text
stringlengths 29
3.31k
| label
sequencelengths 1
11
|
|---|---|
Real-world applications of object recognition often require the solution of
multiple tasks in a single platform. Under the standard paradigm of network
fine-tuning, an entirely new CNN is learned per task, and the final network
size is independent of task complexity. This is wasteful, since simple tasks
require smaller networks than more complex tasks, and limits the number of
tasks that can be solved simultaneously. To address these problems, we propose
a transfer learning procedure, denoted NetTailor, in which layers of a
pre-trained CNN are used as universal blocks that can be combined with small
task-specific layers to generate new networks. Besides minimizing
classification error, the new network is trained to mimic the internal
activations of a strong unconstrained CNN, and minimize its complexity by the
combination of 1) a soft-attention mechanism over blocks and 2) complexity
regularization constraints. In this way, NetTailor can adapt the network
architecture, not just its weights, to the target task. Experiments show that
networks adapted to simple tasks, such as character or traffic sign
recognition, become significantly smaller than those adapted to hard tasks,
such as fine-grained recognition. More importantly, due to the modular nature
of the procedure, this reduction in network complexity is achieved without
compromise of either parameter sharing across tasks, or classification
accuracy. | [
"cs.CV",
"cs.LG"
] |
Most of the existing denoising algorithms are developed for grayscale images,
while it is not a trivial work to extend them for color image denoising because
the noise statistics in R, G, B channels can be very different for real noisy
images. In this paper, we propose a multi-channel (MC) optimization model for
real color image denoising under the weighted nuclear norm minimization (WNNM)
framework. We concatenate the RGB patches to make use of the channel
redundancy, and introduce a weight matrix to balance the data fidelity of the
three channels in consideration of their different noise statistics. The
proposed MC-WNNM model does not have an analytical solution. We reformulate it
into a linear equality-constrained problem and solve it with the alternating
direction method of multipliers. Each alternative updating step has closed-form
solution and the convergence can be guaranteed. Extensive experiments on both
synthetic and real noisy image datasets demonstrate the superiority of the
proposed MC-WNNM over state-of-the-art denoising methods. | [
"cs.CV"
] |
Machine Learning (ML) models are being used in all facets of today's society
to make high stake decisions like bail granting or credit lending, with very
minimal regulations. Such systems are extremely vulnerable to both propagating
and amplifying social biases, and have therefore been subject to growing
research interest. One of the main issues with conventional fairness metrics is
their narrow definitions which hide the complete extent of the bias by focusing
primarily on positive and/or negative outcomes, whilst not paying attention to
the overall distributional shape. Moreover, these metrics are often
contradictory to each other, are severely restrained by the contextual and
legal landscape of the problem, have technical constraints like poor support
for continuous outputs, the requirement of class labels, and are not
explainable.
In this paper, we present Quantile Demographic Drift, which addresses the
shortcomings mentioned above. This metric can also be used to measure
intra-group privilege. It is easily interpretable via existing attribution
techniques, and also extends naturally to individual fairness via the principle
of like-for-like comparison. We make this new fairness score the basis of a new
system that is designed to detect bias in production ML models without the need
for labels. We call the system FairCanary because of its capability to detect
bias in a live deployed model and narrow down the alert to the responsible set
of features, like the proverbial canary in a coal mine. | [
"cs.LG",
"cs.CY"
] |
This work addresses a new problem that learns generative adversarial networks
(GANs) from multiple data collections that are each i) owned separately by
different clients and ii) drawn from a non-identical distribution that
comprises different classes. Given such non-iid data as input, we aim to learn
a distribution involving all the classes input data can belong to, while
keeping the data decentralized in each client storage. Our key contribution to
this end is a new decentralized approach for learning GANs from non-iid data
called Forgiver-First Update (F2U), which a) asks clients to train an
individual discriminator with their own data and b) updates a generator to fool
the most `forgiving' discriminators who deem generated samples as the most
real. Our theoretical analysis proves that this updating strategy allows the
decentralized GAN to achieve a generator's distribution with all the input
classes as its global optimum based on f-divergence minimization. Moreover, we
propose a relaxed version of F2U called Forgiver-First Aggregation (F2A) that
performs well in practice, which adaptively aggregates the discriminators while
emphasizing forgiving ones. Our empirical evaluations with image generation
tasks demonstrated the effectiveness of our approach over state-of-the-art
decentralized learning methods. | [
"cs.LG",
"stat.ML"
] |
The concept of utilizing multi-step returns for updating value functions has
been adopted in deep reinforcement learning (DRL) for a number of years.
Updating value functions with different backup lengths provides advantages in
different aspects, including bias and variance of value estimates, convergence
speed, and exploration behavior of the agent. Conventional methods such as
TD-lambda leverage these advantages by using a target value equivalent to an
exponential average of different step returns. Nevertheless, integrating step
returns into a single target sacrifices the diversity of the advantages offered
by different step return targets. To address this issue, we propose Mixture
Bootstrapped DQN (MB-DQN) built on top of bootstrapped DQN, and uses different
backup lengths for different bootstrapped heads. MB-DQN enables heterogeneity
of the target values that is unavailable in approaches relying only on a single
target value. As a result, it is able to maintain the advantages offered by
different backup lengths. In this paper, we first discuss the motivational
insights through a simple maze environment. In order to validate the
effectiveness of MB-DQN, we perform experiments on the Atari 2600 benchmark
environments, and demonstrate the performance improvement of MB-DQN over a
number of baseline methods. We further provide a set of ablation studies to
examine the impacts of different design configurations of MB-DQN. | [
"cs.LG",
"cs.AI",
"stat.ML"
] |
Semi-Supervised Learning (SSL) has seen success in many application domains,
but this success often hinges on the availability of task-specific unlabeled
data. Knowledge distillation (KD) has enabled compressing deep networks and
ensembles, achieving the best results when distilling knowledge on fresh
task-specific unlabeled examples. However, task-specific unlabeled data can be
challenging to find. We present a general framework called "generate, annotate,
and learn (GAL)" that uses unconditional generative models to synthesize
in-domain unlabeled data, helping advance SSL and KD on different tasks. To
obtain strong task-specific generative models, we adopt generic generative
models, pretrained on open-domain data, and fine-tune them on inputs from
specific tasks. Then, we use existing classifiers to annotate generated
unlabeled examples with soft pseudo labels, which are used for additional
training. When self-training is combined with samples generated from
GPT2-large, fine-tuned on the inputs of each GLUE task, we outperform a strong
RoBERTa-large baseline on the GLUE benchmark. Moreover, KD on GPT-2 samples
yields a new state-of-the-art for 6-layer transformers on the GLUE leaderboard.
Finally, self-training with GAL offers significant gains on image
classification on CIFAR-10 and four tabular tasks from the UCI repository | [
"cs.LG"
] |
Person re-identification (re-id) aims to match the same person from images
taken across multiple cameras. Most existing person re-id methods generally
require a large amount of identity labeled data to act as discriminative
guideline for representation learning. Difficulty in manually collecting
identity labeled data leads to poor adaptability in practical scenarios. To
overcome this problem, we propose an unsupervised center-based clustering
approach capable of progressively learning and exploiting the underlying re-id
discriminative information from temporal continuity within a camera. We call
our framework Temporal Continuity based Unsupervised Learning (TCUL).
Specifically, TCUL simultaneously does center based clustering of unlabeled
(target) dataset and fine-tunes a convolutional neural network (CNN)
pre-trained on irrelevant labeled (source) dataset to enhance discriminative
capability of the CNN for the target dataset. Furthermore, it exploits
temporally continuous nature of images within-camera jointly with spatial
similarity of feature maps across-cameras to generate reliable pseudo-labels
for training a re-identification model. As the training progresses, number of
reliable samples keep on growing adaptively which in turn boosts representation
ability of the CNN. Extensive experiments on three large-scale person re-id
benchmark datasets are conducted to compare our framework with state-of-the-art
techniques, which demonstrate superiority of TCUL over existing methods. | [
"cs.CV"
] |
Learning Enabled Components (LECs) are widely being used in a variety of
perception based autonomy tasks like image segmentation, object detection,
end-to-end driving, etc. These components are trained with large image datasets
with multimodal factors like weather conditions, time-of-day, traffic-density,
etc. The LECs learn from these factors during training, and while testing if
there is variation in any of these factors, the components get confused
resulting in low confidence predictions. The images with factors not seen
during training is commonly referred to as Out-of-Distribution (OOD). For safe
autonomy it is important to identify the OOD images, so that a suitable
mitigation strategy can be performed. Classical one-class classifiers like SVM
and SVDD are used to perform OOD detection. However, the multiple labels
attached to the images in these datasets, restricts the direct application of
these techniques. We address this problem using the latent space of the
$\beta$-Variational Autoencoder ($\beta$-VAE). We use the fact that compact
latent space generated by an appropriately selected $\beta$-VAE will encode the
information about these factors in a few latent variables, and that can be used
for computationally inexpensive detection. We evaluate our approach on the
nuScenes dataset, and our results shows the latent space of $\beta$-VAE is
sensitive to encode changes in the values of the generative factor. | [
"cs.CV",
"cs.LG"
] |
To perform well on unseen and potentially out-of-distribution samples, it is
desirable for machine learning models to have a predictable response with
respect to transformations affecting the factors of variation of the input.
Invariance is commonly achieved through hand-engineered data augmentation, but
do standard data augmentations address transformations that explain variations
in real data? While prior work has focused on synthetic data, we attempt here
to characterize the factors of variation in a real dataset, ImageNet, and study
the invariance of both standard residual networks and the recently proposed
vision transformer with respect to changes in these factors. We show standard
augmentation relies on a precise combination of translation and scale, with
translation recapturing most of the performance improvement -- despite the
(approximate) translation invariance built in to convolutional architectures,
such as residual networks. In fact, we found that scale and translation
invariance was similar across residual networks and vision transformer models
despite their markedly different inductive biases. We show the training data
itself is the main source of invariance, and that data augmentation only
further increases the learned invariances. Interestingly, the invariances
brought from the training process align with the ImageNet factors of variation
we found. Finally, we find that the main factors of variation in ImageNet
mostly relate to appearance and are specific to each class. | [
"cs.CV"
] |
This paper presents an end-to-end differentiable algorithm for robust and
detail-preserving surface normal estimation on unstructured point-clouds. We
utilize graph neural networks to iteratively parameterize an adaptive
anisotropic kernel that produces point weights for weighted least-squares plane
fitting in local neighborhoods. The approach retains the interpretability and
efficiency of traditional sequential plane fitting while benefiting from
adaptation to data set statistics through deep learning. This results in a
state-of-the-art surface normal estimator that is robust to noise, outliers and
point density variation, preserves sharp features through anisotropic kernels
and equivariance through a local quaternion-based spatial transformer. Contrary
to previous deep learning methods, the proposed approach does not require any
hand-crafted features or preprocessing. It improves on the state-of-the-art
results while being more than two orders of magnitude faster and more parameter
efficient. | [
"cs.CV",
"cs.CG"
] |
It has been well recognized that modeling human-object or object-object
relations would be helpful for detection task. Nevertheless, the problem is not
trivial especially when exploring the interactions between human actor, object
and scene (collectively as human-context) to boost video action detectors. The
difficulty originates from the aspect that reliable relations in a video should
depend on not only short-term human-context relation in the present clip but
also the temporal dynamics distilled over a long-range span of the video. This
motivates us to capture both short-term and long-term relations in a video. In
this paper, we present a new Long Short-Term Relation Networks, dubbed as LSTR,
that novelly aggregates and propagates relation to augment features for video
action detection. Technically, Region Proposal Networks (RPN) is remoulded to
first produce 3D bounding boxes, i.e., tubelets, in each video clip. LSTR then
models short-term human-context interactions within each clip through
spatio-temporal attention mechanism and reasons long-term temporal dynamics
across video clips via Graph Convolutional Networks (GCN) in a cascaded manner.
Extensive experiments are conducted on four benchmark datasets, and superior
results are reported when comparing to state-of-the-art methods. | [
"cs.CV"
] |
Reinforcement Learning (RL) has made remarkable achievements, but it still
suffers from inadequate exploration strategies, sparse reward signals, and
deceptive reward functions. These problems motivate the need for a more
efficient and directed exploration. For solving this, a Population-guided
Novelty Search (PNS) parallel learning method is proposed. In PNS, the
population is divided into multiple sub-populations, each of which has one
chief agent and several exploring agents. The role of the chief agent is to
evaluate the policies learned by exploring agents and to share the optimal
policy with all sub-populations. The role of exploring agents is to learn their
policies in collaboration with the guidance of the optimal policy and,
simultaneously, upload their policies to the chief agent. To balance
exploration and exploitation, the Novelty Search (NS) is employed in chief
agents to encourage policies with high novelty while maximizing per-episode
performance. The introduction of sub-populations and NS mechanisms promote
directed exploration and enables better policy search. In the numerical
experiment section, the proposed scheme is applied to the twin delayed deep
deterministic (TD3) policy gradient algorithm, and the effectiveness of PNS to
promote exploration and improve performance in both continuous control domains
and discrete control domains is demonstrated. Notably, the proposed method
achieves rewards that far exceed the SOTA methods in Delayed MoJoCo
environments. | [
"cs.LG",
"cs.AI"
] |
Reinforcement learning agents are faced with two types of uncertainty.
Epistemic uncertainty stems from limited data and is useful for exploration,
whereas aleatoric uncertainty arises from stochastic environments and must be
accounted for in risk-sensitive applications. We highlight the challenges
involved in simultaneously estimating both of them, and propose a framework for
disentangling and estimating these uncertainties on learned Q-values. We derive
unbiased estimators of these uncertainties and introduce an uncertainty-aware
DQN algorithm, which we show exhibits safe learning behavior and outperforms
other DQN variants on the MinAtar testbed. | [
"cs.LG",
"cs.AI",
"stat.ML"
] |
Traditional approaches to interpolate/extrapolate frames in a video sequence
require accurate pixel correspondences between images, e.g., using optical
flow. Their results stem on the accuracy of optical flow estimation, and could
generate heavy artifacts when flow estimation failed. Recently methods using
auto-encoder has shown impressive progress, however they are usually trained
for specific interpolation/extrapolation settings and lack of flexibility and
In order to reduce these limitations, we propose a unified network to
parameterize the interest frame position and therefore infer
interpolate/extrapolate frames within the same framework. To achieve this, we
introduce a transitive consistency loss to better regularize the network. We
adopt a multi-scale structure for the network so that the parameters can be
shared across multi-layers. Our approach avoids expensive global optimization
of optical flow methods, and is efficient and flexible for video
interpolation/extrapolation applications. Experimental results have shown that
our method performs favorably against state-of-the-art methods. | [
"cs.CV"
] |
There has been a surge of recent interest in learning representations for
graph-structured data. Graph representation learning methods have generally
fallen into three main categories, based on the availability of labeled data.
The first, network embedding (such as shallow graph embedding or graph
auto-encoders), focuses on learning unsupervised representations of relational
structure. The second, graph regularized neural networks, leverages graphs to
augment neural network losses with a regularization objective for
semi-supervised learning. The third, graph neural networks, aims to learn
differentiable functions over discrete topologies with arbitrary structure.
However, despite the popularity of these areas there has been surprisingly
little work on unifying the three paradigms. Here, we aim to bridge the gap
between graph neural networks, network embedding and graph regularization
models. We propose a comprehensive taxonomy of representation learning methods
for graph-structured data, aiming to unify several disparate bodies of work.
Specifically, we propose a Graph Encoder Decoder Model (GRAPHEDM), which
generalizes popular algorithms for semi-supervised learning on graphs (e.g.
GraphSage, Graph Convolutional Networks, Graph Attention Networks), and
unsupervised learning of graph representations (e.g. DeepWalk, node2vec, etc)
into a single consistent approach. To illustrate the generality of this
approach, we fit over thirty existing methods into this framework. We believe
that this unifying view both provides a solid foundation for understanding the
intuition behind these methods, and enables future research in the area. | [
"cs.LG",
"cs.NE",
"cs.SI",
"stat.ML"
] |
Visual surveillance aims to stably detect a foreground object using a
continuous image acquired from a fixed camera. Recent deep learning methods
based on supervised learning show superior performance compared to classical
background subtraction algorithms. However, there is still a room for
improvement in static foreground, dynamic background, hard shadow, illumination
changes, camouflage, etc. In addition, most of the deep learning-based methods
operates well on environments similar to training. If the testing environments
are different from training ones, their performance degrades. As a result,
additional training on those operating environments is required to ensure a
good performance. Our previous work which uses spatio-temporal input data
consisted of a number of past images, background images and current image
showed promising results in different environments from training, although it
uses a simple U-NET structure. In this paper, we propose a data augmentation
technique suitable for visual surveillance for additional performance
improvement using the same network used in our previous work. In deep learning,
most data augmentation techniques deal with spatial-level data augmentation
techniques for use in image classification and object detection. In this paper,
we propose a new method of data augmentation in the spatio-temporal dimension
suitable for our previous work. Two data augmentation methods of adjusting
background model images and past images are proposed. Through this, it is shown
that performance can be improved in difficult areas such as static foreground
and ghost objects, compared to previous studies. Through quantitative and
qualitative evaluation using SBI, LASIESTA, and our own dataset, we show that
it gives superior performance compared to deep learning-based algorithms and
background subtraction algorithms. | [
"cs.CV"
] |
Generalization and adaptation of learned skills to novel situations is a core
requirement for intelligent autonomous robots. Although contextual
reinforcement learning provides a principled framework for learning and
generalization of behaviors across related tasks, it generally relies on
uninformed sampling of environments from an unknown, uncontrolled context
distribution, thus missing the benefits of structured, sequential learning. We
introduce a novel relative entropy reinforcement learning algorithm that gives
the agent the freedom to control the intermediate task distribution, allowing
for its gradual progression towards the target context distribution. Empirical
evaluation shows that the proposed curriculum learning scheme drastically
improves sample efficiency and enables learning in scenarios with both broad
and sharp target context distributions in which classical approaches perform
sub-optimally. | [
"cs.LG",
"stat.ML"
] |
Facial attractiveness enhancement has been an interesting application in
Computer Vision and Graphics over these years. It aims to generate a more
attractive face via manipulations on image and geometry structure while
preserving face identity. In this paper, we propose the first Generative
Adversarial Networks (GANs) for enhancing facial attractiveness in both
geometry and appearance aspects, which we call "FA-GANs". FA-GANs contain two
branches and enhance facial attractiveness in two perspectives: facial geometry
and facial appearance. Each branch consists of individual GANs with the
appearance branch adjusting the facial image and the geometry branch adjusting
the facial landmarks in appearance and geometry aspects, respectively. Unlike
the traditional facial manipulations learning from paired faces, which are
infeasible to collect before and after enhancement of the same individual, we
achieve this by learning the features of attractiveness faces through
unsupervised adversarial learning. The proposed FA-GANs are able to extract
attractiveness features and impose them on the enhancement results. To better
enhance faces, both the geometry and appearance networks are considered to
refine the facial attractiveness by adjusting the geometry layout of faces and
the appearance of faces independently. To the best of our knowledge, we are the
first to enhance the facial attractiveness with GANs in both geometry and
appearance aspects. The experimental results suggest that our FA-GANs can
generate compelling perceptual results in both geometry structure and facial
appearance and outperform current state-of-the-art methods. | [
"cs.CV"
] |
A major component of overfitting in model-free reinforcement learning (RL)
involves the case where the agent may mistakenly correlate reward with certain
spurious features from the observations generated by the Markov Decision
Process (MDP). We provide a general framework for analyzing this scenario,
which we use to design multiple synthetic benchmarks from only modifying the
observation space of an MDP. When an agent overfits to different observation
spaces even if the underlying MDP dynamics is fixed, we term this observational
overfitting. Our experiments expose intriguing properties especially with
regards to implicit regularization, and also corroborate results from previous
works in RL generalization and supervised learning (SL). | [
"cs.LG",
"cs.AI",
"stat.ML"
] |
In some memory-constrained settings like IoT devices and over-the-network
data pipelines, it can be advantageous to have smaller contextual embeddings.
We investigate the efficacy of projecting contextual embedding data (BERT) onto
a manifold, and using nonlinear dimensionality reduction techniques to compress
these embeddings. In particular, we propose a novel post-processing approach,
applying a combination of Isomap and PCA. We find that the geodesic distance
estimations, estimates of the shortest path on a Riemannian manifold, from
Isomap's k-Nearest Neighbors graph bolstered the performance of the compressed
embeddings to be comparable to the original BERT embeddings. On one dataset, we
find that despite a 12-fold dimensionality reduction, the compressed embeddings
performed within 0.1% of the original BERT embeddings on a downstream
classification task. In addition, we find that this approach works particularly
well on tasks reliant on syntactic data, when compared with linear
dimensionality reduction. These results show promise for a novel geometric
approach to achieve lower dimensional text embeddings from existing
transformers and pave the way for data-specific and application-specific
embedding compressions. | [
"cs.LG",
"cs.CL"
] |
Consider a two-class classification problem where the number of features is
much larger than the sample size. The features are masked by Gaussian noise
with mean zero and covariance matrix $\Sigma$, where the precision matrix
$\Omega=\Sigma^{-1}$ is unknown but is presumably sparse. The useful features,
also unknown, are sparse and each contributes weakly (i.e., rare and weak) to
the classification decision. By obtaining a reasonably good estimate of
$\Omega$, we formulate the setting as a linear regression model. We propose a
two-stage classification method where we first select features by the method of
Innovated Thresholding (IT), and then use the retained features and Fisher's
LDA for classification. In this approach, a crucial problem is how to set the
threshold of IT. We approach this problem by adapting the recent innovation of
Higher Criticism Thresholding (HCT). We find that when useful features are rare
and weak, the limiting behavior of HCT is essentially just as good as the
limiting behavior of ideal threshold, the threshold one would choose if the
underlying distribution of the signals is known (if only). Somewhat
surprisingly, when $\Omega$ is sufficiently sparse, its off-diagonal
coordinates usually do not have a major influence over the classification
decision. Compared to recent work in the case where $\Omega$ is the identity
matrix [Proc. Natl. Acad. Sci. USA 105 (2008) 14790-14795; Philos. Trans. R.
Soc. Lond. Ser. A Math. Phys. Eng. Sci. 367 (2009) 4449-4470], the current
setting is much more general, which needs a new approach and much more
sophisticated analysis. One key component of the analysis is the intimate
relationship between HCT and Fisher's separation. Another key component is the
tight large-deviation bounds for empirical processes for data with
unconventional correlation structures, where graph theory on vertex coloring
plays an important role. | [
"stat.ML",
"math.ST",
"stat.TH"
] |
The "interpretation through synthesis" approach to analyze face images,
particularly Active Appearance Models (AAMs) method, has become one of the most
successful face modeling approaches over the last two decades. AAM models have
ability to represent face images through synthesis using a controllable
parameterized Principal Component Analysis (PCA) model. However, the accuracy
and robustness of the synthesized faces of AAM are highly depended on the
training sets and inherently on the generalizability of PCA subspaces. This
paper presents a novel Deep Appearance Models (DAMs) approach, an efficient
replacement for AAMs, to accurately capture both shape and texture of face
images under large variations. In this approach, three crucial components
represented in hierarchical layers are modeled using the Deep Boltzmann
Machines (DBM) to robustly capture the variations of facial shapes and
appearances. DAMs are therefore superior to AAMs in inferencing a
representation for new face images under various challenging conditions. The
proposed approach is evaluated in various applications to demonstrate its
robustness and capabilities, i.e. facial super-resolution reconstruction,
facial off-angle reconstruction or face frontalization, facial occlusion
removal and age estimation using challenging face databases, i.e. Labeled Face
Parts in the Wild (LFPW), Helen and FG-NET. Comparing to AAMs and other deep
learning based approaches, the proposed DAMs achieve competitive results in
those applications, thus this showed their advantages in handling occlusions,
facial representation, and reconstruction. | [
"cs.CV"
] |
Despite the fact that deep reinforcement learning (RL) has surpassed
human-level performances in various tasks, it still has several fundamental
challenges. First, most RL methods require intensive data from the exploration
of the environment to achieve satisfactory performance. Second, the use of
neural networks in RL renders it hard to interpret the internals of the system
in a way that humans can understand. To address these two challenges, we
propose a framework that enables an RL agent to reason over its exploration
process and distill high-level knowledge for effectively guiding its future
explorations. Specifically, we propose a novel RL algorithm that learns
high-level knowledge in the form of a finite reward automaton by using the L*
learning algorithm. We prove that in episodic RL, a finite reward automaton can
express any non-Markovian bounded reward functions with finitely many reward
values and approximate any non-Markovian bounded reward function (with
infinitely many reward values) with arbitrary precision. We also provide a
lower bound for the episode length such that the proposed RL approach almost
surely converges to an optimal policy in the limit. We test this approach on
two RL environments with non-Markovian reward functions, choosing a variety of
tasks with increasing complexity for each environment. We compare our algorithm
with the state-of-the-art RL algorithms for non-Markovian reward functions,
such as Joint Inference of Reward machines and Policies for RL (JIRP), Learning
Reward Machine (LRM), and Proximal Policy Optimization (PPO2). Our results show
that our algorithm converges to an optimal policy faster than other baseline
methods. | [
"cs.LG",
"cs.AI"
] |
State-of-the-art face super-resolution methods employ deep convolutional
neural networks to learn a mapping between low- and high- resolution facial
patterns by exploring local appearance knowledge. However, most of these
methods do not well exploit facial structures and identity information, and
struggle to deal with facial images that exhibit large pose variations. In this
paper, we propose a novel face super-resolution method that explicitly
incorporates 3D facial priors which grasp the sharp facial structures. Our work
is the first to explore 3D morphable knowledge based on the fusion of
parametric descriptions of face attributes (e.g., identity, facial expression,
texture, illumination, and face pose). Furthermore, the priors can easily be
incorporated into any network and are extremely efficient in improving the
performance and accelerating the convergence speed. Firstly, a 3D face
rendering branch is set up to obtain 3D priors of salient facial structures and
identity knowledge. Secondly, the Spatial Attention Module is used to better
exploit this hierarchical information (i.e., intensity similarity, 3D facial
structure, and identity content) for the super-resolution problem. Extensive
experiments demonstrate that the proposed 3D priors achieve superior face
super-resolution results over the state-of-the-arts. | [
"cs.CV",
"cs.AI",
"cs.LG",
"eess.IV"
] |
In this paper we present a novel approach for depth map enhancement from an
RGB-D video sequence. The basic idea is to exploit the shading information in
the color image. Instead of making assumption about surface albedo or
controlled object motion and lighting, we use the lighting variations
introduced by casual object movement. We are effectively calculating
photometric stereo from a moving object under natural illuminations. The key
technical challenge is to establish correspondences over the entire image set.
We therefore develop a lighting insensitive robust pixel matching technique
that out-performs optical flow method in presence of lighting variations. In
addition we present an expectation-maximization framework to recover the
surface normal and albedo simultaneously, without any regularization term. We
have validated our method on both synthetic and real datasets to show its
superior performance on both surface details recovery and intrinsic
decomposition. | [
"cs.CV"
] |
Detecting customized moments and highlights from videos given natural
language (NL) user queries is an important but under-studied topic. One of the
challenges in pursuing this direction is the lack of annotated data. To address
this issue, we present the Query-based Video Highlights (QVHighlights) dataset.
It consists of over 10,000 YouTube videos, covering a wide range of topics,
from everyday activities and travel in lifestyle vlog videos to social and
political activities in news videos. Each video in the dataset is annotated
with: (1) a human-written free-form NL query, (2) relevant moments in the video
w.r.t. the query, and (3) five-point scale saliency scores for all
query-relevant clips. This comprehensive annotation enables us to develop and
evaluate systems that detect relevant moments as well as salient highlights for
diverse, flexible user queries. We also present a strong baseline for this
task, Moment-DETR, a transformer encoder-decoder model that views moment
retrieval as a direct set prediction problem, taking extracted video and query
representations as inputs and predicting moment coordinates and saliency scores
end-to-end. While our model does not utilize any human prior, we show that it
performs competitively when compared to well-engineered architectures. With
weakly supervised pretraining using ASR captions, Moment-DETR substantially
outperforms previous methods. Lastly, we present several ablations and
visualizations of Moment-DETR. Data and code is publicly available at
https://github.com/jayleicn/moment_detr | [
"cs.CV",
"cs.AI",
"cs.CL"
] |
Self-attention techniques, and specifically Transformers, are dominating the
field of text processing and are becoming increasingly popular in computer
vision classification tasks. In order to visualize the parts of the image that
led to a certain classification, existing methods either rely on the obtained
attention maps or employ heuristic propagation along the attention graph. In
this work, we propose a novel way to compute relevancy for Transformer
networks. The method assigns local relevance based on the Deep Taylor
Decomposition principle and then propagates these relevancy scores through the
layers. This propagation involves attention layers and skip connections, which
challenge existing methods. Our solution is based on a specific formulation
that is shown to maintain the total relevancy across layers. We benchmark our
method on very recent visual Transformer networks, as well as on a text
classification problem, and demonstrate a clear advantage over the existing
explainability methods. | [
"cs.CV"
] |
Compared with Generative Adversarial Networks (GAN), Energy-Based generative
Models (EBMs) possess two appealing properties: i) they can be directly
optimized without requiring an auxiliary network during the learning and
synthesizing; ii) they can better approximate underlying distribution of the
observed data by learning explicitly potential functions. This paper studies a
branch of EBMs, i.e., energy-based Generative ConvNets (GCNs), which minimize
their energy function defined by a bottom-up ConvNet. From the perspective of
particle physics, we solve the problem of unstable energy dissipation that
might damage the quality of the synthesized samples during the maximum
likelihood learning. Specifically, we firstly establish a connection between
classical FRAME model [1] and dynamic physics process and generalize the GCN in
discrete flow with a certain metric measure from particle perspective. To
address KL-vanishing issue, we then reformulate GCN from the KL discrete flow
with KL divergence measure to a Jordan-Kinderleher-Otto (JKO) discrete flow
with Wasserastein distance metric and derive a Wasserastein GCN (wGCN). Based
on these theoretical studies on GCN, we finally derive a Generalized GCN (GGCN)
to further improve the model generalization and learning capability. GGCN
introduces a hidden space mapping strategy by employing a normal distribution
for the reference distribution to address the learning bias issue. Due to MCMC
sampling in GCNs, it still suffers from a serious time-consuming issue when
sampling steps increase; thus a trainable non-linear upsampling function and an
amortized learning are proposed to improve the learning efficiency. Our
proposed GGCN is trained in a symmetrical learning manner. Our method surpass
the existing models in both model stability and the quality of generated
samples on several widely-used face and natural image datasets. | [
"cs.CV"
] |
We present a new multi-stream 3D mesh reconstruction network (MSMR-Net) for
hand pose estimation from a single RGB image. Our model consists of an image
encoder followed by a mesh-convolution decoder composed of connected graph
convolution layers. In contrast to previous models that form a single mesh
decoding path, our decoder network incorporates multiple cross-resolution
trajectories that are executed in parallel. Thus, global and local information
are shared to form rich decoding representations at minor additional parameter
cost compared to the single trajectory network. We demonstrate the
effectiveness of our method in hand-hand and hand-object interaction scenarios
at various levels of interaction. To evaluate the former scenario, we propose a
method to generate RGB images of closely interacting hands. Moreoever, we
suggest a metric to quantify the degree of interaction and show that close hand
interactions are particularly challenging. Experimental results show that the
MSMR-Net outperforms existing algorithms on the hand-object FreiHAND dataset as
well as on our own hand-hand dataset. | [
"cs.CV"
] |
We present RangeRCNN, a novel and effective 3D object detection framework
based on the range image representation. Most existing methods are voxel-based
or point-based. Though several optimizations have been introduced to ease the
sparsity issue and speed up the running time, the two representations are still
computationally inefficient. Compared to them, the range image representation
is dense and compact which can exploit powerful 2D convolution. Even so, the
range image is not preferred in 3D object detection due to scale variation and
occlusion. In this paper, we utilize the dilated residual block (DRB) to better
adapt different object scales and obtain a more flexible receptive field.
Considering scale variation and occlusion, we propose the RV-PV-BEV (range
view-point view-bird's eye view) module to transfer features from RV to BEV.
The anchor is defined in BEV which avoids scale variation and occlusion.
Neither RV nor BEV can provide enough information for height estimation;
therefore, we propose a two-stage RCNN for better 3D detection performance. The
aforementioned point view not only serves as a bridge from RV to BEV but also
provides pointwise features for RCNN. Experiments show that RangeRCNN achieves
state-of-the-art performance on the KITTI dataset and the Waymo Open dataset,
and provides more possibilities for real-time 3D object detection. We further
introduce and discuss the data augmentation strategy for the range image based
method, which will be very valuable for future research on range image. | [
"cs.CV"
] |
Classification and differentiation of small pathological objects may greatly
vary among human raters due to differences in training, expertise and their
consistency over time. In a radiological setting, objects commonly have high
within-class appearance variability whilst sharing certain characteristics
across different classes, making their distinction even more difficult. As an
example, markers of cerebral small vessel disease, such as enlarged
perivascular spaces (EPVS) and lacunes, can be very varied in their appearance
while exhibiting high inter-class similarity, making this task highly
challenging for human raters. In this work, we investigate joint models of
individual rater behaviour and multirater consensus in a deep learning setting,
and apply it to a brain lesion object-detection task. Results show that jointly
modelling both individual and consensus estimates leads to significant
improvements in performance when compared to directly predicting consensus
labels, while also allowing the characterization of human-rater consistency. | [
"cs.CV",
"cs.AI"
] |
Piecewise constant image approximations of sequential number of segments or
clusters of disconnected pixels are treated. The method of majorizing of
optimal approximation sequence by hierarchical sequence of image approximations
is proposed. A generalization for multidimensional case of color and
multispectral images is foreseen. | [
"cs.CV"
] |
In this paper, we propose a novel iterative convolution-thresholding method
(ICTM) that is applicable to a range of variational models for image
segmentation. A variational model usually minimizes an energy functional
consisting of a fidelity term and a regularization term. In the ICTM, the
interface between two different segment domains is implicitly represented by
their characteristic functions. The fidelity term is then usually written as a
linear functional of the characteristic functions and the regularized term is
approximated by a functional of characteristic functions in terms of heat
kernel convolution. This allows us to design an iterative
convolution-thresholding method to minimize the approximate energy. The method
is simple, efficient and enjoys the energy-decaying property. Numerical
experiments show that the method is easy to implement, robust and applicable to
various image segmentation models. | [
"cs.CV"
] |
Single image inverse problem is a notoriously challenging ill-posed problem
that aims to restore the original image from one of its corrupted versions.
Recently, this field has been immensely influenced by the emergence of
deep-learning techniques. Deep Image Prior (DIP) offers a new approach that
forces the recovered image to be synthesized from a given deep architecture.
While DIP is quite an effective unsupervised approach, it is deprecated in
real-world applications because of the requirement of human assistance. In this
work, we aim to find the best-recovered image without the assistance of humans
by adding a stopping criterion, which will reach maximum when the iteration no
longer improves the image quality. More specifically, we propose to add a
pseudo noise to the corrupted image and measure the pseudo-noise component in
the recovered image by the orthogonality between signal and noise. The accuracy
of the orthogonal stopping criterion has been demonstrated for several tested
problems such as denoising, super-resolution, and inpainting, in which 38 out
of 40 experiments are higher than 95%. | [
"cs.CV",
"eess.IV"
] |
We present lambda layers -- an alternative framework to self-attention -- for
capturing long-range interactions between an input and structured contextual
information (e.g. a pixel surrounded by other pixels). Lambda layers capture
such interactions by transforming available contexts into linear functions,
termed lambdas, and applying these linear functions to each input separately.
Similar to linear attention, lambda layers bypass expensive attention maps, but
in contrast, they model both content and position-based interactions which
enables their application to large structured inputs such as images. The
resulting neural network architectures, LambdaNetworks, significantly
outperform their convolutional and attentional counterparts on ImageNet
classification, COCO object detection and COCO instance segmentation, while
being more computationally efficient. Additionally, we design LambdaResNets, a
family of hybrid architectures across different scales, that considerably
improves the speed-accuracy tradeoff of image classification models.
LambdaResNets reach excellent accuracies on ImageNet while being 3.2 - 4.4x
faster than the popular EfficientNets on modern machine learning accelerators.
When training with an additional 130M pseudo-labeled images, LambdaResNets
achieve up to a 9.5x speed-up over the corresponding EfficientNet checkpoints. | [
"cs.CV",
"cs.LG"
] |
Visual steel surface defect detection is an essential step in steel sheet
manufacturing. Several machine learning-based automated visual inspection (AVI)
methods have been studied in recent years. However, most steel manufacturing
industries still use manual visual inspection due to training time and
inaccuracies involved with AVI methods. Automatic steel defect detection
methods could be useful in less expensive and faster quality control and
feedback. But preparing the annotated training data for segmentation and
classification could be a costly process. In this work, we propose to use the
Transfer Learning-based U-Net (TLU-Net) framework for steel surface defect
detection. We use a U-Net architecture as the base and explore two kinds of
encoders: ResNet and DenseNet. We compare these nets' performance using random
initialization and the pre-trained networks trained using the ImageNet data
set. The experiments are performed using Severstal data. The results
demonstrate that the transfer learning performs 5% (absolute) better than that
of the random initialization in defect classification. We found that the
transfer learning performs 26% (relative) better than that of the random
initialization in defect segmentation. We also found the gain of transfer
learning increases as the training data decreases, and the convergence rate
with transfer learning is better than that of the random initialization. | [
"cs.CV",
"cs.AI",
"cs.LG",
"eess.IV"
] |
Graph representation learning has attracted increasing research attention.
However, most existing studies fuse all structural features and node attributes
to provide an overarching view of graphs, neglecting finer substructures'
semantics, and suffering from interpretation enigmas. This paper presents a
novel hierarchical subgraph-level selection and embedding based graph neural
network for graph classification, namely SUGAR, to learn more discriminative
subgraph representations and respond in an explanatory way. SUGAR reconstructs
a sketched graph by extracting striking subgraphs as the representative part of
the original graph to reveal subgraph-level patterns. To adaptively select
striking subgraphs without prior knowledge, we develop a reinforcement pooling
mechanism, which improves the generalization ability of the model. To
differentiate subgraph representations among graphs, we present a
self-supervised mutual information mechanism to encourage subgraph embedding to
be mindful of the global graph structural properties by maximizing their mutual
information. Extensive experiments on six typical bioinformatics datasets
demonstrate a significant and consistent improvement in model quality with
competitive performance and interpretability. | [
"cs.LG",
"cs.AI"
] |
Recurrent Neural Networks (RNNs) have had considerable success in classifying
and predicting sequences. We demonstrate that RNNs can be effectively used in
order to encode sequences and provide effective representations. The
methodology we use is based on Fisher Vectors, where the RNNs are the
generative probabilistic models and the partial derivatives are computed using
backpropagation. State of the art results are obtained in two central but
distant tasks, which both rely on sequences: video action recognition and image
annotation. We also show a surprising transfer learning result from the task of
image annotation to the task of video action recognition. | [
"cs.CV"
] |
Generative Adversarial Networks (GANs) have become a very popular tool for
implicitly learning high-dimensional probability distributions. Several
improvements have been made to the original GAN formulation to address some of
its shortcomings like mode collapse, convergence issues, entanglement, poor
visual quality etc. While a significant effort has been directed towards
improving the visual quality of images generated by GANs, it is rather
surprising that objective image quality metrics have neither been employed as
cost functions nor as regularizers in GAN objective functions. In this work, we
show how a distance metric that is a variant of the Structural SIMilarity
(SSIM) index (a popular full-reference image quality assessment algorithm), and
a novel quality aware discriminator gradient penalty function that is inspired
by the Natural Image Quality Evaluator (NIQE, a popular no-reference image
quality assessment algorithm) can each be used as excellent regularizers for
GAN objective functions. Specifically, we demonstrate state-of-the-art
performance using the Wasserstein GAN gradient penalty (WGAN-GP) framework over
CIFAR-10, STL10 and CelebA datasets. | [
"cs.CV",
"eess.IV"
] |
In this paper, we elaborate on the use of the Sugeno integral in the context
of machine learning. More specifically, we propose a method for binary
classification, in which the Sugeno integral is used as an aggregation function
that combines several local evaluations of an instance, pertaining to different
features or measurements, into a single global evaluation. Due to the specific
nature of the Sugeno integral, this approach is especially suitable for
learning from ordinal data, that is, when measurements are taken from ordinal
scales. This is a topic that has not received much attention in machine
learning so far. The core of the learning problem itself consists of
identifying the capacity underlying the Sugeno integral. To tackle this
problem, we develop an algorithm based on linear programming. The algorithm
also includes a suitable technique for transforming the original feature values
into local evaluations (local utility scores), as well as a method for tuning a
threshold on the global evaluation. To control the flexibility of the
classifier and mitigate the problem of overfitting the training data, we
generalize our approach toward $k$-maxitive capacities, where $k$ plays the
role of a hyper-parameter of the learner. We present experimental studies, in
which we compare our method with competing approaches on several benchmark data
sets. | [
"cs.LG",
"stat.ML"
] |
Semantic segmentation is essentially important to biomedical image analysis.
Many recent works mainly focus on integrating the Fully Convolutional Network
(FCN) architecture with sophisticated convolution implementation and deep
supervision. In this paper, we propose to decompose the single segmentation
task into three subsequent sub-tasks, including (1) pixel-wise image
segmentation, (2) prediction of the class labels of the objects within the
image, and (3) classification of the scene the image belonging to. While these
three sub-tasks are trained to optimize their individual loss functions of
different perceptual levels, we propose to let them interact by the task-task
context ensemble. Moreover, we propose a novel sync-regularization to penalize
the deviation between the outputs of the pixel-wise segmentation and the class
prediction tasks. These effective regularizations help FCN utilize context
information comprehensively and attain accurate semantic segmentation, even
though the number of the images for training may be limited in many biomedical
applications. We have successfully applied our framework to three diverse 2D/3D
medical image datasets, including Robotic Scene Segmentation Challenge 18
(ROBOT18), Brain Tumor Segmentation Challenge 18 (BRATS18), and Retinal Fundus
Glaucoma Challenge (REFUGE18). We have achieved top-tier performance in all
three challenges. | [
"cs.CV",
"eess.IV"
] |
Temporal Sentence Grounding in Videos (TSGV), i.e., grounding a natural
language sentence which indicates complex human activities in a long and
untrimmed video sequence, has received unprecedented attentions over the last
few years. Although each newly proposed method plausibly can achieve better
performance than previous ones, current TSGV models still tend to capture the
moment annotation biases and fail to take full advantage of multi-modal inputs.
Even more incredibly, several extremely simple baselines without training can
also achieve state-of-the-art performance. In this paper, we take a closer look
at the existing evaluation protocols for TSGV, and find that both the
prevailing dataset splits and evaluation metrics are the devils to cause
unreliable benchmarking. To this end, we propose to re-organize two widely-used
TSGV benchmarks (ActivityNet Captions and Charades-STA). Specifically, we
deliberately make the ground-truth moment distribution different in the
training and test splits, i.e., out-of-distribution (OOD) testing. Meanwhile,
we introduce a new evaluation metric dR@n,IoU@m to calibrate the basic IoU
scores by penalizing on the bias-influenced moment predictions and alleviate
the inflating evaluations caused by the dataset annotation biases such as
overlong ground-truth moments. Under our new evaluation protocol, we conduct
extensive experiments and ablation studies on eight state-of-the-art TSGV
methods. All the results demonstrate that the re-organized dataset splits and
new metric can better monitor the progress in TSGV. Our reorganized datsets are
available at https://github.com/yytzsy/grounding_changing_distribution. | [
"cs.CV"
] |
Text-to-image multimodal tasks, generating/retrieving an image from a given
text description, are extremely challenging tasks since raw text descriptions
cover quite limited information in order to fully describe visually realistic
images. We propose a new visual contextual text representation for
text-to-image multimodal tasks, VICTR, which captures rich visual semantic
information of objects from the text input. First, we use the text description
as initial input and conduct dependency parsing to extract the syntactic
structure and analyse the semantic aspect, including object quantities, to
extract the scene graph. Then, we train the extracted objects, attributes, and
relations in the scene graph and the corresponding geometric relation
information using Graph Convolutional Networks, and it generates text
representation which integrates textual and visual semantic information. The
text representation is aggregated with word-level and sentence-level embedding
to generate both visual contextual word and sentence representation. For the
evaluation, we attached VICTR to the state-of-the-art models in text-to-image
generation.VICTR is easily added to existing models and improves across both
quantitative and qualitative aspects. | [
"cs.CV",
"cs.AI"
] |
Given high-dimensional time series data (e.g., sensor data), how can we
detect anomalous events, such as system faults and attacks? More challengingly,
how can we do this in a way that captures complex inter-sensor relationships,
and detects and explains anomalies which deviate from these relationships?
Recently, deep learning approaches have enabled improvements in anomaly
detection in high-dimensional datasets; however, existing methods do not
explicitly learn the structure of existing relationships between variables, or
use them to predict the expected behavior of time series. Our approach combines
a structure learning approach with graph neural networks, additionally using
attention weights to provide explainability for the detected anomalies.
Experiments on two real-world sensor datasets with ground truth anomalies show
that our method detects anomalies more accurately than baseline approaches,
accurately captures correlations between sensors, and allows users to deduce
the root cause of a detected anomaly. | [
"cs.LG",
"cs.AI"
] |
In deep learning, models typically reuse the same parameters for all inputs.
Mixture of Experts (MoE) defies this and instead selects different parameters
for each incoming example. The result is a sparsely-activated model -- with
outrageous numbers of parameters -- but a constant computational cost. However,
despite several notable successes of MoE, widespread adoption has been hindered
by complexity, communication costs and training instability -- we address these
with the Switch Transformer. We simplify the MoE routing algorithm and design
intuitive improved models with reduced communication and computational costs.
Our proposed training techniques help wrangle the instabilities and we show
large sparse models may be trained, for the first time, with lower precision
(bfloat16) formats. We design models based off T5-Base and T5-Large to obtain
up to 7x increases in pre-training speed with the same computational resources.
These improvements extend into multilingual settings where we measure gains
over the mT5-Base version across all 101 languages. Finally, we advance the
current scale of language models by pre-training up to trillion parameter
models on the "Colossal Clean Crawled Corpus" and achieve a 4x speedup over the
T5-XXL model. | [
"cs.LG",
"cs.AI"
] |
Time series forecasting and spatiotemporal kriging are the two most important
tasks in spatiotemporal data analysis. Recent research on graph neural networks
has made substantial progress in time series forecasting, while little
attention has been paid to the kriging problem -- recovering signals for
unsampled locations/sensors. Most existing scalable kriging methods (e.g.,
matrix/tensor completion) are transductive, and thus full retraining is
required when we have a new sensor to interpolate. In this paper, we develop an
Inductive Graph Neural Network Kriging (IGNNK) model to recover data for
unsampled sensors on a network/graph structure. To generalize the effect of
distance and reachability, we generate random subgraphs as samples and
reconstruct the corresponding adjacency matrix for each sample. By
reconstructing all signals on each sample subgraph, IGNNK can effectively learn
the spatial message passing mechanism. Empirical results on several real-world
spatiotemporal datasets demonstrate the effectiveness of our model. In
addition, we also find that the learned model can be successfully transferred
to the same type of kriging tasks on an unseen dataset. Our results show that:
1) GNN is an efficient and effective tool for spatial kriging; 2) inductive
GNNs can be trained using dynamic adjacency matrices; 3) a trained model can be
transferred to new graph structures and 4) IGNNK can be used to generate
virtual sensors. | [
"cs.LG",
"stat.ML"
] |
Confocal laser endomicroscopy (CLE) allow on-the-fly in vivo intraoperative
imaging in a discreet field of view, especially for brain tumors, rather than
extracting tissue for examination ex vivo with conventional light microscopy.
Fluorescein sodium-driven CLE imaging is more interactive, rapid, and portable
than conventional hematoxylin and eosin (H&E)-staining. However, it has several
limitations: CLE images may be contaminated with artifacts (motion, red blood
cells, noise), and neuropathologists are mainly trained on colorful stained
histology slides like H&E while the CLE images are gray. To improve the
diagnostic quality of CLE, we used a micrograph of an H&E slide from a glioma
tumor biopsy and image style transfer, a neural network method for integrating
the content and style of two images. This was done through minimizing the
deviation of the target image from both the content (CLE) and style (H&E)
images. The style transferred images were assessed and compared to conventional
H&E histology by neurosurgeons and a neuropathologist who then validated the
quality enhancement in 100 pairs of original and transformed images. Average
reviewers' score on test images showed 84 out of 100 transformed images had
fewer artifacts and more noticeable critical structures compared to their
original CLE form. By providing images that are more interpretable than the
original CLE images and more rapidly acquired than H&E slides, the style
transfer method allows a real-time, cellular-level tissue examination using CLE
technology that closely resembles the conventional appearance of H&E staining
and may yield better diagnostic recognition than original CLE grayscale images. | [
"cs.CV",
"eess.IV"
] |
In the last few years, Deep Convolutional Neural Networks (D-CNNs) have shown
state-of-the-art (SOTA) performance for Visual Place Recognition (VPR), a
pivotal component of long-term intelligent robotic vision (vision-aware
localization and navigation systems). The prestigious generalization power of
D-CNNs gained upon training on large scale places datasets and learned
persistent image regions which are found to be robust for specific place
recognition under changing conditions and camera viewpoints. However, against
the computation and power intensive D-CNNs based VPR algorithms that are
employed to determine the approximate location of resource-constrained mobile
robots, lightweight VPR techniques are preferred. This paper presents a
computation- and energy-efficient CAMAL framework that captures place-specific
multi-layer convolutional attentions efficient for environment invariant-VPR.
At 4x lesser power consumption, evaluating the proposed VPR framework on
challenging benchmark place recognition datasets reveal better and comparable
Area under Precision-Recall (AUC-PR) curves with approximately 4x improved
image retrieval performance over the contemporary VPR methodologies. | [
"cs.CV"
] |
Kernel regression is an essential and ubiquitous tool for non-parametric data
analysis, particularly popular among time series and spatial data. However, the
central operation which is performed many times, evaluating a kernel on the
data set, takes linear time. This is impractical for modern large data sets.
In this paper we describe coresets for kernel regression: compressed data
sets which can be used as proxy for the original data and have provably bounded
worst case error. The size of the coresets are independent of the raw number of
data points, rather they only depend on the error guarantee, and in some cases
the size of domain and amount of smoothing. We evaluate our methods on very
large time series and spatial data, and demonstrate that they incur negligible
error, can be constructed extremely efficiently, and allow for great
computational gains. | [
"cs.LG",
"cs.DS"
] |
We introduce dense relational captioning, a novel image captioning task which
aims to generate multiple captions with respect to relational information
between objects in a visual scene. Relational captioning provides explicit
descriptions of each relationship between object combinations. This framework
is advantageous in both diversity and amount of information, leading to a
comprehensive image understanding based on relationships, e.g., relational
proposal generation. For relational understanding between objects, the
part-of-speech (POS, i.e., subject-object-predicate categories) can be a
valuable prior information to guide the causal sequence of words in a caption.
We enforce our framework to not only learn to generate captions but also
predict the POS of each word. To this end, we propose the multi-task
triple-stream network (MTTSNet) which consists of three recurrent units
responsible for each POS which is trained by jointly predicting the correct
captions and POS for each word. In addition, we found that the performance of
MTTSNet can be improved by modulating the object embeddings with an explicit
relational module. We demonstrate that our proposed model can generate more
diverse and richer captions, via extensive experimental analysis on large scale
datasets and several metrics. We additionally extend analysis to an ablation
study, applications on holistic image captioning, scene graph generation, and
retrieval tasks. | [
"cs.CV",
"cs.AI",
"cs.CL"
] |
We propose a method to train generative adversarial networks on mutivariate
feature vectors representing multiple categorical values. In contrast to the
continuous domain, where GAN-based methods have delivered considerable results,
GANs struggle to perform equally well on discrete data. We propose and compare
several architectures based on multiple (Gumbel) softmax output layers taking
into account the structure of the data. We evaluate the performance of our
architecture on datasets with different sparsity, number of features, ranges of
categorical values, and dependencies among the features. Our proposed
architecture and method outperforms existing models. | [
"stat.ML",
"cs.LG"
] |
Previous work showed empirically that large neural networks can be
significantly reduced in size while preserving their accuracy. Model
compression became a central research topic, as it is crucial for deployment of
neural networks on devices with limited computational and memory resources. The
majority of the compression methods are based on heuristics and offer no
worst-case guarantees on the trade-off between the compression rate and the
approximation error for an arbitrarily new sample. We propose the first
efficient, data-independent neural pruning algorithm with a provable trade-off
between its compression rate and the approximation error for any future test
sample. Our method is based on the coreset framework, which finds a small
weighted subset of points that provably approximates the original inputs.
Specifically, we approximate the output of a layer of neurons by a coreset of
neurons in the previous layer and discard the rest. We apply this framework in
a layer-by-layer fashion from the top to the bottom. Unlike previous works, our
coreset is data independent, meaning that it provably guarantees the accuracy
of the function for any input $x\in \mathbb{R}^d$, including an adversarial
one. We demonstrate the effectiveness of our method on popular network
architectures. In particular, our coresets yield 90\% compression of the
LeNet-300-100 architecture on MNIST while improving the accuracy. | [
"cs.LG",
"stat.ML"
] |
Recently there has been an increased interest in unsupervised learning of
disentangled representations using the Variational Autoencoder (VAE) framework.
Most of the existing work has focused largely on modifying the variational cost
function to achieve this goal. We first show that these modifications, e.g.
beta-VAE, simplify the tendency of variational inference to underfit causing
pathological over-pruning and over-orthogonalization of learned components.
Second we propose a complementary approach: to modify the probabilistic model
with a structured latent prior. This prior allows to discover latent variable
representations that are structured into a hierarchy of independent vector
spaces. The proposed prior has three major advantages: First, in contrast to
the standard VAE normal prior the proposed prior is not rotationally invariant.
This resolves the problem of unidentifiability of the standard VAE normal
prior. Second, we demonstrate that the proposed prior encourages a disentangled
latent representation which facilitates learning of disentangled
representations. Third, extensive quantitative experiments demonstrate that the
prior significantly mitigates the trade-off between reconstruction loss and
disentanglement over the state of the art. | [
"stat.ML",
"cs.LG"
] |
In this paper, Object Detection and Tracking System (ODTS) in combination
with a well-known deep learning network, Faster Regional Convolution Neural
Network (Faster R-CNN), for Object Detection and Conventional Object Tracking
algorithm will be introduced and applied for automatic detection and monitoring
of unexpected events on CCTVs in tunnels, which are likely to (1) Wrong-Way
Driving (WWD), (2) Stop, (3) Person out of vehicle in tunnel (4) Fire. ODTS
accepts a video frame in time as an input to obtain Bounding Box (BBox) results
by Object Detection and compares the BBoxs of the current and previous video
frames to assign a unique ID number to each moving and detected object. This
system makes it possible to track a moving object in time, which is not usual
to be achieved in conventional object detection frameworks. A deep learning
model in ODTS was trained with a dataset of event images in tunnels to Average
Precision (AP) values of 0.8479, 0.7161 and 0.9085 for target objects: Car,
Person, and Fire, respectively. Then, based on trained deep learning model, the
ODTS based Tunnel CCTV Accident Detection System was tested using four accident
videos which including each accident. As a result, the system can detect all
accidents within 10 seconds. The more important point is that the detection
capacity of ODTS could be enhanced automatically without any changes in the
program codes as the training dataset becomes rich. | [
"cs.CV",
"cs.LG"
] |
Superpixel-based methodologies have become increasingly popular in computer
vision, especially when the computation is too expensive in time or memory to
perform with a large number of pixels or features. However, rarely is
superpixel segmentation examined within the context of deep convolutional
neural network architectures. This paper presents a novel neural architecture
that exploits the superpixel feature space. The visual feature space is
organized using superpixels to provide the neural network with a substructure
of the images. As the superpixels associate the visual feature space with parts
of the objects in an image, the visual feature space is transformed into a
structured vector representation per superpixel. It is shown that it is
feasible to learn superpixel features using capsules and it is potentially
beneficial to perform image analysis in such a structured manner. This novel
deep learning architecture is examined in the context of an image
classification task, highlighting explicit interpretability (explainability) of
the network's decision making. The results are compared against a baseline deep
neural model, as well as among superpixel capsule networks with a variety of
hyperparameter settings. | [
"cs.CV"
] |
Necessity and sufficiency are the building blocks of all successful
explanations. Yet despite their importance, these notions have been
conceptually underdeveloped and inconsistently applied in explainable
artificial intelligence (XAI), a fast-growing research area that is so far
lacking in firm theoretical foundations. Building on work in logic,
probability, and causality, we establish the central role of necessity and
sufficiency in XAI, unifying seemingly disparate methods in a single formal
framework. We provide a sound and complete algorithm for computing explanatory
factors with respect to a given context, and demonstrate its flexibility and
competitive performance against state of the art alternatives on various tasks. | [
"cs.LG",
"cs.AI"
] |
Since the person re-identification task often suffers from the problem of
pose changes and occlusions, some attentive local features are often suppressed
when training CNNs. In this paper, we propose the Batch DropBlock (BDB) Network
which is a two branch network composed of a conventional ResNet-50 as the
global branch and a feature dropping branch. The global branch encodes the
global salient representations. Meanwhile, the feature dropping branch consists
of an attentive feature learning module called Batch DropBlock, which randomly
drops the same region of all input feature maps in a batch to reinforce the
attentive feature learning of local regions. The network then concatenates
features from both branches and provides a more comprehensive and spatially
distributed feature representation. Albeit simple, our method achieves
state-of-the-art on person re-identification and it is also applicable to
general metric learning tasks. For instance, we achieve 76.4% Rank-1 accuracy
on the CUHK03-Detect dataset and 83.0% Recall-1 score on the Stanford Online
Products dataset, outperforming the existing works by a large margin (more than
6%). | [
"cs.CV"
] |
Prior works on formalizing explanations of a graph neural network (GNN) focus
on a single use case - to preserve the prediction results through identifying
important edges and nodes. In this paper, we develop a multi-purpose
interpretation framework by acquiring a mask that indicates topology
perturbations of the input graphs. We pack the framework into an interactive
visualization system (GNNViz) which can fulfill multiple purposes:
Preserve,Promote, or Attack GNN's predictions. We illustrate our approach's
novelty and effectiveness with three case studies: First, GNNViz can assist non
expert users to easily explore the relationship between graph topology and
GNN's decision (Preserve), or to manipulate the prediction (Promote or Attack)
for an image classification task on MS-COCO; Second, on the Pokec social
network dataset, our framework can uncover unfairness and demographic biases;
Lastly, it compares with state-of-the-art GNN explainer baseline on a synthetic
dataset. | [
"cs.LG"
] |
This work addresses the problem of semantic foggy scene understanding (SFSU).
Although extensive research has been performed on image dehazing and on
semantic scene understanding with clear-weather images, little attention has
been paid to SFSU. Due to the difficulty of collecting and annotating foggy
images, we choose to generate synthetic fog on real images that depict
clear-weather outdoor scenes, and then leverage these partially synthetic data
for SFSU by employing state-of-the-art convolutional neural networks (CNN). In
particular, a complete pipeline to add synthetic fog to real, clear-weather
images using incomplete depth information is developed. We apply our fog
synthesis on the Cityscapes dataset and generate Foggy Cityscapes with 20550
images. SFSU is tackled in two ways: 1) with typical supervised learning, and
2) with a novel type of semi-supervised learning, which combines 1) with an
unsupervised supervision transfer from clear-weather images to their synthetic
foggy counterparts. In addition, we carefully study the usefulness of image
dehazing for SFSU. For evaluation, we present Foggy Driving, a dataset with 101
real-world images depicting foggy driving scenes, which come with ground truth
annotations for semantic segmentation and object detection. Extensive
experiments show that 1) supervised learning with our synthetic data
significantly improves the performance of state-of-the-art CNN for SFSU on
Foggy Driving; 2) our semi-supervised learning strategy further improves
performance; and 3) image dehazing marginally advances SFSU with our learning
strategy. The datasets, models and code are made publicly available. | [
"cs.CV"
] |
The successful deployment of artificial intelligence (AI) in many domains
from healthcare to hiring requires their responsible use, particularly in model
explanations and privacy. Explainable artificial intelligence (XAI) provides
more information to help users to understand model decisions, yet this
additional knowledge exposes additional risks for privacy attacks. Hence,
providing explanation harms privacy. We study this risk for image-based model
inversion attacks and identified several attack architectures with increasing
performance to reconstruct private image data from model explanations. We have
developed several multi-modal transposed CNN architectures that achieve
significantly higher inversion performance than using the target model
prediction only. These XAI-aware inversion models were designed to exploit the
spatial knowledge in image explanations. To understand which explanations have
higher privacy risk, we analyzed how various explanation types and factors
influence inversion performance. In spite of some models not providing
explanations, we further demonstrate increased inversion performance even for
non-explainable target models by exploiting explanations of surrogate models
through attention transfer. This method first inverts an explanation from the
target prediction, then reconstructs the target image. These threats highlight
the urgent and significant privacy risks of explanations and calls attention
for new privacy preservation techniques that balance the dual-requirement for
AI explainability and privacy. | [
"cs.CV",
"cs.CY",
"cs.LG"
] |
We present FedScale, a diverse set of challenging and realistic benchmark
datasets to facilitate scalable, comprehensive, and reproducible federated
learning (FL) research. FedScale datasets are large-scale, encompassing a
diverse range of important FL tasks, such as image classification, object
detection, language modeling, speech recognition, and reinforcement learning.
For each dataset, we provide a unified evaluation protocol using realistic data
splits and evaluation metrics. To meet the pressing need for reproducing
realistic FL at scale, we have also built an efficient evaluation platform to
simplify and standardize the process of FL experimental setup and model
evaluation. Our evaluation platform provides flexible APIs to implement new FL
algorithms and includes new execution backends with minimal developer efforts.
Finally, we perform indepth benchmark experiments on these datasets. Our
experiments suggest fruitful opportunities in heterogeneity-aware
co-optimizations of the system and statistical efficiency under realistic FL
characteristics. FedScale is open-source with permissive licenses and actively
maintained,1 and we welcome feedback and contributions from the community. | [
"cs.LG",
"cs.AI",
"cs.DC",
"cs.PF"
] |
This paper presents the performance of different blockbased discrete cosine
transform (DCT) algorithms for compressing color image. In this RGB component
of color image are converted to YCbCr before DCT transform is applied. Y is
luminance component;Cb and Cr are chrominance components of the image. The
modification of the image data is done based on the classification of image
blocks to edge blocks and non-edge blocks, then the edge block of the image is
compressed with low compression and the nonedge blocks is compressed with high
compression. The analysis results have indicated that the performance of the
suggested method is much better, where the constructed images are less
distorted and compressed with higher factor. | [
"cs.CV"
] |
Dealing with structured data needs the use of expressive representation
formalisms that, however, puts the problem to deal with the computational
complexity of the machine learning process. Furthermore, real world domains
require tools able to manage their typical uncertainty. Many statistical
relational learning approaches try to deal with these problems by combining the
construction of relevant relational features with a probabilistic tool. When
the combination is static (static propositionalization), the constructed
features are considered as boolean features and used offline as input to a
statistical learner; while, when the combination is dynamic (dynamic
propositionalization), the feature construction and probabilistic tool are
combined into a single process. In this paper we propose a selective
propositionalization method that search the optimal set of relational features
to be used by a probabilistic learner in order to minimize a loss function. The
new propositionalization approach has been combined with the random subspace
ensemble method. Experiments on real-world datasets shows the validity of the
proposed method. | [
"cs.LG",
"cs.AI"
] |
Graph self-supervised learning has gained increasing attention due to its
capacity to learn expressive node representations. Many pretext tasks, or loss
functions have been designed from distinct perspectives. However, we observe
that different pretext tasks affect downstream tasks differently cross
datasets, which suggests that searching pretext tasks is crucial for graph
self-supervised learning. Different from existing works focusing on designing
single pretext tasks, this work aims to investigate how to automatically
leverage multiple pretext tasks effectively. Nevertheless, evaluating
representations derived from multiple pretext tasks without direct access to
ground truth labels makes this problem challenging. To address this obstacle,
we make use of a key principle of many real-world graphs, i.e., homophily, or
the principle that ``like attracts like,'' as the guidance to effectively
search various self-supervised pretext tasks. We provide theoretical
understanding and empirical evidence to justify the flexibility of homophily in
this search task. Then we propose the AutoSSL framework which can automatically
search over combinations of various self-supervised tasks. By evaluating the
framework on 7 real-world datasets, our experimental results show that AutoSSL
can significantly boost the performance on downstream tasks including node
clustering and node classification compared with training under individual
tasks. Code will be released at https://github.com/ChandlerBang/AutoSSL. | [
"cs.LG",
"cs.AI"
] |
In this paper, we propose one novel model for point cloud semantic
segmentation, which exploits both the local and global structures within the
point cloud based on the contextual point representations. Specifically, we
enrich each point representation by performing one novel gated fusion on the
point itself and its contextual points. Afterwards, based on the enriched
representation, we propose one novel graph pointnet module, relying on the
graph attention block to dynamically compose and update each point
representation within the local point cloud structure. Finally, we resort to
the spatial-wise and channel-wise attention strategies to exploit the point
cloud global structure and thereby yield the resulting semantic label for each
point. Extensive results on the public point cloud databases, namely the S3DIS
and ScanNet datasets, demonstrate the effectiveness of our proposed model,
outperforming the state-of-the-art approaches. Our code for this paper is
available at https://github.com/fly519/ELGS. | [
"cs.CV"
] |
Among the wide variety of image generative models, two models stand out:
Variational Auto Encoders (VAE) and Generative Adversarial Networks (GAN). GANs
can produce realistic images, but they suffer from mode collapse and do not
provide simple ways to get the latent representation of an image. On the other
hand, VAEs do not have these problems, but they often generate images less
realistic than GANs. In this article, we explain that this lack of realism is
partially due to a common underestimation of the natural image manifold
dimensionality. To solve this issue we introduce a new framework that combines
VAE and GAN in a novel and complementary way to produce an auto-encoding model
that keeps VAEs properties while generating images of GAN-quality. We evaluate
our approach both qualitatively and quantitatively on five image datasets. | [
"cs.CV"
] |
Combining the properties of monovariate internal functions as proposed in
Kolmogorov superimposition theorem, in tandem with the bounds wielded by the
multivariate formulation of Chebyshev inequality, a hybrid model is presented,
that decomposes images into homogeneous probabilistically bounded multivariate
surfaces. Given an image, the model shows a novel way of working on reduced
image representation while processing and capturing the interaction among the
multidimensional information that describes the content of the same. Further,
it tackles the practical issues of preventing leakage by bounding the growth of
surface and reducing the problem sample size. The model if used, also sheds
light on how the Chebyshev parameter relates to the number of pixels and the
dimensionality of the feature space that associates with a pixel. Initial
segmentation results on the Berkeley image segmentation benchmark indicate the
effectiveness of the proposed decomposition algorithm. | [
"cs.CV"
] |
Stereo image pairs can be used to improve the performance of super-resolution
(SR) since additional information is provided from a second viewpoint. However,
it is challenging to incorporate this information for SR since disparities
between stereo images vary significantly. In this paper, we propose a
parallax-attention stereo superresolution network (PASSRnet) to integrate the
information from a stereo image pair for SR. Specifically, we introduce a
parallax-attention mechanism with a global receptive field along the epipolar
line to handle different stereo images with large disparity variations. We also
propose a new and the largest dataset for stereo image SR (namely, Flickr1024).
Extensive experiments demonstrate that the parallax-attention mechanism can
capture correspondence between stereo images to improve SR performance with a
small computational and memory cost. Comparative results show that our PASSRnet
achieves the state-of-the-art performance on the Middlebury, KITTI 2012 and
KITTI 2015 datasets. | [
"cs.CV"
] |
This paper develops a novel two-layer hierarchical classifier that increases
the accuracy of traditional transportation mode classification algorithms. This
paper also enhances classification accuracy by extracting new frequency domain
features. Many researchers have obtained these features from global positioning
system data; however, this data was excluded in this paper, as the system use
might deplete the smartphone's battery and signals may be lost in some areas.
Our proposed two-layer framework differs from previous classification attempts
in three distinct ways: 1) the outputs of the two layers are combined using
Bayes' rule to choose the transportation mode with the largest posterior
probability; 2) the proposed framework combines the new extracted features with
traditionally used time domain features to create a pool of features; and 3) a
different subset of extracted features is used in each layer based on the
classified modes. Several machine learning techniques were used, including
k-nearest neighbor, classification and regression tree, support vector machine,
random forest, and a heterogeneous framework of random forest and support
vector machine. Results show that the classification accuracy of the proposed
framework outperforms traditional approaches. Transforming the time domain
features to the frequency domain also adds new features in a new space and
provides more control on the loss of information. Consequently, combining the
time domain and the frequency domain features in a large pool and then choosing
the best subset results in higher accuracy than using either domain alone. The
proposed two-layer classifier obtained a maximum classification accuracy of
97.02%. | [
"cs.LG",
"cs.CY",
"physics.soc-ph",
"stat.ML"
] |
In this paper we put the visibility transformation on a clear theoretical
footing and show that this transform is able to embed the effect of the
absolute position of the data stream into signature features in a unified and
efficient way. The generated feature set is particularly useful in pattern
recognition tasks, for its simplifying role in allowing the signature feature
set to accommodate nonlinear functions of absolute and relative values. | [
"cs.LG",
"eess.SP",
"stat.ML",
"60L10"
] |
Learning good interventions in a causal graph can be modelled as a stochastic
multi-armed bandit problem with side-information. First, we study this problem
when interventions are more expensive than observations and a budget is
specified. If there are no backdoor paths from an intervenable node to the
reward node then we propose an algorithm to minimize simple regret that
optimally trades-off observations and interventions based on the cost of
intervention. We also propose an algorithm that accounts for the cost of
interventions, utilizes causal side-information, and minimizes the expected
cumulative regret without exceeding the budget. Our cumulative-regret
minimization algorithm performs better than standard algorithms that do not
take side-information into account. Finally, we study the problem of learning
best interventions without budget constraint in general graphs and give an
algorithm that achieves constant expected cumulative regret in terms of the
instance parameters when the parent distribution of the reward variable for
each intervention is known. Our results are experimentally validated and
compared to the best-known bounds in the current literature. | [
"cs.LG",
"stat.ML"
] |
We present a `CLAssifier-DECoder' architecture (\emph{ClaDec}) which
facilitates the comprehension of the output of an arbitrary layer in a neural
network (NN). It uses a decoder to transform the non-interpretable
representation of the given layer to a representation that is more similar to
the domain a human is familiar with. In an image recognition problem, one can
recognize what information is represented by a layer by contrasting
reconstructed images of \emph{ClaDec} with those of a conventional
auto-encoder(AE) serving as reference. We also extend \emph{ClaDec} to allow
the trade-off between human interpretability and fidelity. We evaluate our
approach for image classification using Convolutional NNs. We show that
reconstructed visualizations using encodings from a classifier capture more
relevant information for classification than conventional AEs. Relevant code is
available at \url{https://github.com/JohnTailor/ClaDec} | [
"cs.LG",
"stat.ML"
] |
In the paper the optimal image segmentation by means of piecewise constant
approximations is considered. The optimality is defined by a minimum value of
the total squared error or by equivalent value of standard deviation of the
approximation from the image. The optimal approximations are defined
independently on the method of their obtaining and might be generated in
different algorithms. We investigate the computation of the optimal
approximation on the grounds of stability with respect to a given set of
modifications. To obtain the optimal approximation the Mumford-Shuh model is
generalized and developed, which in the computational part is combined with the
Otsu method in multi-thresholding version. The proposed solution is proved
analytically and experimentally on the example of the standard image. | [
"cs.CV"
] |
As a common visual problem, co-saliency detection within a single image does
not attract enough attention and yet has not been well addressed. Existing
methods often follow a bottom-up strategy to infer co-saliency in an image,
where salient regions are firstly detected using visual primitives such as
color and shape, and then grouped and merged into a co-saliency map. However,
co-saliency is intrinsically perceived in a complex manner with bottom-up and
top-down strategies combined in human vision. To deal with this problem, a
novel end-to-end trainable network is proposed in this paper, which includes a
backbone net and two branch nets. The backbone net uses ground-truth masks as
top-down guidance for saliency prediction, while the two branch nets construct
triplet proposals for feature organization and clustering, which drives the
network to be sensitive to co-salient regions in a bottom-up way. To evaluate
the proposed method, we construct a new dataset of 2,019 nature images with
co-saliency in each image. Experimental results show that the proposed method
achieves a state-of-the-art accuracy with a running speed of 28fps. | [
"cs.CV"
] |
We propose a new method for the numerical solution of a PDE-driven model for
colour image segmentation and give numerical examples of the results. The
method combines the vector-valued Allen-Cahn phase field equation with initial
data fitting terms. This method is known to be closely related to the
Mumford-Shah problem and the level set segmentation by Chan and Vese. Our
numerical solution is performed using a multigrid splitting of a finite element
space, thereby producing an efficient and robust method for the segmentation of
large images. | [
"cs.CV",
"cs.NA",
"I.4.6; G.1.8"
] |
We present a fast and scalable algorithm to induce non-monotonic logic
programs from statistical learning models. We reduce the problem of search for
best clauses to instances of the High-Utility Itemset Mining (HUIM) problem. In
the HUIM problem, feature values and their importance are treated as
transactions and utilities respectively. We make use of TreeExplainer, a fast
and scalable implementation of the Explainable AI tool SHAP, to extract locally
important features and their weights from ensemble tree models. Our experiments
with UCI standard benchmarks suggest a significant improvement in terms of
classification evaluation metrics and running time of the training algorithm
compared to ALEPH, a state-of-the-art Inductive Logic Programming (ILP) system. | [
"cs.LG",
"cs.LO",
"stat.ML"
] |
The automated recognition of music genres from audio information is a
challenging problem, as genre labels are subjective and noisy. Artist labels
are less subjective and less noisy, while certain artists may relate more
strongly to certain genres. At the same time, at prediction time, it is not
guaranteed that artist labels are available for a given audio segment.
Therefore, in this work, we propose to apply the transfer learning framework,
learning artist-related information which will be used at inference time for
genre classification. We consider different types of artist-related
information, expressed through artist group factors, which will allow for more
efficient learning and stronger robustness to potential label noise.
Furthermore, we investigate how to achieve the highest validation accuracy on
the given FMA dataset, by experimenting with various kinds of transfer methods,
including single-task transfer, multi-task transfer and finally multi-task
learning. | [
"cs.LG",
"cs.SD",
"eess.AS",
"stat.ML"
] |
Echo State Networks (ESNs) are a class of single-layer recurrent neural
networks with randomly generated internal weights, and a single layer of
tuneable outer weights, which are usually trained by regularised linear least
squares regression. Remarkably, ESNs still enjoy the universal approximation
property despite the training procedure being entirely linear. In this paper,
we prove that an ESN trained on a sequence of observations from an ergodic
dynamical system (with invariant measure $\mu$) using Tikhonov least squares
regression against a set of targets, will approximate the target function in
the $L^2(\mu)$ norm. In the special case that the targets are future
observations, the ESN is learning the next step map, which allows time series
forecasting. We demonstrate the theory numerically by training an ESN using
Tikhonov least squares on a sequence of scalar observations of the Lorenz
system. | [
"cs.LG",
"math.DS",
"stat.ML"
] |
Distributional reinforcement learning (distributional RL) has seen empirical
success in complex Markov Decision Processes (MDPs) in the setting of nonlinear
function approximation. However, there are many different ways in which one can
leverage the distributional approach to reinforcement learning. In this paper,
we propose GAN Q-learning, a novel distributional RL method based on generative
adversarial networks (GANs) and analyze its performance in simple tabular
environments, as well as OpenAI Gym. We empirically show that our algorithm
leverages the flexibility and blackbox approach of deep learning models while
providing a viable alternative to traditional methods. | [
"stat.ML",
"cs.LG"
] |
Although recent inpainting approaches have demonstrated significant
improvements with deep neural networks, they still suffer from artifacts such
as blunt structures and abrupt colors when filling in the missing regions. To
address these issues, we propose an external-internal inpainting scheme with a
monochromic bottleneck that helps image inpainting models remove these
artifacts. In the external learning stage, we reconstruct missing structures
and details in the monochromic space to reduce the learning dimension. In the
internal learning stage, we propose a novel internal color propagation method
with progressive learning strategies for consistent color restoration.
Extensive experiments demonstrate that our proposed scheme helps image
inpainting models produce more structure-preserved and visually compelling
results. | [
"cs.CV"
] |
We aim to mine temporal causal sequences that explain observed events
(consequents) in time-series traces. Causal explanations of key events in a
time-series has applications in design debugging, anomaly detection, planning,
root-cause analysis and many more. We make use of decision trees and interval
arithmetic to mine sequences that explain defining events in the time-series.
We propose modified decision tree construction metrics to handle the
non-determinism introduced by the temporal dimension. The mined sequences are
expressed in a readable temporal logic language that is easy to interpret. The
application of the proposed methodology is illustrated through various
examples. | [
"cs.LG",
"cs.AI",
"cs.LO"
] |
Weather Recognition plays an important role in our daily lives and many
computer vision applications. However, recognizing the weather conditions from
a single image remains challenging and has not been studied thoroughly.
Generally, most previous works treat weather recognition as a single-label
classification task, namely, determining whether an image belongs to a specific
weather class or not. This treatment is not always appropriate, since more than
one weather conditions may appear simultaneously in a single image. To address
this problem, we make the first attempt to view weather recognition as a
multi-label classification task, i.e., assigning an image more than one labels
according to the displayed weather conditions. Specifically, a CNN-RNN based
multi-label classification approach is proposed in this paper. The
convolutional neural network (CNN) is extended with a channel-wise attention
model to extract the most correlated visual features. The Recurrent Neural
Network (RNN) further processes the features and excavates the dependencies
among weather classes. Finally, the weather labels are predicted step by step.
Besides, we construct two datasets for the weather recognition task and explore
the relationships among different weather conditions. Experimental results
demonstrate the superiority and effectiveness of the proposed approach. The new
constructed datasets will be available at
https://github.com/wzgwzg/Multi-Label-Weather-Recognition. | [
"cs.CV",
"cs.AI"
] |
Autonomous systems possess the features of inferring their own ego-motion,
autonomously understanding their surroundings, and planning trajectories. With
the applications of deep learning and reinforcement learning, the perception
and decision-making abilities of autonomous systems are being efficiently
addressed, and many new learning-based algorithms have surfaced with respect to
autonomous perception and decision-making. In this review, we focus on the
applications of learning-based approaches in perception and decision-making in
autonomous systems, which is different from previous reviews that discussed
traditional methods. First, we delineate the existing classical simultaneous
localization and mapping (SLAM) solutions and review the environmental
perception and understanding methods based on deep learning, including deep
learning-based monocular depth estimation, ego-motion prediction, image
enhancement, object detection, semantic segmentation, and their combinations
with traditional SLAM frameworks. Second, we briefly summarize the existing
motion planning techniques, such as path planning and trajectory planning
methods, and discuss the navigation methods based on reinforcement learning.
Finally, we examine the several challenges and promising directions discussed
and concluded in related research for future works in the era of computer
science, automatic control, and robotics. | [
"cs.CV"
] |
The research project HDV-Mess aims at a currently missing, but very crucial
component for addressing important challenges in the field of connected and
automated driving on public roads. The goal is to record traffic events at
various relevant locations with high accuracy and to collect real traffic data
as a basis for the development and validation of current and future sensor
technologies as well as automated driving functions. For this purpose, it is
necessary to develop a concept for a mobile modular system of measuring
stations for highly accurate traffic data acquisition, which enables a
temporary installation of a sensor and communication infrastructure at
different locations. Within this paper, we first discuss the project goals
before we present our traffic detection concept using mobile modular
intelligent transport systems stations (ITS-Ss). We then explain the approaches
for data processing of sensor raw data to refined trajectories, data
communication, and data validation. | [
"cs.CV"
] |
We propose and study a method called FLOT that estimates scene flow on point
clouds. We start the design of FLOT by noticing that scene flow estimation on
point clouds reduces to estimating a permutation matrix in a perfect world.
Inspired by recent works on graph matching, we build a method to find these
correspondences by borrowing tools from optimal transport. Then, we relax the
transport constraints to take into account real-world imperfections. The
transport cost between two points is given by the pairwise similarity between
deep features extracted by a neural network trained under full supervision
using synthetic datasets. Our main finding is that FLOT can perform as well as
the best existing methods on synthetic and real-world datasets while requiring
much less parameters and without using multiscale analysis. Our second finding
is that, on the training datasets considered, most of the performance can be
explained by the learned transport cost. This yields a simpler method,
FLOT$_0$, which is obtained using a particular choice of optimal transport
parameters and performs nearly as well as FLOT. | [
"cs.CV"
] |
While previous distribution shift detection approaches can identify if a
shift has occurred, these approaches cannot localize which specific features
have caused a distribution shift -- a critical step in diagnosing or fixing any
underlying issue. For example, in military sensor networks, users will want to
detect when one or more of the sensors has been compromised, and critically,
they will want to know which specific sensors might be compromised. Thus, we
first define a formalization of this problem as multiple conditional
distribution hypothesis tests and propose both non-parametric and parametric
statistical tests. For both efficiency and flexibility, we then propose to use
a test statistic based on the density model score function (i.e. gradient with
respect to the input) -- which can easily compute test statistics for all
dimensions in a single forward and backward pass. Any density model could be
used for computing the necessary statistics including deep density models such
as normalizing flows or autoregressive models. We additionally develop methods
for identifying when and where a shift occurs in multivariate time-series data
and show results for multiple scenarios using realistic attack models on both
simulated and real world data. | [
"cs.LG",
"stat.ML"
] |
Digital image forensics aims to detect images that have been digitally
manipulated. Realistic image forgeries involve a combination of splicing,
resampling, region removal, smoothing and other manipulation methods. While
most detection methods in literature focus on detecting a particular type of
manipulation, it is challenging to identify doctored images that involve a host
of manipulations. In this paper, we propose a novel approach to holistically
detect tampered images using a combination of pixel co-occurrence matrices and
deep learning. We extract horizontal and vertical co-occurrence matrices on
three color channels in the pixel domain and train a model using a deep
convolutional neural network (CNN) framework. Our method is agnostic to the
type of manipulation and classifies an image as tampered or untampered. We
train and validate our model on a dataset of more than 86,000 images.
Experimental results show that our approach is promising and achieves more than
0.99 area under the curve (AUC) evaluation metric on the training and
validation subsets. Further, our approach also generalizes well and achieves
around 0.81 AUC on an unseen test dataset comprising more than 19,740 images
released as part of the Media Forensics Challenge (MFC) 2020. Our score was
highest among all other teams that participated in the challenge, at the time
of announcement of the challenge results. | [
"cs.CV"
] |
Deep reinforcement learning techniques have demonstrated superior performance
in a wide variety of environments. As improvements in training algorithms
continue at a brisk pace, theoretical or empirical studies on understanding
what these networks seem to learn, are far behind. In this paper we propose an
interpretable neural network architecture for Q-learning which provides a
global explanation of the model's behavior using key-value memories, attention
and reconstructible embeddings. With a directed exploration strategy, our model
can reach training rewards comparable to the state-of-the-art deep Q-learning
models. However, results suggest that the features extracted by the neural
network are extremely shallow and subsequent testing using out-of-sample
examples shows that the agent can easily overfit to trajectories seen during
training. | [
"cs.LG",
"stat.ML"
] |
Spectral methods have been the mainstay in several domains such as machine
learning and scientific computing. They involve finding a certain kind of
spectral decomposition to obtain basis functions that can capture important
structures for the problem at hand. The most common spectral method is the
principal component analysis (PCA). It utilizes the top eigenvectors of the
data covariance matrix, e.g. to carry out dimensionality reduction. This data
pre-processing step is often effective in separating signal from noise. PCA and
other spectral techniques applied to matrices have several limitations. By
limiting to only pairwise moments, they are effectively making a Gaussian
approximation on the underlying data and fail on data with hidden variables
which lead to non-Gaussianity. However, in most data sets, there are latent
effects that cannot be directly observed, e.g., topics in a document corpus, or
underlying causes of a disease. By extending the spectral decomposition methods
to higher order moments, we demonstrate the ability to learn a wide range of
latent variable models efficiently. Higher-order moments can be represented by
tensors, and intuitively, they can encode more information than just pairwise
moment matrices. More crucially, tensor decomposition can pick up latent
effects that are missed by matrix methods, e.g. uniquely identify
non-orthogonal components. Exploiting these aspects turns out to be fruitful
for provable unsupervised learning of a wide range of latent variable models.
We also outline the computational techniques to design efficient tensor
decomposition methods. We introduce Tensorly, which has a simple python
interface for expressing tensor operations. It has a flexible back-end system
supporting NumPy, PyTorch, TensorFlow and MXNet amongst others, allowing
multi-GPU and CPU operations and seamless integration with deep-learning
functionalities. | [
"cs.LG",
"stat.ML"
] |
The latest advances in computer-assisted precision medicine are making it
feasible to move from population-wide models that are useful to discover
aggregate patterns that hold for group-based analysis to patient-specific
models that can drive patient-specific decisions with regard to treatment
choices, and predictions of outcomes of treatment. Body Composition is
recognized as an important driver and risk factor for a wide variety of
diseases, as well as a predictor of individual patient-specific clinical
outcomes to treatment choices or surgical interventions. 3D CT images are
routinely acquired in the oncological worklows and deliver accurate rendering
of internal anatomy and therefore can be used opportunistically to assess the
amount of skeletal muscle and adipose tissue compartments. Powerful tools of
artificial intelligence such as deep learning are making it feasible now to
segment the entire 3D image and generate accurate measurements of all internal
anatomy. These will enable the overcoming of the severe bottleneck that existed
previously, namely, the need for manual segmentation, which was prohibitive to
scale to the hundreds of 2D axial slices that made up a 3D volumetric image.
Automated tools such as presented here will now enable harvesting whole-body
measurements from 3D CT or MRI images, leading to a new era of discovery of the
drivers of various diseases based on individual tissue, organ volume, shape,
and functional status. These measurements were hitherto unavailable thereby
limiting the field to a very small and limited subset. These discoveries and
the potential to perform individual image segmentation with high speed and
accuracy are likely to lead to the incorporation of these 3D measures into
individual specific treatment planning models related to nutrition, aging,
chemotoxicity, surgery and survival after the onset of a major disease such as
cancer. | [
"cs.CV",
"q-bio.TO"
] |
The application of deep learning to medical image segmentation has been
hampered due to the lack of abundant pixel-level annotated data. Few-shot
Semantic Segmentation (FSS) is a promising strategy for breaking the deadlock.
However, a high-performing FSS model still requires sufficient pixel-level
annotated classes for training to avoid overfitting, which leads to its
performance bottleneck in medical image segmentation due to the unmet need for
annotations. Thus, semi-supervised FSS for medical images is accordingly
proposed to utilize unlabeled data for further performance improvement.
Nevertheless, existing semi-supervised FSS methods has two obvious defects: (1)
neglecting the relationship between the labeled and unlabeled data; (2) using
unlabeled data directly for end-to-end training leads to degenerated
representation learning. To address these problems, we propose a novel
semi-supervised FSS framework for medical image segmentation. The proposed
framework employs Poisson learning for modeling data relationship and
propagating supervision signals, and Spatial Consistency Calibration for
encouraging the model to learn more coherent representations. In this process,
unlabeled samples do not involve in end-to-end training, but provide
supervisory information for query image segmentation through graph-based
learning. We conduct extensive experiments on three medical image segmentation
datasets (i.e. ISIC skin lesion segmentation, abdominal organs segmentation for
MRI and abdominal organs segmentation for CT) to demonstrate the
state-of-the-art performance and broad applicability of the proposed framework. | [
"cs.CV",
"cs.LG"
] |
Although Faster R-CNN and its variants have shown promising performance in
object detection, they only exploit simple first-order representation of object
proposals for final classification and regression. Recent classification
methods demonstrate that the integration of high-order statistics into deep
convolutional neural networks can achieve impressive improvement, but their
goal is to model whole images by discarding location information so that they
cannot be directly adopted to object detection. In this paper, we make an
attempt to exploit high-order statistics in object detection, aiming at
generating more discriminative representations for proposals to enhance the
performance of detectors. To this end, we propose a novel Multi-scale
Location-aware Kernel Representation (MLKP) to capture high-order statistics of
deep features in proposals. Our MLKP can be efficiently computed on a modified
multi-scale feature map using a low-dimensional polynomial kernel
approximation.Moreover, different from existing orderless global
representations based on high-order statistics, our proposed MLKP is location
retentive and sensitive so that it can be flexibly adopted to object detection.
Through integrating into Faster R-CNN schema, the proposed MLKP achieves very
competitive performance with state-of-the-art methods, and improves Faster
R-CNN by 4.9% (mAP), 4.7% (mAP) and 5.0% (AP at IOU=[0.5:0.05:0.95]) on PASCAL
VOC 2007, VOC 2012 and MS COCO benchmarks, respectively. Code is available at:
https://github.com/Hwang64/MLKP. | [
"cs.CV"
] |
Neural networks in the real domain have been studied for a long time and
achieved promising results in many vision tasks for recent years. However, the
extensions of the neural network models in other number fields and their
potential applications are not fully-investigated yet. Focusing on color
images, which can be naturally represented as quaternion matrices, we propose a
quaternion convolutional neural network (QCNN) model to obtain more
representative features. In particular, we redesign the basic modules like
convolution layer and fully-connected layer in the quaternion domain, which can
be used to establish fully-quaternion convolutional neural networks. Moreover,
these modules are compatible with almost all deep learning techniques and can
be plugged into traditional CNNs easily. We test our QCNN models in both color
image classification and denoising tasks. Experimental results show that they
outperform the real-valued CNNs with same structures. | [
"cs.CV"
] |
The human-object interaction (HOI) detection task refers to localizing
humans, localizing objects, and predicting the interactions between each
human-object pair. HOI is considered one of the fundamental steps in truly
understanding complex visual scenes. For detecting HOI, it is important to
utilize relative spatial configurations and object semantics to find salient
spatial regions of images that highlight the interactions between human object
pairs. This issue is addressed by the proposed self-attention based guided
transformer network, GTNet. GTNet encodes this spatial contextual information
in human and object visual features via self-attention while achieving a 4%-6%
improvement over previous state of the art results on both the V-COCO and
HICO-DET datasets. Code will be made available online. | [
"cs.CV"
] |
Corrupting the input and hidden layers of deep neural networks (DNNs) with
multiplicative noise, often drawn from the Bernoulli distribution (or
'dropout'), provides regularization that has significantly contributed to deep
learning's success. However, understanding how multiplicative corruptions
prevent overfitting has been difficult due to the complexity of a DNN's
functional form. In this paper, we show that when a Gaussian prior is placed on
a DNN's weights, applying multiplicative noise induces a Gaussian scale
mixture, which can be reparameterized to circumvent the problematic likelihood
function. Analysis can then proceed by using a type-II maximum likelihood
procedure to derive a closed-form expression revealing how regularization
evolves as a function of the network's weights. Results show that
multiplicative noise forces weights to become either sparse or invariant to
rescaling. We find our analysis has implications for model compression as it
naturally reveals a weight pruning rule that starkly contrasts with the
commonly used signal-to-noise ratio (SNR). While the SNR prunes weights with
large variances, seeing them as noisy, our approach recognizes their robustness
and retains them. We empirically demonstrate our approach has a strong
advantage over the SNR heuristic and is competitive to retraining with soft
targets produced from a teacher model. | [
"stat.ML"
] |
In this paper, we investigate a novel problem of telling the difference
between image pairs in natural language. Compared to previous approaches for
single image captioning, it is challenging to fetch linguistic representation
from two independent visual information. To this end, we have proposed an
effective encoder-decoder caption framework based on Hyper Convolution Net. In
addition, a series of novel feature fusing techniques for pairwise visual
information fusing are introduced and a discriminating referee is proposed to
evaluate the pipeline. Because of the lack of appropriate datasets to support
this task, we have collected and annotated a large new dataset with Amazon
Mechanical Turk (AMT) for generating captions in a pairwise manner (with 14764
images and 26710 image pairs in total). The dataset is the first one on the
relative difference caption task that provides descriptions in free language.
We evaluate the effectiveness of our model on two datasets in the field and it
outperforms the state-of-the-art approach by a large margin. | [
"cs.CV",
"cs.CL"
] |
Recent studies have shown that reinforcement learning (RL) models are
vulnerable in various noisy scenarios. For instance, the observed reward
channel is often subject to noise in practice (e.g., when rewards are collected
through sensors), and is therefore not credible. In addition, for applications
such as robotics, a deep reinforcement learning (DRL) algorithm can be
manipulated to produce arbitrary errors by receiving corrupted rewards. In this
paper, we consider noisy RL problems with perturbed rewards, which can be
approximated with a confusion matrix. We develop a robust RL framework that
enables agents to learn in noisy environments where only perturbed rewards are
observed. Our solution framework builds on existing RL/DRL algorithms and
firstly addresses the biased noisy reward setting without any assumptions on
the true distribution (e.g., zero-mean Gaussian noise as made in previous
works). The core ideas of our solution include estimating a reward confusion
matrix and defining a set of unbiased surrogate rewards. We prove the
convergence and sample complexity of our approach. Extensive experiments on
different DRL platforms show that trained policies based on our estimated
surrogate reward can achieve higher expected rewards, and converge faster than
existing baselines. For instance, the state-of-the-art PPO algorithm is able to
obtain 84.6% and 80.8% improvements on average score for five Atari games, with
error rates as 10% and 30% respectively. | [
"cs.LG",
"cs.CR",
"cs.CV",
"stat.ML"
] |
Segmentation of both large and small white matter hyperintensities/lesions in
brain MR images is a challenging task which has drawn much attention in recent
years. We propose a multi-scale aggregation model framework to deal with
volume-varied lesions. Firstly, we present a specifically-designed network for
small lesion segmentation called Stack-Net, in which multiple convolutional
layers are connected, aiming to preserve rich local spatial information of
small lesions before the sub-sampling layer. Secondly, we aggregate multi-scale
Stack-Nets with different receptive fields to learn multi-scale contextual
information of both large and small lesions. Our model is evaluated on recent
MICCAI WMH Challenge Dataset and outperforms the state-of-the-art on lesion
recall and lesion F1-score under 5-fold cross validation. In addition, we
further test our pre-trained models on a Multiple Sclerosis lesion dataset with
30 subjects under cross-center evaluation. Results show that the aggregation
model is effective in learning multi-scale spatial information.It claimed the
first place on the hidden test set after independent evaluation by the
challenge organizer. In addition, we further test our pre-trained models on a
Multiple Sclerosis lesion dataset with 30 subjects under cross-center
evaluation. Results show that the aggregation model is effective in learning
multi-scale spatial information. | [
"cs.CV"
] |
Novel Object Captioning is a zero-shot Image Captioning task requiring
describing objects not seen in the training captions, but for which information
is available from external object detectors. The key challenge is to select and
describe all salient detected novel objects in the input images. In this paper,
we focus on this challenge and propose the ECOL-R model (Encouraging Copying of
Object Labels with Reinforced Learning), a copy-augmented transformer model
that is encouraged to accurately describe the novel object labels. This is
achieved via a specialised reward function in the SCST reinforcement learning
framework (Rennie et al., 2017) that encourages novel object mentions while
maintaining the caption quality. We further restrict the SCST training to the
images where detected objects are mentioned in reference captions to train the
ECOL-R model. We additionally improve our copy mechanism via Abstract Labels,
which transfer knowledge from known to novel object types, and a Morphological
Selector, which determines the appropriate inflected forms of novel object
labels. The resulting model sets new state-of-the-art on the nocaps (Agrawal et
al., 2019) and held-out COCO (Hendricks et al., 2016) benchmarks. | [
"cs.CV",
"cs.CL"
] |
Existing statistical approaches to natural language problems are very coarse
approximations to the true complexity of language processing. As such, no
single technique will be best for all problem instances. Many researchers are
examining ensemble methods that combine the output of multiple modules to
create more accurate solutions. This paper examines three merging rules for
combining probability distributions: the familiar mixture rule, the logarithmic
rule, and a novel product rule. These rules were applied with state-of-the-art
results to two problems used to assess human mastery of lexical semantics --
synonym questions and analogy questions. All three merging rules result in
ensembles that are more accurate than any of their component modules. The
differences among the three rules are not statistically significant, but it is
suggestive that the popular mixture rule is not the best rule for either of the
two problems. | [
"cs.LG",
"cs.CL",
"cs.IR",
"I.2.6; I.2.7; H.3.1; J.5"
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.