Applications to Crop Classification
|
Geographic variance in satellite imagery impacts the ability of machine
learning models to generalise to new regions. In this paper, we model
geographic generalisation in medium resolution Landsat-8 satellite imagery as a
continuous domain adaptation problem, demonstrating how models generalise
better with appropriate domain knowledge. We develop a dataset spatially
distributed across the entire continental United States, providing macroscopic
insight into the effects of geography on crop classification in multi-spectral
and temporally distributed satellite imagery. Our method demonstrates improved
generalisability from 1) passing geographically correlated climate variables
along with the satellite data to a Transformer model and 2) regressing on the
model features to reconstruct these domain variables. Combined, we provide a
novel perspective on geographic generalisation in satellite imagery and a
simple-yet-effective approach to leverage domain knowledge. Code is available
at: \url{https://github.com/samar-khanna/cropmap}
|
106,850 | 106,850 |
Epileptic Seizure Risk Assessment by Multi-Channel Imaging of the EEG
|
Refractory epileptic patients can suffer a seizure at any moment. Seizure
prediction would substantially improve their lives. In this work, based on
scalp EEG and its transformation into images, the likelihood of an epileptic
seizure occurring at any moment is computed using an average of the softmax
layer output (the likelihood) of a CNN, instead of the output of the
classification layer. Results show that by analyzing the likelihood and
thresholding it, prediction has higher sensitivity or a lower FPR/h. The best
threshold for the likelihood was higher than 50% for 5 patients, and was lower
for the remaining 36. However, more testing is needed, especially in new
seizures, to better assess the real performance of this method. This work is a
proof of concept with a positive outlook.
|
106,851 | 106,851 |
LDPC codes: tracking non-stationary channel noise using sequential
variational Bayesian estimates
|
We present a sequential Bayesian learning method for tracking non-stationary
signal-to-noise ratios in LDPC codes using probabilistic graphical models. We
represent the LDPC code as a cluster graph using a general purpose cluster
graph construction algorithm called the layered trees running intersection
property (LTRIP) algorithm. The channel noise estimator is a global Gamma
cluster, which we extend to allow for Bayesian tracking of non-stationary noise
variation. We evaluate our proposed model on real-world 5G drive test data. Our
results show that our model is capable of tracking non-stationary channel
noise, which outperforms an LDPC code with a fixed knowledge of the actual
average channel noise.
|
106,852 | 106,852 |
Ensemble learning using individual neonatal data for seizure detection
|
Sharing medical data between institutions is difficult in practice due to
data protection laws and official procedures within institutions. Therefore,
most existing algorithms are trained on relatively small electroencephalogram
(EEG) data sets which is likely to be detrimental to prediction accuracy. In
this work, we simulate a case when the data can not be shared by splitting the
publicly available data set into disjoint sets representing data in individual
institutions. We propose to train a (local) detector in each institution and
aggregate their individual predictions into one final prediction. Four
aggregation schemes are compared, namely, the majority vote, the mean, the
weighted mean and the Dawid-Skene method. The approach allows different
detector architectures amongst the institutions. The method was validated on an
independent data set using only a subset of EEG channels. The ensemble reaches
accuracy comparable to a single detector trained on all the data when
sufficient amount of data is available in each institution. The weighted mean
aggregation scheme showed best overall performance, it was only marginally
outperformed by the Dawid-Skene method when local detectors approach
performance of a single detector trained on all available data.
|
106,853 | 106,853 |
Sim-to-Real 6D Object Pose Estimation via Iterative Self-training for
Robotic Bin-picking
|
In this paper, we propose an iterative self-training framework for
sim-to-real 6D object pose estimation to facilitate cost-effective robotic
grasping. Given a bin-picking scenario, we establish a photo-realistic
simulator to synthesize abundant virtual data, and use this to train an initial
pose estimation network. This network then takes the role of a teacher model,
which generates pose predictions for unlabeled real data. With these
predictions, we further design a comprehensive adaptive selection scheme to
distinguish reliable results, and leverage them as pseudo labels to update a
student model for pose estimation on real data. To continuously improve the
quality of pseudo labels, we iterate the above steps by taking the trained
student model as a new teacher and re-label real data using the refined teacher
model. We evaluate our method on a public benchmark and our newly-released
dataset, achieving an ADD(-S) improvement of 11.49% and 22.62% respectively.
Our method is also able to improve robotic bin-picking success by 19.54%,
demonstrating the potential of iterative sim-to-real solutions for robotic
applications.
|
106,854 | 106,854 |
BrainGB: A Benchmark for Brain Network Analysis with Graph Neural
Networks
|
Mapping the connectome of the human brain using structural or functional
connectivity has become one of the most pervasive paradigms for neuroimaging
analysis. Recently, Graph Neural Networks (GNNs) motivated from geometric deep
learning have attracted broad interest due to their established power for
modeling complex networked data. Despite their established performance in other
fields, there has not yet been a systematic study of how to design effective
GNNs for brain network analysis. To bridge this gap, we present BrainGB, a
benchmark for brain network analysis with GNNs. BrainGB standardizes the
process by 1) summarizing brain network construction pipelines for both
functional and structural neuroimaging modalities and 2) modularizing the
implementation of GNN designs. We conduct extensive experiments on datasets
across cohorts and modalities and recommend a set of general recipes for
effective GNN designs on brain networks. To support open and reproducible
research on GNN-based brain network analysis, we also host the BrainGB website
at https:// brainnet.us/ with models, tutorials, examples, as well as an
out-of-box Python package. We hope that this work will provide useful empirical
evidence and offer insights for future research in this novel and promising
direction.
|
106,855 | 106,855 |
Reflective Fiber Faults Detection and Characterization Using
Long-Short-Term Memory
|
To reduce operation-and-maintenance expenses (OPEX) and to ensure optical
network survivability, optical network operators need to detect and diagnose
faults in a timely manner and with high accuracy. With the rapid advancement of
telemetry technology and data analysis techniques, data-driven approaches
leveraging telemetry data to tackle the fault diagnosis problem have been
gaining popularity due to their quick implementation and deployment. In this
paper, we propose a novel multi-task learning model based on long short-term
memory (LSTM) to detect, locate, and estimate the reflectance of fiber
reflective faults (events) including the connectors and the mechanical splices
by extracting insights from monitored data obtained by the optical time domain
reflectometry (OTDR) principle commonly used for troubleshooting of fiber optic
cables or links. The experimental results prove that the proposed method: (i)
achieves a good detection capability and high localization accuracy within
short measurement time even for low SNR values; and (ii) outperforms
conventionally employed techniques.
|
106,856 | 106,856 |
Machine Learning-based Anomaly Detection in Optical Fiber Monitoring
|
Secure and reliable data communication in optical networks is critical for
high-speed Internet. However, optical fibers, serving as the data transmission
medium providing connectivity to billons of users worldwide, are prone to a
variety of anomalies resulting from hard failures (e.g., fiber cuts) and
malicious physical attacks (e.g., optical eavesdropping (fiber tapping)) etc.
Such anomalies may cause network disruption and thereby inducing huge financial
and data losses, or compromise the confidentiality of optical networks by
gaining unauthorized access to the carried data, or gradually degrade the
network operations. Therefore, it is highly required to implement efficient
anomaly detection, diagnosis, and localization schemes for enhancing the
availability and reliability of optical networks. In this paper, we propose a
data driven approach to accurately and quickly detect, diagnose, and localize
fiber anomalies including fiber cuts, and optical eavesdropping attacks. The
proposed method combines an autoencoder-based anomaly detection and an
attention-based bidirectional gated recurrent unit algorithm, whereby the
former is used for fault detection and the latter is adopted for fault
diagnosis and localization once an anomaly is detected by the autoencoder. We
verify the efficiency of our proposed approach by experiments under various
anomaly scenarios using real operational data. The experimental results
demonstrate that: (i) the autoencoder detects any fiber fault or anomaly with
an F1 score of 96.86%; and (ii) the attention-based bidirectional gated
recurrent unit algorithm identifies the the detected anomalies with an average
accuracy of 98.2%, and localizes the faults with an average root mean square
error of 0.19 m.
|
106,857 | 106,857 |
Network state Estimation using Raw Video Analysis: vQoS-GAN based
non-intrusive Deep Learning Approach
|
Content based providers transmits real time complex signal such as video data
from one region to another. During this transmission process, the signals
usually end up distorted or degraded where the actual information present in
the video is lost. This normally happens in the streaming video services
applications. Hence there is a need to know the level of degradation that
happened in the receiver side. This video degradation can be estimated by
network state parameters like data rate and packet loss values. Our proposed
solution vQoS GAN (video Quality of Service Generative Adversarial Network) can
estimate the network state parameters from the degraded received video data
using a deep learning approach of semi supervised generative adversarial
network algorithm. A robust and unique design of deep learning network model
has been trained with the video data along with data rate and packet loss class
labels and achieves over 95 percent of training accuracy. The proposed semi
supervised generative adversarial network can additionally reconstruct the
degraded video data to its original form for a better end user experience.
|
106,858 | 106,858 |
Streamable Neural Audio Synthesis With Non-Causal Convolutions
|
Deep learning models are mostly used in an offline inference fashion.
However, this strongly limits the use of these models inside audio generation
setups, as most creative workflows are based on real-time digital signal
processing. Although approaches based on recurrent networks can be naturally
adapted to this buffer-based computation, the use of convolutions still poses
some serious challenges. To tackle this issue, the use of causal streaming
convolutions have been proposed. However, this requires specific complexified
training and can impact the resulting audio quality.
In this paper, we introduce a new method allowing to produce non-causal
streaming models. This allows to make any convolutional model compatible with
real-time buffer-based processing. As our method is based on a post-training
reconfiguration of the model, we show that it is able to transform models
trained without causal constraints into a streaming model. We show how our
method can be adapted to fit complex architectures with parallel branches. To
evaluate our method, we apply it on the recent RAVE model, which provides
high-quality real-time audio synthesis. We test our approach on multiple music
and speech datasets and show that it is faster than overlap-add methods, while
having no impact on the generation quality. Finally, we introduce two
open-source implementation of our work as Max/MSP and PureData externals, and
as a VST audio plugin. This allows to endow traditional digital audio
workstation with real-time neural audio synthesis on a laptop CPU.
|
106,859 | 106,859 |
EvoSTS Forecasting: Evolutionary Sparse Time-Series Forecasting
|
In this work, we highlight our novel evolutionary sparse time-series
forecasting algorithm also known as EvoSTS. The algorithm attempts to
evolutionary prioritize weights of Long Short-Term Memory (LSTM) Network that
best minimize the reconstruction loss of a predicted signal using a learned
sparse coded dictionary. In each generation of our evolutionary algorithm, a
set number of children with the same initial weights are spawned. Each child
undergoes a training step and adjusts their weights on the same data. Due to
stochastic back-propagation, the set of children has a variety of weights with
different levels of performance. The weights that best minimize the
reconstruction loss with a given signal dictionary are passed to the next
generation. The predictions from the best-performing weights of the first and
last generation are compared. We found improvements while comparing the weights
of these two generations. However, due to several confounding parameters and
hyperparameter limitations, some of the weights had negligible improvements. To
the best of our knowledge, this is the first attempt to use sparse coding in
this way to optimize time series forecasting model weights, such as those of an
LSTM network.
|
106,860 | 106,860 |
A Unified Analysis of Dynamic Interactive Learning
|
In this paper we investigate the problem of learning evolving concepts over a
combinatorial structure. Previous work by Emamjomeh-Zadeh et al. [2020]
introduced dynamics into interactive learning as a way to model non-static user
preferences in clustering problems or recommender systems. We provide many
useful contributions to this problem. First, we give a framework that captures
both of the models analyzed by [Emamjomeh-Zadeh et al., 2020], which allows us
to study any type of concept evolution and matches the same query complexity
bounds and running time guarantees of the previous models. Using this general
model we solve the open problem of closing the gap between the upper and lower
bounds on query complexity. Finally, we study an efficient algorithm where the
learner simply follows the feedback at each round, and we provide mistake
bounds for low diameter graphs such as cliques, stars, and general o(log n)
diameter graphs by using a Markov Chain model.
|
106,861 | 106,861 |
SemiMultiPose: A Semi-supervised Multi-animal Pose Estimation Framework
|
Multi-animal pose estimation is essential for studying animals' social
behaviors in neuroscience and neuroethology. Advanced approaches have been
proposed to support multi-animal estimation and achieve state-of-the-art
performance. However, these models rarely exploit unlabeled data during
training even though real world applications have exponentially more unlabeled
frames than labeled frames. Manually adding dense annotations for a large
number of images or videos is costly and labor-intensive, especially for
multiple instances. Given these deficiencies, we propose a novel
semi-supervised architecture for multi-animal pose estimation, leveraging the
abundant structures pervasive in unlabeled frames in behavior videos to enhance
training, which is critical for sparsely-labeled problems. The resulting
algorithm will provide superior multi-animal pose estimation results on three
animal experiments compared to the state-of-the-art baseline and exhibits more
predictive power in sparsely-labeled data regimes.
|
106,862 | 106,862 |
Leveraging Natural Learning Processing to Uncover Themes in Clinical
Notes of Patients Admitted for Heart Failure
|
Heart failure occurs when the heart is not able to pump blood and oxygen to
support other organs in the body as it should. Treatments include medications
and sometimes hospitalization. Patients with heart failure can have both
cardiovascular as well as non-cardiovascular comorbidities. Clinical notes of
patients with heart failure can be analyzed to gain insight into the topics
discussed in these notes and the major comorbidities in these patients. In this
regard, we apply machine learning techniques, such as topic modeling, to
identify the major themes found in the clinical notes specific to the
procedures performed on 1,200 patients admitted for heart failure at the
University of Illinois Hospital and Health Sciences System (UI Health). Topic
modeling revealed five hidden themes in these clinical notes, including one
related to heart disease comorbidities.
|
106,863 | 106,863 |
Learning and controlling the source-filter representation of speech with
a variational autoencoder
|
Understanding and controlling latent representations in deep generative
models is a challenging yet important problem for analyzing, transforming and
generating various types of data. In speech processing, inspiring from the
anatomical mechanisms of phonation, the source-filter model considers that
speech signals are produced from a few independent and physically meaningful
continuous latent factors, among which the fundamental frequency $f_0$ and the
formants are of primary importance. In this work, we show that the
source-filter model of speech production naturally arises in the latent space
of a variational autoencoder (VAE) trained in an unsupervised manner on a
dataset of natural speech signals. Using only a few seconds of labeled speech
signals generated with an artificial speech synthesizer, we experimentally
illustrate that $f_0$ and the formant frequencies are encoded in orthogonal
subspaces of the VAE latent space and we develop a weakly-supervised method to
accurately and independently control these speech factors of variation within
the learned latent subspaces. Without requiring additional information such as
text or human-labeled data, this results in a deep generative model of speech
spectrograms that is conditioned on $f_0$ and the formant frequencies, and
which is applied to the transformation of speech signals.
|
106,864 | 106,864 |
Generative power of a protein language model trained on multiple
sequence alignments
|
Computational models starting from large ensembles of evolutionarily related
protein sequences capture a representation of protein families and learn
constraints associated to protein structure and function. They thus open the
possibility for generating novel sequences belonging to protein families.
Protein language models trained on multiple sequence alignments, such as MSA
Transformer, are highly attractive candidates to this end. We propose and test
an iterative method that directly uses the masked language modeling objective
to generate sequences using MSA Transformer. We demonstrate that the resulting
sequences generally score better than those generated by Potts models, and even
than natural sequences, for homology, coevolution and structure-based measures.
Moreover, MSA Transformer better reproduces the higher-order statistics and the
distribution of sequences in sequence space of natural data than Potts models,
although Potts models better reproduce first- and second-order statistics. MSA
Transformer is thus a strong candidate for protein sequence generation and
protein design.
|
106,865 | 106,865 |
Exploring Dual Encoder Architectures for Question Answering
|
Dual encoders have been used for question-answering (QA) and information
retrieval (IR) tasks with good results. There are two major types of dual
encoders, Siamese Dual Encoders (SDE), with parameters shared across two
encoders, and Asymmetric Dual Encoder (ADE), with two distinctly parameterized
encoders. In this work, we explore the dual encoder architectures for QA
retrieval tasks. By evaluating on MS MARCO and the MultiReQA benchmark, we show
that SDE performs significantly better than ADE. We further propose three
different improved versions of ADEs. Based on the evaluation of QA retrieval
tasks and direct analysis of the embeddings, we demonstrate that sharing
parameters in projection layers would enable ADEs to perform competitively with
SDEs.
|
106,866 | 106,866 |
MIMO Channel Estimation using Score-Based Generative Models
|
Channel estimation is a critical task in multiple-input multiple-output
digital communications that has effects on end-to-end system performance. In
this work, we introduce a novel approach for channel estimation using deep
score-based generative models. These models are trained to estimate the
gradient of the log-prior distribution, and can be used to iteratively refine
estimates, given observed measurements of a signal. We introduce a framework
for training score-based generative models for wireless channels, as well as
performing channel estimation using posterior sampling at test time. We derive
theoretical robustness guarantees of channel estimation with posterior sampling
in single-input single-output scenarios, and show that the observations
regarding estimation performance are verified experimentally in MIMO channels.
Our results in simulated clustered delay line channels show competitive
in-distribution performance without error floors in the high signal-to-noise
ratio regime, and robust out-of-distribution performance, outperforming
competing deep learning methods by up to 5 dB in end-to-end communication
performance, while the complexity analysis reveals how model architecture can
efficiently trade performance for estimation latency.
|
106,867 | 106,867 |
Learning Optimal Dynamic Treatment Regimes Using Causal Tree Methods in
Medicine
|
Dynamic treatment regimes (DTRs) are used in medicine to tailor sequential
treatment decisions to patients by considering patient heterogeneity. Common
methods for learning optimal DTRs, however, have shortcomings: they are
typically based on outcome prediction and not treatment effect estimation, or
they use linear models that are restrictive for patient data from modern
electronic health records. To address these shortcomings, we develop two novel
methods for learning optimal DTRs that effectively handle complex patient data.
We call our methods DTR-CT and DTR-CF. Our methods are based on a data-driven
estimation of heterogeneous treatment effects using causal tree methods,
specifically causal trees and causal forests, that learn non-linear
relationships, control for time-varying confounding, are doubly robust, and
explainable. To the best of our knowledge, our paper is the first that adapts
causal tree methods for learning optimal DTRs. We evaluate our proposed methods
using synthetic data and then apply them to real-world data from intensive care
units. Our methods outperform state-of-the-art baselines in terms of cumulative
regret and percentage of optimal decisions by a considerable margin. Our work
improves treatment recommendations from electronic health record and is thus of
direct relevance for personalized medicine.
|
106,868 | 106,868 |
Reinforcement Learning Policy Recommendation for Interbank Network
Stability
|
In this paper we analyze the effect of a policy recommendation on the
performances of an artificial interbank market. Financial institutions
stipulate lending agreements following a public recommendation and their
individual information. The former, modeled by a reinforcement learning optimal
policy trying to maximize the long term fitness of the system, gathers
information on the economic environment and directs economic actors to create
credit relationships based on the optimal choice between a low interest rate or
high liquidity supply. The latter, based on the agents' balance sheet, allows
to determine the liquidity supply and interest rate that the banks optimally
offer on the market. Based on the combination between the public and the
private signal, financial institutions create or cut their credit connections
over time via a preferential attachment evolving procedure able to generate a
dynamic network. Our results show that the emergence of a core-periphery
interbank network, combined with a certain level of homogeneity on the size of
lenders and borrowers, are essential features to ensure the resilience of the
system. Moreover, the reinforcement learning optimal policy recommendation
plays a crucial role in mitigating systemic risk with respect to alternative
policy instruments.
|
106,869 | 106,869 |
Scalable and Robust Self-Learning for Skill Routing in Large-Scale
Conversational AI Systems
|
Skill routing is an important component in large-scale conversational
systems. In contrast to traditional rule-based skill routing, state-of-the-art
systems use a model-based approach to enable natural conversations. To provide
supervision signal required to train such models, ideas such as human
annotation, replication of a rule-based system, relabeling based on user
paraphrases, and bandit-based learning were suggested. However, these
approaches: (a) do not scale in terms of the number of skills and skill
on-boarding, (b) require a very costly expert annotation/rule-design, (c)
introduce risks in the user experience with each model update. In this paper,
we present a scalable self-learning approach to explore routing alternatives
without causing abrupt policy changes that break the user experience, learn
from the user interaction, and incrementally improve the routing via frequent
model refreshes. To enable such robust frequent model updates, we suggest a
simple and effective approach that ensures controlled policy updates for
individual domains, followed by an off-policy evaluation for making deployment
decisions without any need for lengthy A/B experimentation. We conduct various
offline and online A/B experiments on a commercial large-scale conversational
system to demonstrate the effectiveness of the proposed method in real-world
production settings.
|
106,870 | 106,870 |
Accelerated Policy Learning with Parallel Differentiable Simulation
|
Deep reinforcement learning can generate complex control policies, but
requires large amounts of training data to work effectively. Recent work has
attempted to address this issue by leveraging differentiable simulators.
However, inherent problems such as local minima and exploding/vanishing
numerical gradients prevent these methods from being generally applied to
control tasks with complex contact-rich dynamics, such as humanoid locomotion
in classical RL benchmarks. In this work we present a high-performance
differentiable simulator and a new policy learning algorithm (SHAC) that can
effectively leverage simulation gradients, even in the presence of
non-smoothness. Our learning algorithm alleviates problems with local minima
through a smooth critic function, avoids vanishing/exploding gradients through
a truncated learning window, and allows many physical environments to be run in
parallel. We evaluate our method on classical RL control tasks, and show
substantial improvements in sample efficiency and wall-clock time over
state-of-the-art RL and differentiable simulation-based algorithms. In
addition, we demonstrate the scalability of our method by applying it to the
challenging high-dimensional problem of muscle-actuated locomotion with a large
action space, achieving a greater than 17x reduction in training time over the
best-performing established RL algorithm.
|
106,871 | 106,871 |
Masked Siamese Networks for Label-Efficient Learning
|
We propose Masked Siamese Networks (MSN), a self-supervised learning
framework for learning image representations. Our approach matches the
representation of an image view containing randomly masked patches to the
representation of the original unmasked image. This self-supervised
pre-training strategy is particularly scalable when applied to Vision
Transformers since only the unmasked patches are processed by the network. As a
result, MSNs improve the scalability of joint-embedding architectures, while
producing representations of a high semantic level that perform competitively
on low-shot image classification. For instance, on ImageNet-1K, with only 5,000
annotated images, our base MSN model achieves 72.4% top-1 accuracy, and with 1%
of ImageNet-1K labels, we achieve 75.7% top-1 accuracy, setting a new
state-of-the-art for self-supervised learning on this benchmark. Our code is
publicly available.
|
106,872 | 106,872 |
CLUES: A Benchmark for Learning Classifiers using Natural Language
Explanations
|
Supervised learning has traditionally focused on inductive learning by
observing labeled examples of a task. In contrast, humans have the ability to
learn new concepts from language. Here, we explore training zero-shot
classifiers for structured data purely from language. For this, we introduce
CLUES, a benchmark for Classifier Learning Using natural language ExplanationS,
consisting of a range of classification tasks over structured data along with
natural language supervision in the form of explanations. CLUES consists of 36
real-world and 144 synthetic classification tasks. It contains crowdsourced
explanations describing real-world tasks from multiple teachers and
programmatically generated explanations for the synthetic tasks. To model the
influence of explanations in classifying an example, we develop ExEnt, an
entailment-based model that learns classifiers using explanations. ExEnt
generalizes up to 18% better (relative) on novel tasks than a baseline that
does not use explanations. We delineate key challenges for automated learning
from explanations, addressing which can lead to progress on CLUES in the
future. Code and datasets are available at: https://clues-benchmark.github.io.
|
106,873 | 106,873 |
Neighborhood Attention Transformer
|
We present Neighborhood Attention Transformer (NAT), an efficient, accurate
and scalable hierarchical transformer that works well on both image
classification and downstream vision tasks. It is built upon Neighborhood
Attention (NA), a simple and flexible attention mechanism that localizes the
receptive field for each query to its nearest neighboring pixels. NA is a
localization of self-attention, and approaches it as the receptive field size
increases. It is also equivalent in FLOPs and memory usage to Swin
Transformer's shifted window attention given the same receptive field size,
while being less constrained. Furthermore, NA includes local inductive biases,
which eliminate the need for extra operations such as pixel shifts.
Experimental results on NAT are competitive; NAT-Tiny reaches 83.2% top-1
accuracy on ImageNet with only 4.3 GFLOPs and 28M parameters, 51.4% mAP on
MS-COCO and 48.4% mIoU on ADE20k. We will open-source our checkpoints, training
script, configurations, and our CUDA kernel at:
https://github.com/SHI-Labs/Neighborhood-Attention-Transformer .
|
106,874 | 106,874 |
Tight Bounds for Quantum State Certification with Incoherent
Measurements
|
We consider the problem of quantum state certification, where we are given
the description of a mixed state $\sigma \in \mathbb{C}^{d \times d}$, $n$
copies of a mixed state $\rho \in \mathbb{C}^{d \times d}$, and $\varepsilon >
0$, and we are asked to determine whether $\rho = \sigma$ or whether $\| \rho -
\sigma \|_1 > \varepsilon$. When $\sigma$ is the maximally mixed state
$\frac{1}{d} I_d$, this is known as mixedness testing. We focus on algorithms
which use incoherent measurements, i.e. which only measure one copy of $\rho$
at a time. Unlike those that use entangled, multi-copy measurements, these can
be implemented without persistent quantum memory and thus represent a large
class of protocols that can be run on current or near-term devices.
For mixedness testing, there is a folklore algorithm which uses incoherent
measurements and only needs $O(d^{3/2} / \varepsilon^2)$ copies. The algorithm
is non-adaptive, that is, its measurements are fixed ahead of time, and is
known to be optimal for non-adaptive algorithms. However, when the algorithm
can make arbitrary incoherent measurements, the best known lower bound is only
$\Omega (d^{4/3} / \varepsilon^2)$ [Bubeck-Chen-Li '20], and it has been an
outstanding open problem to close this polynomial gap. In this work, 1) we
settle the copy complexity of mixedness testing with incoherent measurements
and show that $\Omega (d^{3/2} / \varepsilon^2)$ copies are necessary, and 2)
we show the instance-optimal bounds for state certification to general $\sigma$
first derived by [Chen-Li-O'Donnell '21] for non-adaptive measurements also
hold for arbitrary incoherent measurements.
Qualitatively, our results say that adaptivity does not help at all for these
problems. Our results are based on new techniques that allow us to reduce the
problem to understanding certain matrix martingales, which we believe may be of
independent interest.
|
106,875 | 106,875 |
Any-resolution Training for High-resolution Image Synthesis
|
Generative models operate at fixed resolution, even though natural images
come in a variety of sizes. As high-resolution details are downsampled away,
and low-resolution images are discarded altogether, precious supervision is
lost. We argue that every pixel matters and create datasets with variable-size
images, collected at their native resolutions. Taking advantage of this data is
challenging; high-resolution processing is costly, and current architectures
can only process fixed-resolution data. We introduce continuous-scale training,
a process that samples patches at random scales to train a new generator with
variable output resolutions. First, conditioning the generator on a target
scale allows us to generate higher resolutions images than previously possible,
without adding layers to the model. Second, by conditioning on continuous
coordinates, we can sample patches that still obey a consistent global layout,
which also allows for scalable training at higher resolutions. Controlled FFHQ
experiments show our method takes advantage of the multi-resolution training
data better than discrete multi-scale approaches, achieving better FID scores
and cleaner high-frequency details. We also train on other natural image
domains including churches, mountains, and birds, and demonstrate arbitrary
scale synthesis with both coherent global layouts and realistic local details,
going beyond 2K resolution in our experiments. Our project page is available
at: https://chail.github.io/anyres-gan/.
|
106,876 | 106,876 |
A Level Set Theory for Neural Implicit Evolution under Explicit Flows
|
Coordinate-based neural networks parameterizing implicit surfaces have
emerged as efficient representations of geometry. They effectively act as
parametric level sets with the zero-level set defining the surface of interest.
We present a framework that allows applying deformation operations defined for
triangle meshes onto such implicit surfaces. Several of these operations can be
viewed as energy-minimization problems that induce an instantaneous flow field
on the explicit surface. Our method uses the flow field to deform parametric
implicit surfaces by extending the classical theory of level sets. We also
derive a consolidated view for existing methods on differentiable surface
extraction and rendering, by formalizing connections to the level-set theory.
We show that these methods drift from the theory and that our approach exhibits
improvements for applications like surface smoothing, mean-curvature flow,
inverse rendering and user-defined editing on implicit geometry.
|
106,877 | 106,877 |
Spatio-Temporal Analysis of Transformer based Architecture for Attention
Estimation from EEG
|
For many years now, understanding the brain mechanism has been a great
research subject in many different fields. Brain signal processing and
especially electroencephalogram (EEG) has recently known a growing interest
both in academia and industry. One of the main examples is the increasing
number of Brain-Computer Interfaces (BCI) aiming to link brains and computers.
In this paper, we present a novel framework allowing us to retrieve the
attention state, i.e degree of attention given to a specific task, from EEG
signals. While previous methods often consider the spatial relationship in EEG
through electrodes and process them in recurrent or convolutional based
architecture, we propose here to also exploit the spatial and temporal
information with a transformer-based network that has already shown its
supremacy in many machine-learning (ML) related studies, e.g. machine
translation. In addition to this novel architecture, an extensive study on the
feature extraction methods, frequential bands and temporal windows length has
also been carried out. The proposed network has been trained and validated on
two public datasets and achieves higher results compared to state-of-the-art
models. As well as proposing better results, the framework could be used in
real applications, e.g. Attention Deficit Hyperactivity Disorder (ADHD)
symptoms or vigilance during a driving assessment.
|
106,878 | 106,878 |
Diagnosing and Fixing Manifold Overfitting in Deep Generative Models
|
Likelihood-based, or explicit, deep generative models use neural networks to
construct flexible high-dimensional densities. This formulation directly
contradicts the manifold hypothesis, which states that observed data lies on a
low-dimensional manifold embedded in high-dimensional ambient space. In this
paper we investigate the pathologies of maximum-likelihood training in the
presence of this dimensionality mismatch. We formally prove that degenerate
optima are achieved wherein the manifold itself is learned but not the
distribution on it, a phenomenon we call manifold overfitting. We propose a
class of two-step procedures consisting of a dimensionality reduction step
followed by maximum-likelihood density estimation, and prove that they recover
the data-generating distribution in the nonparametric regime, thus avoiding
manifold overfitting. We also show that these procedures enable density
estimation on the manifolds learned by implicit models, such as generative
adversarial networks, hence addressing a major shortcoming of these models.
Several recently proposed methods are instances of our two-step procedures; we
thus unify, extend, and theoretically justify a large class of models.
|
106,879 | 106,879 |
Active Learning for Regression and Classification by Inverse Distance
Weighting
|
This paper proposes an active learning algorithm for solving regression and
classification problems based on inverse-distance weighting functions for
selecting the feature vectors to query. The algorithm has the following
features: (i) supports both pool-based and population-based sampling; (ii) is
independent of the type of predictor used; (iii) can handle known and unknown
constraints on the queryable feature vectors; and (iv) can run either
sequentially, or in batch mode, depending on how often the predictor is
retrained. The method's potential is shown in numerical tests on illustrative
synthetic problems and real-world regression and classification datasets from
the UCI repository. A Python implementation of the algorithm that we call IDEAL
(Inverse-Distance based Exploration for Active Learning), is available at
\url{http://cse.lab.imtlucca.it/~bemporad/ideal}.
|
106,880 | 106,880 |
Relaxing Equivariance Constraints with Non-stationary Continuous Filters
|
Equivariances provide useful inductive biases in neural network modeling,
with the translation equivariance of convolutional neural networks being a
canonical example. Equivariances can be embedded in architectures through
weight-sharing and place symmetry constraints on the functions a neural network
can represent. The type of symmetry is typically fixed and has to be chosen in
advance. Although some tasks are inherently equivariant, many tasks do not
strictly follow such symmetries. In such cases, equivariance constraints can be
overly restrictive. In this work, we propose a parameter-efficient relaxation
of equivariance that can effectively interpolate between a (i) non-equivariant
linear product, (ii) a strict-equivariant convolution, and (iii) a
strictly-invariant mapping. The proposed parameterization can be thought of as
a building block to allow adjustable symmetry structure in neural networks.
Compared to non-equivariant or strict-equivariant baselines, we experimentally
verify that soft equivariance leads to improved performance in terms of test
accuracy on CIFAR-10 and CIFAR-100 image classification tasks.
|
106,881 | 106,881 |
Brazilian Court Documents Clustered by Similarity Together Using Natural
Language Processing Approaches with Transformers
|
Recent advances in Artificial intelligence (AI) have leveraged promising
results in solving complex problems in the area of Natural Language Processing
(NLP), being an important tool to help in the expeditious resolution of
judicial proceedings in the legal area. In this context, this work targets the
problem of detecting the degree of similarity between judicial documents that
can be achieved in the inference group, by applying six NLP techniques based on
transformers, namely BERT, GPT-2 and RoBERTa pre-trained in the Brazilian
Portuguese language and the same specialized using 210,000 legal proceedings.
Documents were pre-processed and had their content transformed into a vector
representation using these NLP techniques. Unsupervised learning was used to
cluster the lawsuits, calculating the quality of the model based on the cosine
of the distance between the elements of the group to its centroid. We noticed
that models based on transformers present better performance when compared to
previous research, highlighting the RoBERTa model specialized in the Brazilian
Portuguese language, making it possible to advance in the current state of the
art in the area of NLP applied to the legal sector.
|
106,882 | 106,882 |
Testing distributional assumptions of learning algorithms
|
There are many important high dimensional function classes that have fast
agnostic learning algorithms when strong assumptions on the distribution of
examples can be made, such as Gaussianity or uniformity over the domain. But
how can one be sufficiently confident that the data indeed satisfies the
distributional assumption, so that one can trust in the output quality of the
agnostic learning algorithm? We propose a model by which to systematically
study the design of tester-learner pairs $(\mathcal{A},\mathcal{T})$, such that
if the distribution on examples in the data passes the tester $\mathcal{T}$
then one can safely trust the output of the agnostic learner $\mathcal{A}$ on
the data.
To demonstrate the power of the model, we apply it to the classical problem
of agnostically learning halfspaces under the standard Gaussian distribution
and present a tester-learner pair with a combined run-time of
$n^{\tilde{O}(1/\epsilon^4)}$. This qualitatively matches that of the best
known ordinary agnostic learning algorithms for this task. In contrast, finite
sample Gaussian distribution testers do not exist for the $L_1$ and EMD
distance measures. A key step in the analysis is a novel characterization of
concentration and anti-concentration properties of a distribution whose
low-degree moments approximately match those of a Gaussian. We also use tools
from polynomial approximation theory.
In contrast, we show strong lower bounds on the combined run-times of
tester-learner pairs for the problems of agnostically learning convex sets
under the Gaussian distribution and for monotone Boolean functions under the
uniform distribution over $\{0,1\}^n$. Through these lower bounds we exhibit
natural problems where there is a dramatic gap between standard agnostic
learning run-time and the run-time of the best tester-learner pair.
|
106,883 | 106,883 |
Hierarchical Embedded Bayesian Additive Regression Trees
|
We propose a simple yet powerful extension of Bayesian Additive Regression
Trees which we name Hierarchical Embedded BART (HE-BART). The model allows for
random effects to be included at the terminal node level of a set of regression
trees, making HE-BART a non-parametric alternative to mixed effects models
which avoids the need for the user to specify the structure of the random
effects in the model, whilst maintaining the prediction and uncertainty
calibration properties of standard BART. Using simulated and real-world
examples, we demonstrate that this new extension yields superior predictions
for many of the standard mixed effects models' example data sets, and yet still
provides consistent estimates of the random effect variances. In a future
version of this paper, we outline its use in larger, more advanced data sets
and structures.
|
106,884 | 106,884 |
Alternating Mahalanobis Distance Minimization for Stable and Accurate CP
Decomposition
|
CP decomposition (CPD) is prevalent in chemometrics, signal processing, data
mining and many more fields. While many algorithms have been proposed to
compute the CPD, alternating least squares (ALS) remains one of the most widely
used algorithm for computing the decomposition. Recent works have introduced
the notion of eigenvalues and singular values of a tensor and explored
applications of eigenvectors and singular vectors in areas like signal
processing, data analytics and in various other fields. We introduce a new
formulation for deriving singular values and vectors of a tensor by considering
the critical points of a function different from what is used in the previous
work. Computing these critical points in an alternating manner motivates an
alternating optimization algorithm which corresponds to alternating least
squares algorithm in the matrix case. However, for tensors with order greater
than equal to $3$, it minimizes an objective function which is different from
the commonly used least squares loss. Alternating optimization of this new
objective leads to simple updates to the factor matrices with the same
asymptotic computational cost as ALS. We show that a subsweep of this algorithm
can achieve a superlinear convergence rate for exact CPD with known rank and
verify it experimentally. We then view the algorithm as optimizing a
Mahalanobis distance with respect to each factor with ground metric dependent
on the other factors. This perspective allows us to generalize our approach to
interpolate between updates corresponding to the ALS and the new algorithm to
manage the tradeoff between stability and fitness of the decomposition. Our
experimental results show that for approximating synthetic and real-world
tensors, this algorithm and its variants converge to a better conditioned
decomposition with comparable and sometimes better fitness as compared to the
ALS algorithm.
|
106,885 | 106,885 |
Causal Disentanglement with Network Information for Debiased
Recommendations
|
Recommender systems aim to recommend new items to users by learning user and
item representations. In practice, these representations are highly entangled
as they consist of information about multiple factors, including user's
interests, item attributes along with confounding factors such as user
conformity, and item popularity. Considering these entangled representations
for inferring user preference may lead to biased recommendations (e.g., when
the recommender model recommends popular items even if they do not align with
the user's interests).
Recent research proposes to debias by modeling a recommender system from a
causal perspective. The exposure and the ratings are analogous to the treatment
and the outcome in the causal inference framework, respectively. The critical
challenge in this setting is accounting for the hidden confounders. These
confounders are unobserved, making it hard to measure them. On the other hand,
since these confounders affect both the exposure and the ratings, it is
essential to account for them in generating debiased recommendations. To better
approximate hidden confounders, we propose to leverage network information
(i.e., user-social and user-item networks), which are shown to influence how
users discover and interact with an item. Aside from the user conformity,
aspects of confounding such as item popularity present in the network
information is also captured in our method with the aid of \textit{causal
disentanglement} which unravels the learned representations into independent
factors that are responsible for (a) modeling the exposure of an item to the
user, (b) predicting the ratings, and (c) controlling the hidden confounders.
Experiments on real-world datasets validate the effectiveness of the proposed
model for debiasing recommender systems.
|
106,886 | 106,886 |
Learning two-phase microstructure evolution using neural operators and
autoencoder architectures
|
Phase-field modeling is an effective mesoscale method for capturing the
evolution dynamics of materials, e.g., in spinodal decomposition of a two-phase
mixture. However, the accuracy of high-fidelity phase field models comes at a
substantial computational cost. Hence, fast and generalizable surrogate models
are needed to alleviate the cost in computationally taxing processes such as in
optimization and design of materials. The intrinsic discontinuous nature of the
physical phenomena incurred by the presence of sharp phase boundaries makes the
training of the surrogate model cumbersome. We develop a new framework that
integrates a convolutional autoencoder architecture with a deep neural operator
(DeepONet) to learn the dynamic evolution of a two-phase mixture. We utilize
the convolutional autoencoder to provide a compact representation of the
microstructure data in a low-dimensional latent space. DeepONet, which consists
of two sub-networks, one for encoding the input function at a fixed number of
sensors locations (branch net) and another for encoding the locations for the
output functions (trunk net), learns the mesoscale dynamics of the
microstructure evolution in the latent space. The decoder part of the
convolutional autoencoder can then reconstruct the time-evolved microstructure
from the DeepONet predictions. The result is an efficient and accurate
accelerated phase-field framework that outperforms other neural-network-based
approaches while at the same time being robust to noisy inputs.
|
106,887 | 106,887 |
Physics-Aware Recurrent Convolutional (PARC) Neural Networks to
Assimilate Meso-scale Reactive Mechanics of Energetic Materials
|
The thermomechanical properties of energetic materials (EM) are known to be a
function of their microscopic structures, i.e., morphological configurations of
crystals and pores. This microstructural dependency has motivated vigorous
research in the EM community, seeking to engineer material microstructures with
targeted properties and performance under the materials-by-design paradigm.
However, establishing the complex structure-property-performance (SPP)
relationships of EMs demands extensive experimental and simulation efforts, and
assimilating and encapsulating these relationships in usable models is a
challenge. Here, we present a novel deep learning method, Physics-Aware
Recurrent Convolutional (PARC) Neural Network, that can "learn" the mesoscale
thermo-mechanics of EM microstructures during the shock-to-detonation
transition (SDT). We show that this new approach can produce accurate
high-fidelity predictions of time-evolving temperature and pressure fields of
the same quality as the state-of-the-art direct numerical simulations (DNS),
despite the dramatic reduction of computing time, from hours and days on a
high-performance computing cluster (HPC) to a little more than a second on a
commodity laptop. We also demonstrate that PARC can provide physical insights,
i.e., the artificial neurons can illuminate the underlying physics by
identifying which microstructural features led to critical hotspots and what
are the characteristics of "critical" versus "non-critical" microstructures.
This new knowledge generated alongside the capacity to conduct high-throughput
experiments will broaden our theoretical understanding of the initiation
mechanisms of EM detonation, as a step towards engineering EMs with specific
properties.
|
106,888 | 106,888 |
Harnessing Interpretable Machine Learning for Origami Feature Design and
Pattern Selection
|
Engineering design of origami systems is challenging because comparing
different origami patterns requires using categorical features and evaluating
multi-physics behavior targets introduces multi-objective problems. This work
shows that a decision tree machine learning method is particularly suitable for
the inverse design of origami. This interpretable machine learning method can
reveal complex interactions between categorical features and continuous
features for comparing different origami patterns, can tackle multi-objective
problems for designing active origami with multi-physics performance targets,
and can extend existing origami shape fitting algorithms to further consider
non-geometrical performances of origami systems. The proposed framework shows a
holistic way of designing active origami systems for various applications such
as metamaterials, deployable structures, soft robots, biomedical devices, and
many more.
|
106,889 | 106,889 |
Robotic and Generative Adversarial Attacks in Offline Writer-independent
Signature Verification
|
This study explores how robots and generative approaches can be used to mount
successful false-acceptance adversarial attacks on signature verification
systems. Initially, a convolutional neural network topology and data
augmentation strategy are explored and tuned, producing an 87.12% accurate
model for the verification of 2,640 human signatures. Two robots are then
tasked with forging 50 signatures, where 25 are used for the verification
attack, and the remaining 25 are used for tuning of the model to defend against
them. Adversarial attacks on the system show that there exists an information
security risk; the Line-us robotic arm can fool the system 24% of the time and
the iDraw 2.0 robot 32% of the time. A conditional GAN finds similar success,
with around 30% forged signatures misclassified as genuine. Following fine-tune
transfer learning of robotic and generative data, adversarial attacks are
reduced below the model threshold by both robots and the GAN. It is observed
that tuning the model reduces the risk of attack by robots to 8% and 12%, and
that conditional generative adversarial attacks can be reduced to 4% when 25
images are presented and 5% when 1000 images are presented.
|
106,890 | 106,890 |
Minimizing Control for Credit Assignment with Strong Feedback
|
The success of deep learning attracted interest in whether the brain learns
hierarchical representations using gradient-based learning. However, current
biologically plausible methods for gradient-based credit assignment in deep
neural networks need infinitesimally small feedback signals, which is
problematic in biologically realistic noisy environments and at odds with
experimental evidence in neuroscience showing that top-down feedback can
significantly influence neural activity. Building upon deep feedback control
(DFC), a recently proposed credit assignment method, we combine strong feedback
influences on neural activity with gradient-based learning and show that this
naturally leads to a novel view on neural network optimization. Instead of
gradually changing the network weights towards configurations with low output
loss, weight updates gradually minimize the amount of feedback required from a
controller that drives the network to the supervised output label. Moreover, we
show that the use of strong feedback in DFC allows learning forward and
feedback connections simultaneously, using a learning rule fully local in space
and time. We complement our theoretical results with experiments on standard
computer-vision benchmarks, showing competitive performance to backpropagation
as well as robustness to noise. Overall, our work presents a fundamentally
novel view of learning as control minimization, while sidestepping biologically
unrealistic assumptions.
|
106,891 | 106,891 |
Methodical Advice Collection and Reuse in Deep Reinforcement Learning
|
Reinforcement learning (RL) has shown great success in solving many
challenging tasks via use of deep neural networks. Although using deep learning
for RL brings immense representational power, it also causes a well-known
sample-inefficiency problem. This means that the algorithms are data-hungry and
require millions of training samples to converge to an adequate policy. One way
to combat this issue is to use action advising in a teacher-student framework,
where a knowledgeable teacher provides action advice to help the student. This
work considers how to better leverage uncertainties about when a student should
ask for advice and if the student can model the teacher to ask for less advice.
The student could decide to ask for advice when it is uncertain or when both it
and its model of the teacher are uncertain. In addition to this investigation,
this paper introduces a new method to compute uncertainty for a deep RL agent
using a secondary neural network. Our empirical results show that using dual
uncertainties to drive advice collection and reuse may improve learning
performance across several Atari games.
|
106,892 | 106,892 |
Causal Transformer for Estimating Counterfactual Outcomes
|
Estimating counterfactual outcomes over time from observational data is
relevant for many applications (e.g., personalized medicine). Yet,
state-of-the-art methods build upon simple long short-term memory (LSTM)
networks, thus rendering inferences for complex, long-range dependencies
challenging. In this paper, we develop a novel Causal Transformer for
estimating counterfactual outcomes over time. Our model is specifically
designed to capture complex, long-range dependencies among time-varying
confounders. For this, we combine three transformer subnetworks with separate
inputs for time-varying covariates, previous treatments, and previous outcomes
into a joint network with in-between cross-attentions. We further develop a
custom, end-to-end training procedure for our Causal Transformer. Specifically,
we propose a novel counterfactual domain confusion loss to address confounding
bias: it aims to learn adversarial balanced representations, so that they are
predictive of the next outcome but non-predictive of the current treatment
assignment. We evaluate our Causal Transformer based on synthetic and
real-world datasets, where it achieves superior performance over current
baselines. To the best of our knowledge, this is the first work proposing
transformer-based architecture for estimating counterfactual outcomes from
longitudinal data.
|
106,893 | 106,893 |
Convergence and Implicit Regularization Properties of Gradient Descent
for Deep Residual Networks
|
We prove linear convergence of gradient descent to a global minimum for the
training of deep residual networks with constant layer width and smooth
activation function. We further show that the trained weights, as a function of
the layer index, admits a scaling limit which is H\"older continuous as the
depth of the network tends to infinity. The proofs are based on non-asymptotic
estimates of the loss function and of norms of the network weights along the
gradient descent path. We illustrate the relevance of our theoretical results
to practical settings using detailed numerical experiments on supervised
learning problems.
|
106,894 | 106,894 |
auton-survival: an Open-Source Package for Regression, Counterfactual
Estimation, Evaluation and Phenotyping with Censored Time-to-Event Data
|
Applications of machine learning in healthcare often require working with
time-to-event prediction tasks including prognostication of an adverse event,
re-hospitalization or death. Such outcomes are typically subject to censoring
due to loss of follow up. Standard machine learning methods cannot be applied
in a straightforward manner to datasets with censored outcomes. In this paper,
we present auton-survival, an open-source repository of tools to streamline
working with censored time-to-event or survival data. auton-survival includes
tools for survival regression, adjustment in the presence of domain shift,
counterfactual estimation, phenotyping for risk stratification, evaluation, as
well as estimation of treatment effects. Through real world case studies
employing a large subset of the SEER oncology incidence data, we demonstrate
the ability of auton-survival to rapidly support data scientists in answering
complex health and epidemiological questions.
|
106,895 | 106,895 |
Characterizing the Efficiency vs. Accuracy Trade-off for Long-Context
NLP Models
|
With many real-world applications of Natural Language Processing (NLP)
comprising of long texts, there has been a rise in NLP benchmarks that measure
the accuracy of models that can handle longer input sequences. However, these
benchmarks do not consider the trade-offs between accuracy, speed, and power
consumption as input sizes or model sizes are varied. In this work, we perform
a systematic study of this accuracy vs. efficiency trade-off on two widely used
long-sequence models - Longformer-Encoder-Decoder (LED) and Big Bird - during
fine-tuning and inference on four datasets from the SCROLLS benchmark. To study
how this trade-off differs across hyperparameter settings, we compare the
models across four sequence lengths (1024, 2048, 3072, 4096) and two model
sizes (base and large) under a fixed resource budget. We find that LED
consistently achieves better accuracy at lower energy costs than Big Bird. For
summarization, we find that increasing model size is more energy efficient than
increasing sequence length for higher accuracy. However, this comes at the cost
of a large drop in inference speed. For question answering, we find that
smaller models are both more efficient and more accurate due to the larger
training batch sizes possible under a fixed resource budget.
|
106,896 | 106,896 |
The training response law explains how deep neural networks learn
|
Deep neural network is the widely applied technology in this decade. In spite
of the fruitful applications, the mechanism behind that is still to be
elucidated. We study the learning process with a very simple supervised
learning encoding problem. As a result, we found a simple law, in the training
response, which describes neural tangent kernel. The response consists of a
power law like decay multiplied by a simple response kernel. We can construct a
simple mean-field dynamical model with the law, which explains how the network
learns. In the learning, the input space is split into sub-spaces along
competition between the kernels. With the iterated splits and the aging, the
network gets more complexity, but finally loses its plasticity.
|
106,897 | 106,897 |
Unsupervised Probabilistic Models for Sequential Electronic Health
Records
|
We develop an unsupervised probabilistic model for heterogeneous Electronic
Health Record (EHR) data. Utilizing a mixture model formulation, our approach
directly models sequences of arbitrary length, such as medications and
laboratory results. This allows for subgrouping and incorporation of the
dynamics underlying heterogeneous data types. The model consists of a layered
set of latent variables that encode underlying structure in the data. These
variables represent subject subgroups at the top layer, and unobserved states
for sequences in the second layer. We train this model on episodic data from
subjects receiving medical care in the Kaiser Permanente Northern California
integrated healthcare delivery system. The resulting properties of the trained
model generate novel insight from these complex and multifaceted data. In
addition, we show how the model can be used to analyze sequences that
contribute to assessment of mortality likelihood.
|
106,898 | 106,898 |
Towards a Unified Framework for Uncertainty-aware Nonlinear Variable
Selection with Theoretical Guarantees
|
We develop a simple and unified framework for nonlinear variable selection
that incorporates uncertainty in the prediction function and is compatible with
a wide range of machine learning models (e.g., tree ensembles, kernel methods,
neural networks, etc). In particular, for a learned nonlinear model
$f(\mathbf{x})$, we consider quantifying the importance of an input variable
$\mathbf{x}^j$ using the integrated partial derivative $\Psi_j = \Vert
\frac{\partial}{\partial \mathbf{x}^j} f(\mathbf{x})\Vert^2_{P_\mathcal{X}}$.
We then (1) provide a principled approach for quantifying variable selection
uncertainty by deriving its posterior distribution, and (2) show that the
approach is generalizable even to non-differentiable models such as tree
ensembles. Rigorous Bayesian nonparametric theorems are derived to guarantee
the posterior consistency and asymptotic uncertainty of the proposed approach.
Extensive simulations and experiments on healthcare benchmark datasets confirm
that the proposed algorithm outperforms existing classic and recent variable
selection methods.
|
106,899 | 106,899 |
Pushing the Limits of Simple Pipelines for Few-Shot Learning: External
Data and Fine-Tuning Make a Difference
|
Few-shot learning (FSL) is an important and topical problem in computer
vision that has motivated extensive research into numerous methods spanning
from sophisticated meta-learning methods to simple transfer learning baselines.
We seek to push the limits of a simple-but-effective pipeline for more
realistic and practical settings of few-shot image classification. To this end,
we explore few-shot learning from the perspective of neural network
architecture, as well as a three stage pipeline of network updates under
different data supplies, where unsupervised external data is considered for
pre-training, base categories are used to simulate few-shot tasks for
meta-training, and the scarcely labelled data of an novel task is taken for
fine-tuning. We investigate questions such as: (1) How pre-training on external
data benefits FSL? (2) How state-of-the-art transformer architectures can be
exploited? and (3) How fine-tuning mitigates domain shift? Ultimately, we show
that a simple transformer-based pipeline yields surprisingly good performance
on standard benchmarks such as Mini-ImageNet, CIFAR-FS, CDFSL and Meta-Dataset.
Our code and demo are available at https://hushell.github.io/pmf.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.