text
stringlengths 29
3.31k
| label
sequencelengths 1
11
|
|---|---|
We present a new computational model for gaze prediction in egocentric videos
by exploring patterns in temporal shift of gaze fixations (attention
transition) that are dependent on egocentric manipulation tasks. Our assumption
is that the high-level context of how a task is completed in a certain way has
a strong influence on attention transition and should be modeled for gaze
prediction in natural dynamic scenes. Specifically, we propose a hybrid model
based on deep neural networks which integrates task-dependent attention
transition with bottom-up saliency prediction. In particular, the
task-dependent attention transition is learned with a recurrent neural network
to exploit the temporal context of gaze fixations, e.g. looking at a cup after
moving gaze away from a grasped bottle. Experiments on public egocentric
activity datasets show that our model significantly outperforms
state-of-the-art gaze prediction methods and is able to learn meaningful
transition of human attention. | [
"cs.CV"
] |
An increasing amount of studies have investigated the decision-making process
of VQA models. Many of these studies focus on the reason behind the correct
answer chosen by a model. Yet, the reason why the distracting answer chose by a
model has rarely been studied. To this end, we introduce a novel task called
\textit{textual Distractors Generation for VQA} (DG-VQA) that explaining the
decision boundaries of existing VQA models. The goal of DG-VQA is to generate
the most confusing set of textual distractors in multi-choice VQA tasks which
expose the vulnerability of existing models (i.e. to generate distractors that
lure existing models to fail). We show that DG-VQA can be formulated as a
Markov Decision Process, and present a reinforcement learning solution to come
up with distractors in an unsupervised manner. The solution addresses the lack
of large annotated corpus issues in previous distractor generation methods. Our
proposed model receives reward signals from fully-trained multi-choice VQA
models and updates its parameters via policy gradient. The empirical results
show that the generated textual distractors can successfully attack several
popular VQA models with an average $20\%$ accuracy drop from $64\%$.
Furthermore, we conduct adversarial training to improve the robustness of VQA
models by incorporating the generated distractors. Empirical results validate
the effectiveness of adversarial training by showing a performance improvement
of $27\%$ for the multi-choice VQA task. | [
"cs.CV",
"cs.CL"
] |
One of the most exciting technology breakthroughs in the last few years has
been the rise of deep learning. State-of-the-art deep learning models are being
widely deployed in academia and industry, across a variety of areas, from image
analysis to natural language processing. These models have grown from fledgling
research subjects to mature techniques in real-world use. The increasing scale
of data, computational power and the associated algorithmic innovations are the
main drivers for the progress we see in this field. These developments also
have a huge potential for the automotive industry and therefore the interest in
deep learning-based technology is growing. A lot of the product innovations,
such as self-driving cars, parking and lane-change assist or safety functions,
such as autonomous emergency braking, are powered by deep learning algorithms.
Deep learning is poised to offer gains in performance and functionality for
most ADAS (Advanced Driver Assistance System) solutions. Virtual sensing for
vehicle dynamics application, vehicle inspection/heath monitoring, automated
driving and data-driven product development are key areas that are expected to
get the most attention. This article provides an overview of the recent
advances and some associated challenges in deep learning techniques in the
context of automotive applications. | [
"cs.LG",
"cs.RO",
"eess.SP",
"stat.ML"
] |
How quickly can a given class of concepts be learned from examples? It is
common to measure the performance of a supervised machine learning algorithm by
plotting its "learning curve", that is, the decay of the error rate as a
function of the number of training examples. However, the classical theoretical
framework for understanding learnability, the PAC model of Vapnik-Chervonenkis
and Valiant, does not explain the behavior of learning curves: the
distribution-free PAC model of learning can only bound the upper envelope of
the learning curves over all possible data distributions. This does not match
the practice of machine learning, where the data source is typically fixed in
any given scenario, while the learner may choose the number of training
examples on the basis of factors such as computational resources and desired
accuracy.
In this paper, we study an alternative learning model that better captures
such practical aspects of machine learning, but still gives rise to a complete
theory of the learnable in the spirit of the PAC model. More precisely, we
consider the problem of universal learning, which aims to understand the
performance of learning algorithms on every data distribution, but without
requiring uniformity over the distribution. The main result of this paper is a
remarkable trichotomy: there are only three possible rates of universal
learning. More precisely, we show that the learning curves of any given concept
class decay either at an exponential, linear, or arbitrarily slow rates.
Moreover, each of these cases is completely characterized by appropriate
combinatorial parameters, and we exhibit optimal learning algorithms that
achieve the best possible rate in each case.
For concreteness, we consider in this paper only the realizable case, though
analogous results are expected to extend to more general learning scenarios. | [
"cs.LG",
"cs.DS",
"math.ST",
"stat.ML",
"stat.TH"
] |
The automatic generation of radiology reports given medical radiographs has
significant potential to operationally and improve clinical patient care. A
number of prior works have focused on this problem, employing advanced methods
from computer vision and natural language generation to produce readable
reports. However, these works often fail to account for the particular nuances
of the radiology domain, and, in particular, the critical importance of
clinical accuracy in the resulting generated reports. In this work, we present
a domain-aware automatic chest X-ray radiology report generation system which
first predicts what topics will be discussed in the report, then conditionally
generates sentences corresponding to these topics. The resulting system is
fine-tuned using reinforcement learning, considering both readability and
clinical accuracy, as assessed by the proposed Clinically Coherent Reward. We
verify this system on two datasets, Open-I and MIMIC-CXR, and demonstrate that
our model offers marked improvements on both language generation metrics and
CheXpert assessed accuracy over a variety of competitive baselines. | [
"cs.CV",
"cs.CL"
] |
Adversarial formulations such as generative adversarial networks (GANs) have
rekindled interest in two-player min-max games. A central obstacle in the
optimization of such games is the rotational dynamics that hinder their
convergence. Existing methods typically employ intuitive, carefully
hand-designed mechanisms for controlling such rotations. In this paper, we take
a novel approach to address this issue by casting min-max optimization as a
physical system. We leverage tools from physics to introduce LEAD (Least-Action
Dynamics), a second-order optimizer for min-max games. Next, using Lyapunov
stability theory and spectral analysis, we study LEAD's convergence properties
in continuous and discrete-time settings for bilinear games to demonstrate
linear convergence to the Nash equilibrium. Finally, we empirically evaluate
our method on synthetic setups and CIFAR-10 image generation to demonstrate
improvements over baseline methods. | [
"cs.LG",
"cs.GT",
"cs.MA",
"math.OC"
] |
Generative adversarial network (GAN) has become one of the most important
neural network models for classical unsupervised machine learning. A variety of
discriminator loss functions have been developed to train GAN's discriminators
and they all have a common structure: a sum of real and fake losses that only
depends on the actual and generated data respectively. One challenge associated
with an equally weighted sum of two losses is that the training may benefit one
loss but harm the other, which we show causes instability and mode collapse. In
this paper, we introduce a new family of discriminator loss functions that
adopts a weighted sum of real and fake parts, which we call adaptive weighted
loss functions or aw-loss functions. Using the gradients of the real and fake
parts of the loss, we can adaptively choose weights to train a discriminator in
the direction that benefits the GAN's stability. Our method can be potentially
applied to any discriminator model with a loss that is a sum of the real and
fake parts. Experiments validated the effectiveness of our loss functions on an
unconditional image generation task, improving the baseline results by a
significant margin on CIFAR-10, STL-10, and CIFAR-100 datasets in Inception
Scores and FID. | [
"cs.LG"
] |
Missing value imputation is a challenging and well-researched topic in data
mining. In this paper, we propose IFGAN, a missing value imputation algorithm
based on Feature-specific Generative Adversarial Networks (GAN). Our idea is
intuitive yet effective: a feature-specific generator is trained to impute
missing values, while a discriminator is expected to distinguish the imputed
values from observed ones. The proposed architecture is capable of handling
different data types, data distributions, missing mechanisms, and missing
rates. It also improves post-imputation analysis by preserving inter-feature
correlations. We empirically show on several real-life datasets that IFGAN
outperforms current state-of-the-art algorithm under various missing
conditions. | [
"cs.LG",
"stat.ML"
] |
We consider representation learning from 3D graphs in which each node is
associated with a spatial position in 3D. This is an under explored area of
research, and a principled framework is currently lacking. In this work, we
propose a generic framework, known as the 3D graph network (3DGN), to provide a
unified interface at different levels of granularity for 3D graphs. Built on
3DGN, we propose the spherical message passing (SMP) as a novel and specific
scheme for realizing the 3DGN framework in the spherical coordinate system
(SCS). We conduct formal analyses and show that the relative location of each
node in 3D graphs is uniquely defined in the SMP scheme. Thus, our SMP
represents a complete and accurate architecture for learning from 3D graphs in
the SCS. We derive physically-based representations of geometric information
and propose the SphereNet for learning representations of 3D graphs. We show
that existing 3D deep models can be viewed as special cases of the SphereNet.
Experimental results demonstrate that the use of complete and accurate 3D
information in 3DGN and SphereNet leads to significant performance improvements
in prediction tasks. | [
"cs.LG"
] |
Recently, generative adversarial networks have gained a lot of popularity for
image generation tasks. However, such models are associated with complex
learning mechanisms and demand very large relevant datasets. This work borrows
concepts from image and video captioning models to form an image generative
framework. The model is trained in a similar fashion as recurrent captioning
model and uses the learned weights for image generation. This is done in an
inverse direction, where the input is a caption and the output is an image. The
vector representation of the sentence and frames are extracted from an
encoder-decoder model which is initially trained on similar sentence and image
pairs. Our model conditions image generation on a natural language caption. We
leverage a sequence-to-sequence model to generate synthetic captions that have
the same meaning for having a robust image generation. One key advantage of our
method is that the traditional image captioning datasets can be used for
synthetic sentence paraphrases. Results indicate that images generated through
multiple captions are better at capturing the semantic meaning of the family of
captions. | [
"cs.LG",
"stat.ML"
] |
Generative models are now capable of producing highly realistic images that
look nearly indistinguishable from the data on which they are trained. This
raises the question: if we have good enough generative models, do we still need
datasets? We investigate this question in the setting of learning
general-purpose visual representations from a black-box generative model rather
than directly from data. Given an off-the-shelf image generator without any
access to its training data, we train representations from the samples output
by this generator. We compare several representation learning methods that can
be applied to this setting, using the latent space of the generator to generate
multiple "views" of the same semantic content. We show that for contrastive
methods, this multiview data can naturally be used to identify positive pairs
(nearby in latent space) and negative pairs (far apart in latent space). We
find that the resulting representations rival those learned directly from real
data, but that good performance requires care in the sampling strategy applied
and the training method. Generative models can be viewed as a compressed and
organized copy of a dataset, and we envision a future where more and more
"model zoos" proliferate while datasets become increasingly unwieldy, missing,
or private. This paper suggests several techniques for dealing with visual
representation learning in such a future. Code is released on our project page:
https://ali-design.github.io/GenRep/ | [
"cs.CV"
] |
An increasing need of running Convolutional Neural Network (CNN) models on
mobile devices with limited computing power and memory resource encourages
studies on efficient model design. A number of efficient architectures have
been proposed in recent years, for example, MobileNet, ShuffleNet, and
MobileNetV2. However, all these models are heavily dependent on depthwise
separable convolution which lacks efficient implementation in most deep
learning frameworks. In this study, we propose an efficient architecture named
PeleeNet, which is built with conventional convolution instead. On ImageNet
ILSVRC 2012 dataset, our proposed PeleeNet achieves a higher accuracy and over
1.8 times faster speed than MobileNet and MobileNetV2 on NVIDIA TX2. Meanwhile,
PeleeNet is only 66% of the model size of MobileNet. We then propose a
real-time object detection system by combining PeleeNet with Single Shot
MultiBox Detector (SSD) method and optimizing the architecture for fast speed.
Our proposed detection system2, named Pelee, achieves 76.4% mAP (mean average
precision) on PASCAL VOC2007 and 22.4 mAP on MS COCO dataset at the speed of
23.6 FPS on iPhone 8 and 125 FPS on NVIDIA TX2. The result on COCO outperforms
YOLOv2 in consideration of a higher precision, 13.6 times lower computational
cost and 11.3 times smaller model size. | [
"cs.CV"
] |
Active learning approaches in computer vision generally involve querying
strong labels for data. However, previous works have shown that weak
supervision can be effective in training models for vision tasks while greatly
reducing annotation costs. Using this knowledge, we propose an adaptive
supervision framework for active learning and demonstrate its effectiveness on
the task of object detection. Instead of directly querying bounding box
annotations (strong labels) for the most informative samples, we first query
weak labels and optimize the model. Using a switching condition, the required
supervision level can be increased. Our framework requires little to no change
in model architecture. Our extensive experiments show that the proposed
framework can be used to train good generalizable models with much lesser
annotation costs than the state of the art active learning approaches for
object detection. | [
"cs.CV"
] |
State-of-the-art named entity recognition (NER) systems have been improving
continuously using neural architectures over the past several years. However,
many tasks including NER require large sets of annotated data to achieve such
performance. In particular, we focus on NER from clinical notes, which is one
of the most fundamental and critical problems for medical text analysis. Our
work centers on effectively adapting these neural architectures towards
low-resource settings using parameter transfer methods. We complement a
standard hierarchical NER model with a general transfer learning framework
consisting of parameter sharing between the source and target tasks, and
showcase scores significantly above the baseline architecture. These sharing
schemes require an exponential search over tied parameter sets to generate an
optimal configuration. To mitigate the problem of exhaustively searching for
model optimization, we propose the Dynamic Transfer Networks (DTN), a gated
architecture which learns the appropriate parameter sharing scheme between
source and target datasets. DTN achieves the improvements of the optimized
transfer learning framework with just a single training setting, effectively
removing the need for exponential search. | [
"cs.LG",
"cs.CL",
"stat.ML"
] |
Plane detection in 3D point clouds is a crucial pre-processing step for
applications such as point cloud segmentation, semantic mapping and SLAM. In
contrast to many recent plane detection methods that are only applicable on
organized point clouds, our work is targeted to unorganized point clouds that
do not permit a 2D parametrization. We compare three methods for detecting
planes in point clouds efficiently. One is a novel method proposed in this
paper that generates plane hypotheses by sampling from a set of points with
estimated normals. We named this method Oriented Point Sampling (OPS) to
contrast with more conventional techniques that require the sampling of three
unoriented points to generate plane hypotheses. We also implemented an
efficient plane detection method based on local sampling of three unoriented
points and compared it with OPS and the 3D-KHT algorithm, which is based on
octrees, on the detection of planes on 10,000 point clouds from the SUN RGB-D
dataset. | [
"cs.CV",
"cs.RO"
] |
This study aims to automatically diagnose thoracic diseases depicted on the
chest x-ray (CXR) images using deep convolutional neural networks. The existing
methods generally used the entire CXR images for training purposes, but this
strategy may suffer from two drawbacks. First, potential misalignment or the
existence of irrelevant objects in the entire CXR images may cause unnecessary
noise and thus limit the network performance. Second, the relatively low image
resolution caused by the resizing operation, which is a common preprocessing
procedure for training neural networks, may lead to the loss of image details,
making it difficult to detect pathologies with small lesion regions. To address
these issues, we present a novel method termed as segmentation-based deep
fusion network (SDFN), which leverages the domain knowledge and the
higherresolution information of local lung regions. Specifically, the local
lung regions were identified and cropped by the Lung Region Generator (LRG).
Two CNN-based classification models were then used as feature extractors to
obtain the discriminative features of the entire CXR images and the cropped
lung region images. Lastly, the obtained features were fused by the feature
fusion module for disease classification. Evaluated by the NIH benchmark split
on the Chest X-ray 14 Dataset, our experimental result demonstrated that the
developed method achieved more accurate disease classification compared with
the available approaches via the receiver operating characteristic (ROC)
analyses. It was also found that the SDFN could localize the lesion regions
more precisely as compared to the traditional method. | [
"cs.CV"
] |
Encoder-decoder architectures are widely adopted for medical image
segmentation tasks. With the lateral skip connection, the models can obtain and
fuse both semantic and resolution information in deep layers to achieve more
accurate segmentation performance. However, in many applications (e.g., blurry
boundary images), these models often cannot precisely locate complex boundaries
and segment tiny isolated parts. To solve this challenging problem, we firstly
analyze why simple skip connections are not enough to help accurately locate
indistinct boundaries and argue that it is due to the fuzzy information in the
skip connection provided in the encoder layers. Then we propose a
semantic-guided encoder feature learning strategy to learn both high resolution
and rich semantic encoder features so that we can more accurately locate the
blurry boundaries, which can also enhance the network by selectively learning
discriminative features. Besides, we further propose a soft contour constraint
mechanism to model the blurry boundary detection. Experimental results on real
clinical datasets show that our proposed method can achieve state-of-the-art
segmentation accuracy, especially for the blurry regions. Further analysis also
indicates that our proposed network components indeed contribute to the
improvement of performance. Experiments on additional datasets validate the
generalization ability of our proposed method. | [
"cs.CV"
] |
Modelers use automatic differentiation (AD) of computation graphs to
implement complex Deep Learning models without defining gradient computations.
Stochastic AD extends AD to stochastic computation graphs with sampling steps,
which arise when modelers handle the intractable expectations common in
Reinforcement Learning and Variational Inference. However, current methods for
stochastic AD are limited: They are either only applicable to continuous random
variables and differentiable functions, or can only use simple but high
variance score-function estimators. To overcome these limitations, we introduce
Storchastic, a new framework for AD of stochastic computation graphs.
Storchastic allows the modeler to choose from a wide variety of gradient
estimation methods at each sampling step, to optimally reduce the variance of
the gradient estimates. Furthermore, Storchastic is provably unbiased for
estimation of any-order gradients, and generalizes variance reduction
techniques to higher-order gradient estimates. Finally, we implement
Storchastic as a PyTorch library. | [
"stat.ML",
"cs.AI",
"cs.LG"
] |
Point clouds have attracted increasing attention. Significant progress has
been made in methods for point cloud analysis, which often requires costly
human annotation as supervision. To address this issue, we propose a novel
self-contrastive learning for self-supervised point cloud representation
learning, aiming to capture both local geometric patterns and nonlocal semantic
primitives based on the nonlocal self-similarity of point clouds. The
contributions are two-fold: on the one hand, instead of contrasting among
different point clouds as commonly employed in contrastive learning, we exploit
self-similar point cloud patches within a single point cloud as positive
samples and otherwise negative ones to facilitate the task of contrastive
learning. On the other hand, we actively learn hard negative samples that are
close to positive samples for discriminative feature learning. Experimental
results show that the proposed method achieves state-of-the-art performance on
widely used benchmark datasets for self-supervised point cloud segmentation and
transfer learning for classification. | [
"cs.CV"
] |
We introduce NExT-QA, a rigorously designed video question answering
(VideoQA) benchmark to advance video understanding from describing to
explaining the temporal actions. Based on the dataset, we set up multi-choice
and open-ended QA tasks targeting causal action reasoning, temporal action
reasoning, and common scene comprehension. Through extensive analysis of
baselines and established VideoQA techniques, we find that top-performing
methods excel at shallow scene descriptions but are weak in causal and temporal
action reasoning. Furthermore, the models that are effective on multi-choice
QA, when adapted to open-ended QA, still struggle in generalizing the answers.
This raises doubt on the ability of these models to reason and highlights
possibilities for improvement. With detailed results for different question
types and heuristic observations for future works, we hope NExT-QA will guide
the next generation of VQA research to go beyond superficial scene description
towards a deeper understanding of videos. (The dataset and related resources
are available at https://github.com/doc-doc/NExT-QA.git) | [
"cs.CV",
"cs.AI"
] |
Designing a new drug is a lengthy and expensive process. As the space of
potential molecules is very large (10^23-10^60), a common technique during drug
discovery is to start from a molecule which already has some of the desired
properties. An interdisciplinary team of scientists generates hypothesis about
the required changes to the prototype. In this work, we develop an algorithmic
unsupervised-approach that automatically generates potential drug molecules
given a prototype drug. We show that the molecules generated by the system are
valid molecules and significantly different from the prototype drug. Out of the
compounds generated by the system, we identified 35 FDA-approved drugs. As an
example, our system generated Isoniazid - one of the main drugs for
Tuberculosis. The system is currently being deployed for use in collaboration
with pharmaceutical companies to further analyze the additional generated
molecules. | [
"cs.LG",
"stat.ML"
] |
Unsupervised (or self-supervised) graph representation learning is essential
to facilitate various graph data mining tasks when external supervision is
unavailable. The challenge is to encode the information about the graph
structure and the attributes associated with the nodes and edges into a low
dimensional space. Most existing unsupervised methods promote similar
representations across nodes that are topologically close. Recently, it was
shown that leveraging additional graph-level information, e.g., information
that is shared among all nodes, encourages the representations to be mindful of
the global properties of the graph, which greatly improves their quality.
However, in most graphs, there is significantly more structure that can be
captured, e.g., nodes tend to belong to (multiple) clusters that represent
structurally similar nodes. Motivated by this observation, we propose a graph
representation learning method called Graph InfoClust (GIC), that seeks to
additionally capture cluster-level information content. These clusters are
computed by a differentiable K-means method and are jointly optimized by
maximizing the mutual information between nodes of the same clusters. This
optimization leads the node representations to capture richer information and
nodal interactions, which improves their quality. Experiments show that GIC
outperforms state-of-art methods in various downstream tasks (node
classification, link prediction, and node clustering) with a 0.9% to 6.1% gain
over the best competing approach, on average. | [
"cs.LG",
"stat.ML"
] |
A major impediment in rapidly deploying object detection models for instance
detection is the lack of large annotated datasets. For example, finding a large
labeled dataset containing instances in a particular kitchen is unlikely. Each
new environment with new instances requires expensive data collection and
annotation. In this paper, we propose a simple approach to generate large
annotated instance datasets with minimal effort. Our key insight is that
ensuring only patch-level realism provides enough training signal for current
object detector models. We automatically `cut' object instances and `paste'
them on random backgrounds. A naive way to do this results in pixel artifacts
which result in poor performance for trained models. We show how to make
detectors ignore these artifacts during training and generate data that gives
competitive performance on real data. Our method outperforms existing synthesis
approaches and when combined with real images improves relative performance by
more than 21% on benchmark datasets. In a cross-domain setting, our synthetic
data combined with just 10% real data outperforms models trained on all real
data. | [
"cs.CV"
] |
Image reconstruction including image restoration and denoising is a
challenging problem in the field of image computing. We present a new method,
called X-GANs, for reconstruction of arbitrary corrupted resource based on a
variant of conditional generative adversarial networks (conditional GANs). In
our method, a novel generator and multi-scale discriminators are proposed, as
well as the combined adversarial losses, which integrate a VGG perceptual loss,
an adversarial perceptual loss, and an elaborate corresponding point loss
together based on the analysis of image feature. Our conditional GANs have
enabled a variety of applications in image reconstruction, including image
denoising, image restoration from quite a sparse sampling, image inpainting,
image recovery from the severely polluted block or even color-noise dominated
images, which are extreme cases and haven't been addressed in the status quo.
We have significantly improved the accuracy and quality of image
reconstruction. Extensive perceptual experiments on datasets ranging from human
faces to natural scenes demonstrate that images reconstructed by the presented
approach are considerably more realistic than alternative work. Our method can
also be extended to handle high-ratio image compression. | [
"cs.CV",
"cs.MM"
] |
In recent years, deep learning has become prevalent to solve applications
from multiple domains. Convolutional Neural Networks (CNNs) particularly have
demonstrated state of the art performance for the task of image classification.
However, the decisions made by these networks are not transparent and cannot be
directly interpreted by a human. Several approaches have been proposed to
explain to understand the reasoning behind a prediction made by a network. In
this paper, we propose a topology of grouping these methods based on their
assumptions and implementations. We focus primarily on white box methods that
leverage the information of the internal architecture of a network to explain
its decision. Given the task of image classification and a trained CNN, this
work aims to provide a comprehensive and detailed overview of a set of methods
that can be used to create explanation maps for a particular image, that assign
an importance score to each pixel of the image based on its contribution to the
decision of the network. We also propose a further classification of the white
box methods based on their implementations to enable better comparisons and
help researchers find methods best suited for different scenarios. | [
"cs.CV",
"cs.AI",
"cs.LG"
] |
Learning interpretable and disentangled representations is a crucial yet
challenging task in representation learning. In this work, we focus on
semi-supervised disentanglement learning and extend work by Locatello et al.
(2019) by introducing another source of supervision that we denote as label
replacement. Specifically, during training, we replace the inferred
representation associated with a data point with its ground-truth
representation whenever it is available. Our extension is theoretically
inspired by our proposed general framework of semi-supervised disentanglement
learning in the context of VAEs which naturally motivates the supervised terms
commonly used in existing semi-supervised VAEs (but not for disentanglement
learning). Extensive experiments on synthetic and real datasets demonstrate
both quantitatively and qualitatively the ability of our extension to
significantly and consistently improve disentanglement with very limited
supervision. | [
"cs.LG",
"stat.ML"
] |
Impressive progress has been made in the fields of computer vision and
natural language processing. However, it remains a challenge to find the best
point of interaction for these very different modalities. In this chapter we
discuss how attributes allow us to exchange information between the two
modalities and in this way lead to an interaction on a semantic level.
Specifically we discuss how attributes allow using knowledge mined from
language resources for recognizing novel visual categories, how we can generate
sentence description about images and video, how we can ground natural language
in visual content, and finally, how we can answer natural language questions
about images. | [
"cs.CV",
"cs.CL"
] |
Automatically describing an image with a natural language has been an
emerging challenge in both fields of computer vision and natural language
processing. In this paper, we present Long Short-Term Memory with Attributes
(LSTM-A) - a novel architecture that integrates attributes into the successful
Convolutional Neural Networks (CNNs) plus Recurrent Neural Networks (RNNs)
image captioning framework, by training them in an end-to-end manner. To
incorporate attributes, we construct variants of architectures by feeding image
representations and attributes into RNNs in different ways to explore the
mutual but also fuzzy relationship between them. Extensive experiments are
conducted on COCO image captioning dataset and our framework achieves superior
results when compared to state-of-the-art deep models. Most remarkably, we
obtain METEOR/CIDEr-D of 25.2%/98.6% on testing data of widely used and
publicly available splits in (Karpathy & Fei-Fei, 2015) when extracting image
representations by GoogleNet and achieve to date top-1 performance on COCO
captioning Leaderboard. | [
"cs.CV"
] |
We explore total scene capture -- recording, modeling, and rerendering a
scene under varying appearance such as season and time of day. Starting from
internet photos of a tourist landmark, we apply traditional 3D reconstruction
to register the photos and approximate the scene as a point cloud. For each
photo, we render the scene points into a deep framebuffer, and train a neural
network to learn the mapping of these initial renderings to the actual photos.
This rerendering network also takes as input a latent appearance vector and a
semantic mask indicating the location of transient objects like pedestrians.
The model is evaluated on several datasets of publicly available images
spanning a broad range of illumination conditions. We create short videos
demonstrating realistic manipulation of the image viewpoint, appearance, and
semantic labeling. We also compare results with prior work on scene
reconstruction from internet photos. | [
"cs.CV",
"cs.GR"
] |
Recognising detailed clothing characteristics (fine-grained attributes) in
unconstrained images of people in-the-wild is a challenging task for computer
vision, especially when there is only limited training data from the wild
whilst most data available for model learning are captured in well-controlled
environments using fashion models (well lit, no background clutter, frontal
view, high-resolution). In this work, we develop a deep learning framework
capable of model transfer learning from well-controlled shop clothing images
collected from web retailers to in-the-wild images from the street.
Specifically, we formulate a novel Multi-Task Curriculum Transfer (MTCT) deep
learning method to explore multiple sources of different types of web
annotations with multi-labelled fine-grained attributes. Our multi-task loss
function is designed to extract more discriminative representations in training
by jointly learning all attributes, and our curriculum strategy exploits the
staged easy-to-complex transfer learning motivated by cognitive studies. We
demonstrate the advantages of the MTCT model over the state-of-the-art methods
on the X-Domain benchmark, a large scale clothing attribute dataset. Moreover,
we show that the MTCT model has a notable advantage over contemporary models
when the training data size is small. | [
"cs.CV"
] |
Conditional quantile estimation is a key statistical learning challenge
motivated by the need to quantify uncertainty in predictions or to model a
diverse population without being overly reductive. As such, many models have
been developed for this problem. Adopting a meta viewpoint, we propose a
general framework (inspired by neural network optimization) for aggregating any
number of conditional quantile models in order to boost predictive accuracy. We
consider weighted ensembling strategies of increasing flexibility where the
weights may vary over individual models, quantile levels, and feature values.
An appeal of our approach is its portability: we ensure that estimated
quantiles at adjacent levels do not cross by applying simple transformations
through which gradients can be backpropagated, and this allows us to leverage
the modern deep learning toolkit for building quantile ensembles. Our
experiments confirm that ensembling can lead to big gains in accuracy, even
when the constituent models are themselves powerful and flexible. | [
"stat.ML",
"cs.LG"
] |
In this paper we propose a geometry-aware model for video object detection.
Specifically, we consider the setting that cameras can be well approximated as
static, e.g. in video surveillance scenarios, and scene pseudo depth maps can
therefore be inferred easily from the object scale on the image plane. We make
the following contributions: First, we extend the recent anchor-free detector
(CornerNet [17]) to video object detections. In order to exploit the
spatial-temporal information while maintaining high efficiency, the proposed
model accepts video clips as input, and only makes predictions for the starting
and the ending frames, i.e. heatmaps of object bounding box corners and the
corresponding embeddings for grouping. Second, to tackle the challenge from
scale variations in object detection, scene geometry information, e.g. derived
depth maps, is explicitly incorporated into deep networks for multi-scale
feature selection and for the network prediction. Third, we validate the
proposed architectures on an autonomous driving dataset generated from the
Carla simulator [5], and on a real dataset for human detection (DukeMTMC
dataset [28]). When comparing with the existing competitive single-stage or
two-stage detectors, the proposed geometry-aware spatio-temporal network
achieves significantly better results. | [
"cs.CV"
] |
Inspired by recent advances in multimodal learning and machine translation,
we introduce an encoder-decoder pipeline that learns (a): a multimodal joint
embedding space with images and text and (b): a novel language model for
decoding distributed representations from our space. Our pipeline effectively
unifies joint image-text embedding models with multimodal neural language
models. We introduce the structure-content neural language model that
disentangles the structure of a sentence to its content, conditioned on
representations produced by the encoder. The encoder allows one to rank images
and sentences while the decoder can generate novel descriptions from scratch.
Using LSTM to encode sentences, we match the state-of-the-art performance on
Flickr8K and Flickr30K without using object detections. We also set new best
results when using the 19-layer Oxford convolutional network. Furthermore we
show that with linear encoders, the learned embedding space captures multimodal
regularities in terms of vector space arithmetic e.g. *image of a blue car* -
"blue" + "red" is near images of red cars. Sample captions generated for 800
images are made available for comparison. | [
"cs.LG",
"cs.CL",
"cs.CV"
] |
Despite impressive performance on many text classification tasks, deep neural
networks tend to learn frequent superficial patterns that are specific to the
training data and do not always generalize well. In this work, we observe this
limitation with respect to the task of native language identification. We find
that standard text classifiers which perform well on the test set end up
learning topical features which are confounds of the prediction task (e.g., if
the input text mentions Sweden, the classifier predicts that the author's
native language is Swedish). We propose a method that represents the latent
topical confounds and a model which "unlearns" confounding features by
predicting both the label of the input text and the confound; but we train the
two predictors adversarially in an alternating fashion to learn a text
representation that predicts the correct label but is less prone to using
information about the confound. We show that this model generalizes better and
learns features that are indicative of the writing style rather than the
content. | [
"cs.LG",
"cs.CL",
"stat.ML"
] |
In this paper, we introduce InstantEmbedding, an efficient method for
generating single-node representations using local PageRank computations. We
theoretically prove that our approach produces globally consistent
representations in sublinear time. We demonstrate this empirically by
conducting extensive experiments on real-world datasets with over a billion
edges. Our experiments confirm that InstantEmbedding requires drastically less
computation time (over 9,000 times faster) and less memory (by over 8,000
times) to produce a single node's embedding than traditional methods including
DeepWalk, node2vec, VERSE, and FastRP. We also show that our method produces
high quality representations, demonstrating results that meet or exceed the
state of the art for unsupervised representation learning on tasks like node
classification and link prediction. | [
"cs.LG",
"cs.AI",
"cs.SI",
"stat.ML"
] |
There has been recently a growing interest in studying adversarial examples
on natural language models in the black-box setting. These methods attack
natural language classifiers by perturbing certain important words until the
classifier label is changed. In order to find these important words, these
methods rank all words by importance by querying the target model word by word
for each input sentence, resulting in high query inefficiency. A new
interesting approach was introduced that addresses this problem through
interpretable learning to learn the word ranking instead of previous expensive
search. The main advantage of using this approach is that it achieves
comparable attack rates to the state-of-the-art methods, yet faster and with
fewer queries, where fewer queries are desirable to avoid suspicion towards the
attacking agent. Nonetheless, this approach sacrificed the useful information
that could be leveraged from the target classifier for that sake of query
efficiency. In this paper we study the effect of leveraging the target model
outputs and data on both attack rates and average number of queries, and we
show that both can be improved, with a limited overhead of additional queries. | [
"cs.LG",
"cs.AI",
"cs.CL",
"cs.CR",
"I.5.0, I.2.0"
] |
A simple and widely adopted approach to extend Gaussian processes (GPs) to
multiple outputs is to model each output as a linear combination of a
collection of shared, unobserved latent GPs. An issue with this approach is
choosing the number of latent processes and their kernels. These choices are
typically done manually, which can be time consuming and prone to human biases.
We propose Gaussian Process Automatic Latent Process Selection (GP-ALPS), which
automatically chooses the latent processes by turning off those that do not
meaningfully contribute to explaining the data. We develop a variational
inference scheme, assess the quality of the variational posterior by comparing
it against the gold standard MCMC, and demonstrate the suitability of GP-ALPS
in a set of preliminary experiments. | [
"stat.ML",
"cs.LG",
"stat.ME"
] |
Convolutional Neural Networks (ConvNets) at present achieve remarkable
performance in image classification tasks. However, current ConvNets cannot
guarantee the capabilities of the mammalian visual systems such as invariance
to contrast and illumination changes. Some ideas to overcome the illumination
and contrast variations usually have to be tuned manually and tend to fail when
tested with other types of data degradation. In this context, we present a new
bio-inspired {entry} layer, M6, which detects low-level geometric features
(lines, edges, and orientations) which are similar to patterns detected by the
V1 visual cortex. This new trainable layer is capable of coping with image
classification even with large contrast variations. The explanation for this
behavior is the monogenic signal geometry, which represents each pixel value in
a 3D space using quaternions, a fact that confers a degree of explainability to
the networks. We compare M6 with a conventional convolutional layer (C) and a
deterministic quaternion local phase layer (Q9). The experimental setup {is
designed to evaluate the robustness} of our M6 enriched ConvNet model and
includes three architectures, four datasets, three types of contrast
degradation (including non-uniform haze degradations). The numerical results
reveal that the models with M6 are the most robust in front of any kind of
contrast variations. This amounts to a significant enhancement of the C models,
which usually have reasonably good performance only when the same training and
test degradation are used, except for the case of maximum degradation.
Moreover, the Structural Similarity Index Measure (SSIM) is used to analyze and
explain the robustness effect of the M6 feature maps under any kind of contrast
degradations. | [
"cs.CV",
"cs.AI"
] |
Processing an input signal that contains arbitrary structures, e.g.,
superpixels and point clouds, remains a big challenge in computer vision.
Linear diffusion, an effective model for image processing, has been recently
integrated with deep learning algorithms. In this paper, we propose to learn
pairwise relations among data points in a global fashion to improve semantic
segmentation with arbitrarily-structured data, through spatial generalized
propagation networks (SGPN). The network propagates information on a group of
graphs, which represent the arbitrarily-structured data, through a learned,
linear diffusion process. The module is flexible to be embedded and jointly
trained with many types of networks, e.g., CNNs. We experiment with semantic
segmentation networks, where we use our propagation module to jointly train on
different data -- images, superpixels and point clouds. We show that SGPN
consistently improves the performance of both pixel and point cloud
segmentation, compared to networks that do not contain this module. Our method
suggests an effective way to model the global pairwise relations for
arbitrarily-structured data. | [
"cs.CV"
] |
Electrocardiogram (ECG) detection and delineation are key steps for numerous
tasks in clinical practice, as ECG is the most performed non-invasive test for
assessing cardiac condition. State-of-the-art algorithms employ digital signal
processing (DSP), which require laborious rule adaptation to new morphologies.
In contrast, deep learning (DL) algorithms, especially for classification, are
gaining weight in academic and industrial settings. However, the lack of model
explainability and small databases hinder their applicability. We demonstrate
DL can be successfully applied to low interpretative tasks by embedding ECG
detection and delineation onto a segmentation framework. For this purpose, we
adapted and validated the most used neural network architecture for image
segmentation, the U-Net, to one-dimensional data. The model was trained using
PhysioNet's QT database, comprised of 105 ambulatory ECG recordings, for
single- and multi-lead scenarios. To alleviate data scarcity, data
regularization techniques such as pre-training with low-quality data labels,
performing ECG-based data augmentation and applying strong model regularizers
to the architecture were attempted. Other variations in the model's capacity
(U-Net's depth and width), alongside the application of state-of-the-art
additions, were evaluated. These variations were exhaustively validated in a
5-fold cross-validation manner. The best performing configuration reached
precisions of 90.12%, 99.14% and 98.25% and recalls of 98.73%, 99.94% and
99.88% for the P, QRS and T waves, respectively, on par with DSP-based
approaches. Despite being a data-hungry technique trained on a small dataset,
DL-based approaches demonstrate to be a viable alternative to traditional
DSP-based ECG processing techniques. | [
"cs.LG",
"eess.SP",
"stat.ML"
] |
Many vision and language models suffer from poor visual grounding - often
falling back on easy-to-learn language priors rather than basing their
decisions on visual concepts in the image. In this work, we propose a generic
approach called Human Importance-aware Network Tuning (HINT) that effectively
leverages human demonstrations to improve visual grounding. HINT encourages
deep networks to be sensitive to the same input regions as humans. Our approach
optimizes the alignment between human attention maps and gradient-based network
importances - ensuring that models learn not just to look at but rather rely on
visual concepts that humans found relevant for a task when making predictions.
We apply HINT to Visual Question Answering and Image Captioning tasks,
outperforming top approaches on splits that penalize over-reliance on language
priors (VQA-CP and robust captioning) using human attention demonstrations for
just 6% of the training data. | [
"cs.CV"
] |
Consecutive frames in a video are highly redundant. Therefore, to perform the
task of video object detection, executing single frame detectors on every frame
without reusing any information is quite wasteful. It is with this idea in mind
that we propose RN-VID (standing for RetinaNet-VIDeo), a novel approach to
video object detection. Our contributions are twofold. First, we propose a new
architecture that allows the usage of information from nearby frames to enhance
feature maps. Second, we propose a novel module to merge feature maps of same
dimensions using re-ordering of channels and 1 x 1 convolutions. We then
demonstrate that RN-VID achieves better mean average precision (mAP) than
corresponding single frame detectors with little additional cost during
inference. | [
"cs.CV"
] |
A pharmacological effect of a drug on cells, organs and systems refers to the
specific biochemical interaction produced by a drug substance, which is called
its mechanism of action. Drug repositioning (or drug repurposing) is a
fundamental problem for the identification of new opportunities for the use of
already approved or failed drugs. In this paper, we present a method based on a
multi-relation unsupervised graph embedding model that learns latent
representations for drugs and diseases so that the distance between these
representations reveals repositioning opportunities. Once representations for
drugs and diseases are obtained we learn the likelihood of new links (that is,
new indications) between drugs and diseases. Known drug indications are used
for learning a model that predicts potential indications. Compared with
existing unsupervised graph embedding methods our method shows superior
prediction performance in terms of area under the ROC curve, and we present
examples of repositioning opportunities found on recent biomedical literature
that were also predicted by our method. | [
"cs.LG",
"q-bio.QM",
"stat.ML"
] |
The work in this paper is driven by the question how to exploit the temporal
cues available in videos for their accurate classification, and for human
action recognition in particular? Thus far, the vision community has focused on
spatio-temporal approaches with fixed temporal convolution kernel depths. We
introduce a new temporal layer that models variable temporal convolution kernel
depths. We embed this new temporal layer in our proposed 3D CNN. We extend the
DenseNet architecture - which normally is 2D - with 3D filters and pooling
kernels. We name our proposed video convolutional network `Temporal 3D
ConvNet'~(T3D) and its new temporal layer `Temporal Transition Layer'~(TTL).
Our experiments show that T3D outperforms the current state-of-the-art methods
on the HMDB51, UCF101 and Kinetics datasets.
The other issue in training 3D ConvNets is about training them from scratch
with a huge labeled dataset to get a reasonable performance. So the knowledge
learned in 2D ConvNets is completely ignored. Another contribution in this work
is a simple and effective technique to transfer knowledge from a pre-trained 2D
CNN to a randomly initialized 3D CNN for a stable weight initialization. This
allows us to significantly reduce the number of training samples for 3D CNNs.
Thus, by finetuning this network, we beat the performance of generic and recent
methods in 3D CNNs, which were trained on large video datasets, e.g. Sports-1M,
and finetuned on the target datasets, e.g. HMDB51/UCF101. The T3D codes will be
released | [
"cs.CV"
] |
Current graph neural networks (GNNs) lack generalizability with respect to
scales (graph sizes, graph diameters, edge weights, etc..) when solving many
graph analysis problems. Taking the perspective of synthesizing graph theory
programs, we propose several extensions to address the issue. First, inspired
by the dependency of the iteration number of common graph theory algorithms on
graph size, we learn to terminate the message passing process in GNNs
adaptively according to the computation progress. Second, inspired by the fact
that many graph theory algorithms are homogeneous with respect to graph
weights, we introduce homogeneous transformation layers that are universal
homogeneous function approximators, to convert ordinary GNNs to be homogeneous.
Experimentally, we show that our GNN can be trained from small-scale graphs but
generalize well to large-scale graphs for a number of basic graph theory
problems. It also shows generalizability for applications of multi-body
physical simulation and image-based navigation problems. | [
"cs.LG",
"cs.CV"
] |
The availability of labeled image datasets has been shown critical for
high-level image understanding, which continuously drives the progress of
feature designing and models developing. However, constructing labeled image
datasets is laborious and monotonous. To eliminate manual annotation, in this
work, we propose a novel image dataset construction framework by employing
multiple textual queries. We aim at collecting diverse and accurate images for
given queries from the Web. Specifically, we formulate noisy textual queries
removing and noisy images filtering as a multi-view and multi-instance learning
problem separately. Our proposed approach not only improves the accuracy but
also enhances the diversity of the selected images. To verify the effectiveness
of our proposed approach, we construct an image dataset with 100 categories.
The experiments show significant performance gains by using the generated data
of our approach on several tasks, such as image classification, cross-dataset
generalization, and object detection. The proposed method also consistently
outperforms existing weakly supervised and web-supervised approaches. | [
"cs.CV",
"cs.MM"
] |
Depth data has a widespread use since the popularity of high-resolution 3D
sensors. In multi-view sequences, depth information is used to supplement the
color data of each view. This article proposes a joint encoding of multiple
depth maps with a unique representation. Color and depth images of each view
are segmented independently and combined in an optimal Rate-Distortion fashion.
The resulting partitions are projected to a reference view where a coherent
hierarchy for the multiple views is built. A Rate-Distortionoptimization is
applied to obtain the final segmentation choosing nodes of the hierarchy. The
consistent segmentation is used to robustly encode depth maps of multiple views
obtaining competitive results with HEVC coding standards. Available at:
http://link.springer.com/article/10.1007/s11042-017-5409-z | [
"cs.CV"
] |
Person re-identification (Re-ID) in real-world scenarios usually suffers from
various degradation factors, e.g., low-resolution, weak illumination, blurring
and adverse weather. On the one hand, these degradations lead to severe
discriminative information loss, which significantly obstructs identity
representation learning; on the other hand, the feature mismatch problem caused
by low-level visual variations greatly reduces retrieval performance. An
intuitive solution to this problem is to utilize low-level image restoration
methods to improve the image quality. However, existing restoration methods
cannot directly serve to real-world Re-ID due to various limitations, e.g., the
requirements of reference samples, domain gap between synthesis and reality,
and incompatibility between low-level and high-level methods. In this paper, to
solve the above problem, we propose a degradation invariance learning framework
for real-world person Re-ID. By introducing a self-supervised disentangled
representation learning strategy, our method is able to simultaneously extract
identity-related robust features and remove real-world degradations without
extra supervision. We use low-resolution images as the main demonstration, and
experiments show that our approach is able to achieve state-of-the-art
performance on several Re-ID benchmarks. In addition, our framework can be
easily extended to other real-world degradation factors, such as weak
illumination, with only a few modifications. | [
"cs.CV"
] |
We study an unsupervised domain adaptation problem for the semantic labeling
of 3D point clouds, with a particular focus on domain discrepancies induced by
different LiDAR sensors. Based on the observation that sparse 3D point clouds
are sampled from 3D surfaces, we take a Complete and Label approach to recover
the underlying surfaces before passing them to a segmentation network.
Specifically, we design a Sparse Voxel Completion Network (SVCN) to complete
the 3D surfaces of a sparse point cloud. Unlike semantic labels, to obtain
training pairs for SVCN requires no manual labeling. We also introduce local
adversarial learning to model the surface prior. The recovered 3D surfaces
serve as a canonical domain, from which semantic labels can transfer across
different LiDAR sensors. Experiments and ablation studies with our new
benchmark for cross-domain semantic labeling of LiDAR data show that the
proposed approach provides 8.2-36.6% better performance than previous domain
adaptation methods. | [
"cs.CV",
"cs.LG"
] |
Over the past decade, Deep Convolutional Neural Networks have been widely
adopted for medical image segmentation and shown to achieve adequate
performance. However, due to the inherent inductive biases present in the
convolutional architectures, they lack understanding of long-range dependencies
in the image. Recently proposed Transformer-based architectures that leverage
self-attention mechanism encode long-range dependencies and learn
representations that are highly expressive. This motivates us to explore
Transformer-based solutions and study the feasibility of using
Transformer-based network architectures for medical image segmentation tasks.
Majority of existing Transformer-based network architectures proposed for
vision applications require large-scale datasets to train properly. However,
compared to the datasets for vision applications, for medical imaging the
number of data samples is relatively low, making it difficult to efficiently
train transformers for medical applications. To this end, we propose a Gated
Axial-Attention model which extends the existing architectures by introducing
an additional control mechanism in the self-attention module. Furthermore, to
train the model effectively on medical images, we propose a Local-Global
training strategy (LoGo) which further improves the performance. Specifically,
we operate on the whole image and patches to learn global and local features,
respectively. The proposed Medical Transformer (MedT) is evaluated on three
different medical image segmentation datasets and it is shown that it achieves
better performance than the convolutional and other related transformer-based
architectures. Code: https://github.com/jeya-maria-jose/Medical-Transformer | [
"cs.CV"
] |
Recent work proposed $\delta$-relevant inputs (or sets) as a probabilistic
explanation for the predictions made by a classifier on a given input.
$\delta$-relevant sets are significant because they serve to relate
(model-agnostic) Anchors with (model-accurate) PI- explanations, among other
explanation approaches. Unfortunately, the computation of smallest size
$\delta$-relevant sets is complete for ${NP}^{PP}$, rendering their computation
largely infeasible in practice. This paper investigates solutions for tackling
the practical limitations of $\delta$-relevant sets. First, the paper
alternatively considers the computation of subset-minimal sets. Second, the
paper studies concrete families of classifiers, including decision trees among
others. For these cases, the paper shows that the computation of subset-minimal
$\delta$-relevant sets is in NP, and can be solved with a polynomial number of
calls to an NP oracle. The experimental evaluation compares the proposed
approach with heuristic explainers for the concrete case of the classifiers
studied in the paper, and confirms the advantage of the proposed solution over
the state of the art. | [
"cs.LG",
"cs.AI"
] |
X-ray baggage security screening is widely used to maintain aviation and
transport security. Of particular interest is the focus on automated security
X-ray analysis for particular classes of object such as electronics, electrical
items, and liquids. However, manual inspection of such items is challenging
when dealing with potentially anomalous items. Here we present a dual
convolutional neural network (CNN) architecture for automatic anomaly detection
within complex security X-ray imagery. We leverage recent advances in
region-based (R-CNN), mask-based CNN (Mask R-CNN) and detection architectures
such as RetinaNet to provide object localisation variants for specific object
classes of interest. Subsequently, leveraging a range of established CNN object
and fine-grained category classification approaches we formulate within object
anomaly detection as a two-class problem (anomalous or benign). While the best
performing object localisation method is able to perform with 97.9% mean
average precision (mAP) over a six-class X-ray object detection problem,
subsequent two-class anomaly/benign classification is able to achieve 66%
performance for within object anomaly detection. Overall, this performance
illustrates both the challenge and promise of object-wise anomaly detection
within the context of cluttered X-ray security imagery. | [
"cs.CV"
] |
Competition between times series often arises in sales prediction, when
similar products are on sale on a marketplace. This article provides a model of
the presence of cannibalization between times series. This model creates a
"competitiveness" function that depends on external features such as price and
margin. It also provides a theoretical guaranty on the error of the model under
some reasonable conditions, and implement this model using a neural network to
compute this competitiveness function. This implementation outperforms other
traditional time series methods and classical neural networks for market share
prediction on a real-world data set. | [
"stat.ML",
"cs.LG",
"stat.AP"
] |
With the emergence of Virtual and Mixed Reality (XR) devices, eye tracking
has received significant attention in the computer vision community. Eye gaze
estimation is a crucial component in XR -- enabling energy efficient rendering,
multi-focal displays, and effective interaction with content. In head-mounted
XR devices, the eyes are imaged off-axis to avoid blocking the field of view.
This leads to increased challenges in inferring eye related quantities and
simultaneously provides an opportunity to develop accurate and robust learning
based approaches. To this end, we present MagicEyes, the first large scale eye
dataset collected using real MR devices with comprehensive ground truth
labeling. MagicEyes includes $587$ subjects with $80,000$ images of
human-labeled ground truth and over $800,000$ images with gaze target labels.
We evaluate several state-of-the-art methods on MagicEyes and also propose a
new multi-task EyeNet model designed for detecting the cornea, glints and pupil
along with eye segmentation in a single forward pass. | [
"cs.CV",
"cs.HC",
"cs.LG",
"eess.IV"
] |
We introduce a simple (one line of code) modification to the Generative
Adversarial Network (GAN) training algorithm that materially improves results
with no increase in computational cost: When updating the generator parameters,
we simply zero out the gradient contributions from the elements of the batch
that the critic scores as `least realistic'. Through experiments on many
different GAN variants, we show that this `top-k update' procedure is a
generally applicable improvement. In order to understand the nature of the
improvement, we conduct extensive analysis on a simple mixture-of-Gaussians
dataset and discover several interesting phenomena. Among these is that, when
gradient updates are computed using the worst-scoring batch elements, samples
can actually be pushed further away from their nearest mode. We also apply our
method to recent GAN variants and improve state-of-the-art FID for conditional
generation from 9.21 to 8.57 on CIFAR-10. | [
"stat.ML",
"cs.LG"
] |
This paper introduces the PettingZoo library and the accompanying Agent
Environment Cycle ("AEC") games model. PettingZoo is a library of diverse sets
of multi-agent environments with a universal, elegant Python API. PettingZoo
was developed with the goal of accelerating research in Multi-Agent
Reinforcement Learning ("MARL"), by making work more interchangeable,
accessible and reproducible akin to what OpenAI's Gym library did for
single-agent reinforcement learning. PettingZoo's API, while inheriting many
features of Gym, is unique amongst MARL APIs in that it's based around the
novel AEC games model. We argue, in part through case studies on major problems
in popular MARL environments, that the popular game models are poor conceptual
models of the games commonly used with MARL, that they promote severe bugs that
are hard to detect, and that the AEC games model addresses these problems. | [
"cs.LG",
"cs.MA",
"stat.ML"
] |
We study how neural networks trained by gradient descent extrapolate, i.e.,
what they learn outside the support of the training distribution. Previous
works report mixed empirical results when extrapolating with neural networks:
while feedforward neural networks, a.k.a. multilayer perceptrons (MLPs), do not
extrapolate well in certain simple tasks, Graph Neural Networks (GNNs) --
structured networks with MLP modules -- have shown some success in more complex
tasks. Working towards a theoretical explanation, we identify conditions under
which MLPs and GNNs extrapolate well. First, we quantify the observation that
ReLU MLPs quickly converge to linear functions along any direction from the
origin, which implies that ReLU MLPs do not extrapolate most nonlinear
functions. But, they can provably learn a linear target function when the
training distribution is sufficiently "diverse". Second, in connection to
analyzing the successes and limitations of GNNs, these results suggest a
hypothesis for which we provide theoretical and empirical evidence: the success
of GNNs in extrapolating algorithmic tasks to new data (e.g., larger graphs or
edge weights) relies on encoding task-specific non-linearities in the
architecture or features. Our theoretical analysis builds on a connection of
over-parameterized networks to the neural tangent kernel. Empirically, our
theory holds across different training settings. | [
"cs.LG",
"cs.AI",
"cs.CV",
"stat.ML"
] |
With the success of deep neural networks, knowledge distillation which guides
the learning of a small student network from a large teacher network is being
actively studied for model compression and transfer learning. However, few
studies have been performed to resolve the poor learning issue of the student
network when the student and teacher model sizes significantly differ. In this
paper, we propose a densely guided knowledge distillation using multiple
teacher assistants that gradually decreases the model size to efficiently
bridge the large gap between the teacher and student networks. To stimulate
more efficient learning of the student network, we guide each teacher assistant
to every other smaller teacher assistants iteratively. Specifically, when
teaching a smaller teacher assistant at the next step, the existing larger
teacher assistants from the previous step are used as well as the teacher
network. Moreover, we design stochastic teaching where, for each mini-batch, a
teacher or teacher assistants are randomly dropped. This acts as a regularizer
to improve the efficiency of teaching of the student network. Thus, the student
can always learn salient distilled knowledge from the multiple sources. We
verified the effectiveness of the proposed method for a classification task
using CIFAR-10, CIFAR-100, and ImageNet. We also achieved significant
performance improvements with various backbone architectures such as ResNet,
WideResNet, and VGG. | [
"cs.CV"
] |
Machine learning on graph-structured data has attracted high research
interest due to the emergence of Graph Neural Networks (GNNs). Most of the
proposed GNNs are based on the node homophily, i.e neighboring nodes share
similar characteristics. However, in many complex networks, nodes that lie to
distant parts of the graph share structurally equivalent characteristics and
exhibit similar roles (e.g chemical properties of distant atoms in a molecule,
type of social network users). A growing literature proposed representations
that identify structurally equivalent nodes. However, most of the existing
methods require high time and space complexity. In this paper, we propose
VNEstruct, a simple approach, based on entropy measures of the neighborhood's
topology, for generating low-dimensional structural representations, that is
time-efficient and robust to graph perturbations. Empirically, we observe that
VNEstruct exhibits robustness on structural role identification tasks.
Moreover, VNEstruct can achieve state-of-the-art performance on graph
classification, without incorporating the graph structure information in the
optimization, in contrast to GNN competitors. | [
"cs.LG",
"cs.SI"
] |
Empirically it has been observed that the performance of deep neural networks
steadily improves as we increase model size, contradicting the classical view
on overfitting and generalization. Recently, the double descent phenomena has
been proposed to reconcile this observation with theory, suggesting that the
test error has a second descent when the model becomes sufficiently
overparameterized, as the model size itself acts as an implicit regularizer. In
this paper we add to the growing body of work in this space, providing a
careful study of learning dynamics as a function of model size for the least
squares scenario. We show an excess risk bound for the gradient descent
solution of the least squares objective. The bound depends on the smallest
non-zero eigenvalue of the covariance matrix of the input features, via a
functional form that has the double descent behavior. This gives a new
perspective on the double descent curves reported in the literature. Our
analysis of the excess risk allows to decouple the effect of optimization and
generalization error. In particular, we find that in case of noiseless
regression, double descent is explained solely by optimization-related
quantities, which was missed in studies focusing on the Moore-Penrose
pseudoinverse solution. We believe that our derivation provides an alternative
view compared to existing work, shedding some light on a possible cause of this
phenomena, at least in the considered least squares setting. We empirically
explore if our predictions hold for neural networks, in particular whether the
covariance of intermediary hidden activations has a similar behavior as the one
predicted by our derivations. | [
"cs.LG",
"math.OC",
"stat.ML"
] |
One of the main motivations for training high quality image generative models
is their potential use as tools for image manipulation. Recently, generative
adversarial networks (GANs) have been able to generate images of remarkable
quality. Unfortunately, adversarially-trained unconditional generator networks
have not been successful as image priors. One of the main requirements for a
network to act as a generative image prior, is being able to generate every
possible image from the target distribution. Adversarial learning often
experiences mode-collapse, which manifests in generators that cannot generate
some modes of the target distribution. Another requirement often not satisfied
is invertibility i.e. having an efficient way of finding a valid input latent
code given a required output image. In this work, we show that differently from
earlier GANs, the very recently proposed style-generators are quite easy to
invert. We use this important observation to propose style generators as
general purpose image priors. We show that style generators outperform other
GANs as well as Deep Image Prior as priors for image enhancement tasks. The
latent space spanned by style-generators satisfies linear identity-pose
relations. The latent space linearity, combined with invertibility, allows us
to animate still facial images without supervision. Extensive experiments are
performed to support the main contributions of this paper. | [
"cs.CV",
"cs.LG",
"stat.ML"
] |
Predicting chemical properties from the structure of a molecule is of great
importance in many applications including drug discovery and material design.
Machine learning based molecular property prediction holds the promise of
enabling accurate predictions at much less complexity, when compared to, for
example Density Functional Theory (DFT) calculations. Features extracted from
molecular graphs, using graph neural nets in a supervised manner, have emerged
as strong baselines for such tasks. However, the vast chemical space together
with the limited availability of labels makes supervised learning challenging,
calling for learning a general-purpose molecular representation. Recently,
pre-trained transformer-based language models (PTLMs) on large unlabeled corpus
have produced state-of-the-art results in many downstream natural language
processing tasks. Inspired by this development, here we present molecular
embeddings obtained by training an efficient transformer encoder model,
referred to as MoLFormer. This model was employed with a linear attention
mechanism and highly paralleized training on 1D SMILES sequences of 1.1 billion
unlabeled molecules from the PubChem and ZINC datasets. Experiments show that
the learned molecular representation performs competitively, when compared to
existing graph-based and fingerprint-based supervised learning baselines, on
the challenging tasks of predicting properties of QM8 and QM9 molecules.
Further task-specific fine-tuning of the MoLFormerr representation improves
performance on several of those property prediction benchmarks. These results
provide encouraging evidence that large-scale molecular language models can
capture sufficient structural information to be able to accurately predict
quantum chemical properties and beyond. | [
"cs.LG",
"cs.CL",
"q-bio.BM"
] |
The objective of Open set recognition (OSR) is to learn a classifier that can
reject the unknown samples while classifying the known classes accurately. In
this paper, we propose a self-supervision method, Detransformation Autoencoder
(DTAE), for the OSR problem. This proposed method engages in learning
representations that are invariant to the transformations of the input data.
Experiments on several standard image datasets indicate that the pre-training
process significantly improves the model performance in the OSR tasks.
Meanwhile, our proposed self-supervision method achieves significant gains in
detecting the unknown class and classifying the known classes. Moreover, our
analysis indicates that DTAE can yield representations that contain more target
class information and less transformation information than RotNet. | [
"cs.LG",
"cs.CV"
] |
In reinforcement learning, an agent learns to reach a set of goals by means
of an external reward signal. In the natural world, intelligent organisms learn
from internal drives, bypassing the need for external signals, which is
beneficial for a wide range of tasks. Motivated by this observation, we propose
to formulate an intrinsic objective as the mutual information between the goal
states and the controllable states. This objective encourages the agent to take
control of its environment. Subsequently, we derive a surrogate objective of
the proposed reward function, which can be optimized efficiently. Lastly, we
evaluate the developed framework in different robotic manipulation and
navigation tasks and demonstrate the efficacy of our approach. A video showing
experimental results is available at https://youtu.be/CT4CKMWBYz0 | [
"cs.LG",
"stat.ML"
] |
With the recent advances in complex networks theory, graph-based techniques
for image segmentation has attracted great attention recently. In order to
segment the image into meaningful connected components, this paper proposes an
image segmentation general framework using complex networks based community
detection algorithms. If we consider regions as communities, using community
detection algorithms directly can lead to an over-segmented image. To address
this problem, we start by splitting the image into small regions using an
initial segmentation. The obtained regions are used for building the complex
network. To produce meaningful connected components and detect homogeneous
communities, some combinations of color and texture based features are employed
in order to quantify the regions similarities. To sum up, the network of
regions is constructed adaptively to avoid many small regions in the image, and
then, community detection algorithms are applied on the resulting adaptive
similarity matrix to obtain the final segmented image. Experiments are
conducted on Berkeley Segmentation Dataset and four of the most influential
community detection algorithms are tested. Experimental results have shown that
the proposed general framework increases the segmentation performances compared
to some existing methods. | [
"cs.CV",
"cs.LG",
"stat.ML"
] |
Semantic segmentation has been a long standing challenging task in computer
vision. It aims at assigning a label to each image pixel and needs significant
number of pixellevel annotated data, which is often unavailable. To address
this lack, in this paper, we leverage, on one hand, massive amount of available
unlabeled or weakly labeled data, and on the other hand, non-real images
created through Generative Adversarial Networks. In particular, we propose a
semi-supervised framework ,based on Generative Adversarial Networks (GANs),
which consists of a generator network to provide extra training examples to a
multi-class classifier, acting as discriminator in the GAN framework, that
assigns sample a label y from the K possible classes or marks it as a fake
sample (extra class). The underlying idea is that adding large fake visual data
forces real samples to be close in the feature space, enabling a bottom-up
clustering process, which, in turn, improves multiclass pixel classification.
To ensure higher quality of generated images for GANs with consequent improved
pixel classification, we extend the above framework by adding weakly annotated
data, i.e., we provide class level information to the generator. We tested our
approaches on several challenging benchmarking visual datasets, i.e. PASCAL,
SiftFLow, Stanford and CamVid, achieving competitive performance also compared
to state-of-the-art semantic segmentation method | [
"cs.CV"
] |
Temporal action localization (TAL) is a fundamental yet challenging task in
video understanding. Existing TAL methods rely on pre-training a video encoder
through action classification supervision. This results in a task discrepancy
problem for the video encoder -- trained for action classification, but used
for TAL. Intuitively, end-to-end model optimization is a good solution.
However, this is not operable for TAL subject to the GPU memory constraints,
due to the prohibitive computational cost in processing long untrimmed videos.
In this paper, we resolve this challenge by introducing a novel low-fidelity
end-to-end (LoFi) video encoder pre-training method. Instead of always using
the full training configurations for TAL learning, we propose to reduce the
mini-batch composition in terms of temporal, spatial or spatio-temporal
resolution so that end-to-end optimization for the video encoder becomes
operable under the memory conditions of a mid-range hardware budget. Crucially,
this enables the gradient to flow backward through the video encoder from a TAL
loss supervision, favourably solving the task discrepancy problem and providing
more effective feature representations. Extensive experiments show that the
proposed LoFi pre-training approach can significantly enhance the performance
of existing TAL methods. Encouragingly, even with a lightweight ResNet18 based
video encoder in a single RGB stream, our method surpasses two-stream ResNet50
based alternatives with expensive optical flow, often by a good margin. | [
"cs.CV"
] |
Sleep staging is fundamental for sleep assessment and disease diagnosis.
Although previous attempts to classify sleep stages have achieved high
classification performance, several challenges remain open: 1) How to
effectively extract salient waves in multimodal sleep data; 2) How to capture
the multi-scale transition rules among sleep stages; 3) How to adaptively seize
the key role of specific modality for sleep staging. To address these
challenges, we propose SalientSleepNet, a multimodal salient wave detection
network for sleep staging. Specifically, SalientSleepNet is a temporal fully
convolutional network based on the $\rm U^2$-Net architecture that is
originally proposed for salient object detection in computer vision. It is
mainly composed of two independent $\rm U^2$-like streams to extract the
salient features from multimodal data, respectively. Meanwhile, the multi-scale
extraction module is designed to capture multi-scale transition rules among
sleep stages. Besides, the multimodal attention module is proposed to
adaptively capture valuable information from multimodal data for the specific
sleep stage. Experiments on the two datasets demonstrate that SalientSleepNet
outperforms the state-of-the-art baselines. It is worth noting that this model
has the least amount of parameters compared with the existing deep neural
network models. | [
"cs.LG",
"eess.SP"
] |
In this work, we propose an interactive system to design diverse high-quality
garment images from fashion sketches and the texture information. The major
challenge behind this system is to generate high-quality and detailed texture
according to the user-provided texture information. Prior works mainly use the
texture patch representation and try to map a small texture patch to a whole
garment image, hence unable to generate high-quality details. In contrast,
inspired by intrinsic image decomposition, we decompose this task into texture
synthesis and shading enhancement. In particular, we propose a novel bi-colored
edge texture representation to synthesize textured garment images and a shading
enhancer to render shading based on the grayscale edges. The bi-colored edge
representation provides simple but effective texture cues and color
constraints, so that the details can be better reconstructed. Moreover, with
the rendered shading, the synthesized garment image becomes more vivid. | [
"cs.CV"
] |
Recently, several approaches have been proposed to solve language generation
problems. Transformer is currently state-of-the-art seq-to-seq model in
language generation. Reinforcement Learning (RL) is useful in solving exposure
bias and the optimisation on non-differentiable metrics in seq-to-seq language
learning. However, Transformer is hard to combine with RL as the costly
computing resource is required for sampling. We tackle this problem by
proposing an off-policy RL learning algorithm where a behaviour policy
represented by GRUs performs the sampling. We reduce the high variance of
importance sampling (IS) by applying the truncated relative importance sampling
(TRIS) technique and Kullback-Leibler (KL)-control concept. TRIS is a simple
yet effective technique, and there is a theoretical proof that KL-control helps
to reduce the variance of IS. We formulate this off-policy RL based on
self-critical sequence training. Specifically, we use a Transformer-based
captioning model as the target policy and use an image-guided language
auto-encoder as the behaviour policy to explore the environment. The proposed
algorithm achieves state-of-the-art performance on the visual paragraph
generation and improved results on image captioning. | [
"cs.CV",
"cs.LG"
] |
Driven by deep learning and the large volume of data, scene text recognition
has evolved rapidly in recent years. Formerly, RNN-attention based methods have
dominated this field, but suffer from the problem of \textit{attention drift}
in certain situations. Lately, semantic segmentation based algorithms have
proven effective at recognizing text of different forms (horizontal, oriented
and curved). However, these methods may produce spurious characters or miss
genuine characters, as they rely heavily on a thresholding procedure operated
on segmentation maps. To tackle these challenges, we propose in this paper an
alternative approach, called TextScanner, for scene text recognition.
TextScanner bears three characteristics: (1) Basically, it belongs to the
semantic segmentation family, as it generates pixel-wise, multi-channel
segmentation maps for character class, position and order; (2) Meanwhile, akin
to RNN-attention based methods, it also adopts RNN for context modeling; (3)
Moreover, it performs paralleled prediction for character position and class,
and ensures that characters are transcripted in correct order. The experiments
on standard benchmark datasets demonstrate that TextScanner outperforms the
state-of-the-art methods. Moreover, TextScanner shows its superiority in
recognizing more difficult text such Chinese transcripts and aligning with
target characters. | [
"cs.CV",
"cs.CL",
"cs.LG"
] |
In many real-world prediction tasks, class labels include information about
the relative ordering between labels, which is not captured by commonly-used
loss functions such as multi-category cross-entropy. Recently, the deep
learning community adopted ordinal regression frameworks to take such ordering
information into account. Neural networks were equipped with ordinal regression
capabilities by transforming ordinal targets into binary classification
subtasks. However, this method suffers from inconsistencies among the different
binary classifiers. To resolve these inconsistencies, we propose the COnsistent
RAnk Logits (CORAL) framework with strong theoretical guarantees for
rank-monotonicity and consistent confidence scores. Moreover, the proposed
method is architecture-agnostic and can extend arbitrary state-of-the-art deep
neural network classifiers for ordinal regression tasks. The empirical
evaluation of the proposed rank-consistent method on a range of face-image
datasets for age prediction shows a substantial reduction of the prediction
error compared to the reference ordinal regression network. | [
"cs.LG",
"stat.ML"
] |
In this paper, we first provide a new perspective to divide existing high
performance object detection methods into direct and indirect regressions.
Direct regression performs boundary regression by predicting the offsets from a
given point, while indirect regression predicts the offsets from some bounding
box proposals. Then we analyze the drawbacks of the indirect regression, which
the recent state-of-the-art detection structures like Faster-RCNN and SSD
follows, for multi-oriented scene text detection, and point out the potential
superiority of direct regression. To verify this point of view, we propose a
deep direct regression based method for multi-oriented scene text detection.
Our detection framework is simple and effective with a fully convolutional
network and one-step post processing. The fully convolutional network is
optimized in an end-to-end way and has bi-task outputs where one is pixel-wise
classification between text and non-text, and the other is direct regression to
determine the vertex coordinates of quadrilateral text boundaries. The proposed
method is particularly beneficial for localizing incidental scene texts. On the
ICDAR2015 Incidental Scene Text benchmark, our method achieves the F1-measure
of 81%, which is a new state-of-the-art and significantly outperforms previous
approaches. On other standard datasets with focused scene texts, our method
also reaches the state-of-the-art performance. | [
"cs.CV"
] |
In order to address the issue that medical image would suffer from severe
blurring caused by the lack of high-frequency details in the process of image
super-resolution reconstruction, a novel medical image super-resolution method
based on dense neural network and blended attention mechanism is proposed. The
proposed method adds blended attention blocks to dense neural
network(DenseNet), so that the neural network can concentrate more attention to
the regions and channels with sufficient high-frequency details. Batch
normalization layers are removed to avoid loss of high-frequency texture
details. Final obtained high resolution medical image are obtained using
deconvolutional layers at the very end of the network as up-sampling operators.
Experimental results show that the proposed method has an improvement of 0.05db
to 11.25dB and 0.6% to 14.04% on the peak signal-to-noise ratio(PSNR) metric
and structural similarity index(SSIM) metric, respectively, compared with the
mainstream image super-resolution methods. This work provides a new idea for
theoretical studies of medical image super-resolution reconstruction. | [
"cs.CV",
"physics.med-ph"
] |
Recent years have witnessed the tremendous research interests in network
embedding. Extant works have taken the neighborhood formation as the critical
information to reveal the inherent dynamics of network structures, and
suggested encoding temporal edge formation sequences to capture the historical
influences of neighbors. In this paper, however, we argue that the edge
formation can be attributed to a variety of driving factors including the
temporal influence, which is better referred to as multiple aspects. As a
matter of fact, different node aspects can drive the formation of distinctive
neighbors, giving birth to the multi-aspect embedding that relates to but goes
beyond a temporal scope. Along this vein, we propose a Mixture of Hawkes-based
Temporal Network Embeddings (MHNE) model to capture the aspect-driven
neighborhood formation of networks. In MHNE, we encode the multi-aspect
embeddings into the mixture of Hawkes processes to gain the advantages in
modeling the excitation effects and the latent aspects. Specifically, a graph
attention mechanism is used to assign different weights to account for the
excitation effects of history events, while a Gumbel-Softmax is plugged in to
derive the distribution over the aspects. Extensive experiments on 8 different
temporal networks have demonstrated the great performance of the multi-aspect
embeddings obtained by MHNE in comparison with the state-of-the-art methods. | [
"cs.LG"
] |
With the widespread success of deep learning in biomedical image
segmentation, domain shift becomes a critical and challenging problem, as the
gap between two domains can severely affect model performance when deployed to
unseen data with heterogeneous features. To alleviate this problem, we present
a novel unsupervised domain adaptation network, for generalizing models learned
from the labeled source domain to the unlabeled target domain for
cross-modality biomedical image segmentation. Specifically, our approach
consists of two key modules, a conditional domain discriminator~(CDD) and a
category-centric prototype aligner~(CCPA). The CDD, extended from conditional
domain adversarial networks in classifier tasks, is effective and robust in
handling complex cross-modality biomedical images. The CCPA, improved from the
graph-induced prototype alignment mechanism in cross-domain object detection,
can exploit precise instance-level features through an elaborate prototype
representation. In addition, it can address the negative effect of class
imbalance via entropy-based loss. Extensive experiments on a public benchmark
for the cardiac substructure segmentation task demonstrate that our method
significantly improves performance on the target domain. | [
"cs.CV"
] |
Solving the visual symbol grounding problem has long been a goal of
artificial intelligence. The field appears to be advancing closer to this goal
with recent breakthroughs in deep learning for natural language grounding in
static images. In this paper, we propose to translate videos directly to
sentences using a unified deep neural network with both convolutional and
recurrent structure. Described video datasets are scarce, and most existing
methods have been applied to toy domains with a small vocabulary of possible
words. By transferring knowledge from 1.2M+ images with category labels and
100,000+ images with captions, our method is able to create sentence
descriptions of open-domain videos with large vocabularies. We compare our
approach with recent work using language generation metrics, subject, verb, and
object prediction accuracy, and a human evaluation. | [
"cs.CV",
"cs.CL"
] |
Bird species classification has received more and more attention in the field
of computer vision, for its promising applications in biology and environmental
studies. Recognizing bird species is difficult due to the challenges of
discriminative region localization and fine-grained feature learning. In this
paper, we have introduced a Transfer learning based method with multistage
training. We have used both Pre-Trained Mask-RCNN and an ensemble model
consisting of Inception Nets (InceptionV3 & InceptionResNetV2 ) to get
localization and species of the bird from the images respectively. Our final
model achieves an F1 score of 0.5567 or 55.67 % on the dataset provided in CVIP
2018 Challenge. | [
"cs.CV",
"cs.LG"
] |
Self-supervised learning of graph neural networks (GNN) is in great need
because of the widespread label scarcity issue in real-world graph/network
data. Graph contrastive learning (GCL), by training GNNs to maximize the
correspondence between the representations of the same graph in its different
augmented forms, may yield robust and transferable GNNs even without using
labels. However, GNNs trained by traditional GCL often risk capturing redundant
graph features and thus may be brittle and provide sub-par performance in
downstream tasks. Here, we propose a novel principle, termed adversarial-GCL
(AD-GCL), which enables GNNs to avoid capturing redundant information during
the training by optimizing adversarial graph augmentation strategies used in
GCL. We pair AD-GCL with theoretical explanations and design a practical
instantiation based on trainable edge-dropping graph augmentation. We
experimentally validate AD-GCL by comparing with the state-of-the-art GCL
methods and achieve performance gains of up-to $14\%$ in unsupervised, $6\%$ in
transfer, and $3\%$ in semi-supervised learning settings overall with 18
different benchmark datasets for the tasks of molecule property regression and
classification, and social network classification. | [
"cs.LG",
"cs.AI"
] |
Remote sensing image fusion (also known as pan-sharpening) aims at generating
high resolution multi-spectral (MS) image from inputs of a high spatial
resolution single band panchromatic (PAN) image and a low spatial resolution
multi-spectral image. Inspired by the astounding achievements of convolutional
neural networks (CNNs) in a variety of computer vision tasks, in this paper, we
propose a two-stream fusion network (TFNet) to address the problem of
pan-sharpening. Unlike previous CNN based methods that consider pan-sharpening
as a super resolution problem and perform pan-sharpening in pixel level, the
proposed TFNet aims to fuse PAN and MS images in feature level and reconstruct
the pan-sharpened image from the fused features. The TFNet mainly consists of
three parts. The first part is comprised of two networks extracting features
from PAN and MS images, respectively. The subsequent network fuses them
together to form compact features that represent both spatial and spectral
information of PAN and MS images, simultaneously. Finally, the desired high
spatial resolution MS image is recovered from the fused features through an
image reconstruction network. Experiments on Quickbird and \mbox{GaoFen-1}
satellite images demonstrate that the proposed TFNet can fuse PAN and MS
images, effectively, and produce pan-sharpened images competitive with even
superior to state of the arts. | [
"cs.CV"
] |
Metric learning has the aim to improve classification accuracy by learning a
distance measure which brings data points from the same class closer together
and pushes data points from different classes further apart. Recent research
has demonstrated that metric learning approaches can also be applied to trees,
such as molecular structures, abstract syntax trees of computer programs, or
syntax trees of natural language, by learning the cost function of an edit
distance, i.e. the costs of replacing, deleting, or inserting nodes in a tree.
However, learning such costs directly may yield an edit distance which violates
metric axioms, is challenging to interpret, and may not generalize well. In
this contribution, we propose a novel metric learning approach for trees which
we call embedding edit distance learning (BEDL) and which learns an edit
distance indirectly by embedding the tree nodes as vectors, such that the
Euclidean distance between those vectors supports class discrimination. We
learn such embeddings by reducing the distance to prototypical trees from the
same class and increasing the distance to prototypical trees from different
classes. In our experiments, we show that BEDL improves upon the
state-of-the-art in metric learning for trees on six benchmark data sets,
ranging from computer science over biomedical data to a natural-language
processing data set containing over 300,000 nodes. | [
"cs.LG",
"stat.ML"
] |
This paper describes the details of Sighthound's fully automated vehicle
make, model and color recognition system. The backbone of our system is a deep
convolutional neural network that is not only computationally inexpensive, but
also provides state-of-the-art results on several competitive benchmarks.
Additionally, our deep network is trained on a large dataset of several million
images which are labeled through a semi-automated process. Finally we test our
system on several public datasets as well as our own internal test dataset. Our
results show that we outperform other methods on all benchmarks by significant
margins. Our model is available to developers through the Sighthound Cloud API
at https://www.sighthound.com/products/cloud | [
"cs.CV",
"cs.AI"
] |
We propose Skip-Convolutions to leverage the large amount of redundancies in
video streams and save computations. Each video is represented as a series of
changes across frames and network activations, denoted as residuals. We
reformulate standard convolution to be efficiently computed on residual frames:
each layer is coupled with a binary gate deciding whether a residual is
important to the model prediction,~\eg foreground regions, or it can be safely
skipped, e.g. background regions. These gates can either be implemented as an
efficient network trained jointly with convolution kernels, or can simply skip
the residuals based on their magnitude. Gating functions can also incorporate
block-wise sparsity structures, as required for efficient implementation on
hardware platforms. By replacing all convolutions with Skip-Convolutions in two
state-of-the-art architectures, namely EfficientDet and HRNet, we reduce their
computational cost consistently by a factor of 3~4x for two different tasks,
without any accuracy drop. Extensive comparisons with existing model
compression, as well as image and video efficiency methods demonstrate that
Skip-Convolutions set a new state-of-the-art by effectively exploiting the
temporal redundancies in videos. | [
"cs.CV",
"cs.LG"
] |
Respiratory rate (RR) is a clinical sign representing ventilation. An
abnormal change in RR is often the first sign of health deterioration as the
body attempts to maintain oxygen delivery to its tissues. There has been a
growing interest in remotely monitoring of RR in everyday settings which has
made photoplethysmography (PPG) monitoring wearable devices an attractive
choice. PPG signals are useful sources for RR extraction due to the presence of
respiration-induced modulations in them. The existing PPG-based RR estimation
methods mainly rely on hand-crafted rules and manual parameters tuning. An
end-to-end deep learning approach was recently proposed, however, despite its
automatic nature, the performance of this method is not ideal using the real
world data. In this paper, we present an end-to-end and accurate pipeline for
RR estimation using Cycle Generative Adversarial Networks (CycleGAN) to
reconstruct respiratory signals from raw PPG signals. Our results demonstrate a
higher RR estimation accuracy of up to 2$\times$ (mean absolute error of
1.9$\pm$0.3 using five fold cross validation) compared to the state-of-th-art
using a identical publicly available dataset. Our results suggest that CycleGAN
can be a valuable method for RR estimation from raw PPG signals. | [
"cs.LG",
"eess.SP"
] |
Moving Object Detection (MOD) is an important task for achieving robust
autonomous driving. An autonomous vehicle has to estimate collision risk with
other interacting objects in the environment and calculate an optional
trajectory. Collision risk is typically higher for moving objects than static
ones due to the need to estimate the future states and poses of the objects for
decision making. This is particularly important for near-range objects around
the vehicle which are typically detected by a fisheye surround-view system that
captures a 360{\deg} view of the scene. In this work, we propose a CNN
architecture for moving object detection using fisheye images that were
captured in autonomous driving environment. As motion geometry is highly
non-linear and unique for fisheye cameras, we will make an improved version of
the current dataset public to encourage further research. To target embedded
deployment, we design a lightweight encoder sharing weights across sequential
images. The proposed network runs at 15 fps on a 1 teraflops automotive
embedded system at accuracy of 40% IoU and 69.5% mIoU. | [
"cs.CV",
"eess.IV"
] |
Generative adversarial networks have been able to generate striking results
in various domains. This generation capability can be general while the
networks gain deep understanding regarding the data distribution. In many
domains, this data distribution consists of anomalies and normal data, with the
anomalies commonly occurring relatively less, creating datasets that are
imbalanced. The capabilities that generative adversarial networks offer can be
leveraged to examine these anomalies and help alleviate the challenge that
imbalanced datasets propose via creating synthetic anomalies. This anomaly
generation can be specifically beneficial in domains that have costly data
creation processes as well as inherently imbalanced datasets. One of the
domains that fits this description is the host-based intrusion detection
domain. In this work, ADFA-LD dataset is chosen as the dataset of interest
containing system calls of small foot-print next generation attacks. The data
is first converted into images, and then a Cycle-GAN is used to create images
of anomalous data from images of normal data. The generated data is combined
with the original dataset and is used to train a model to detect anomalies. By
doing so, it is shown that the classification results are improved, with the
AUC rising from 0.55 to 0.71, and the anomaly detection rate rising from 17.07%
to 80.49%. The results are also compared to SMOTE, showing the potential
presented by generative adversarial networks in anomaly generation. | [
"cs.LG",
"cs.CR",
"cs.CV",
"stat.ML"
] |
Cross-domain sentiment classification (CDSC) is an importance task in domain
adaptation and sentiment classification. Due to the domain discrepancy, a
sentiment classifier trained on source domain data may not works well on target
domain data. In recent years, many researchers have used deep neural network
models for cross-domain sentiment classification task, many of which use
Gradient Reversal Layer (GRL) to design an adversarial network structure to
train a domain-shared sentiment classifier. Different from those methods, we
proposed Hierarchical Attention Generative Adversarial Networks (HAGAN) which
alternately trains a generator and a discriminator in order to produce a
document representation which is sentiment-distinguishable but
domain-indistinguishable. Besides, the HAGAN model applies Bidirectional Gated
Recurrent Unit (Bi-GRU) to encode the contextual information of a word and a
sentence into the document representation. In addition, the HAGAN model use
hierarchical attention mechanism to optimize the document representation and
automatically capture the pivots and non-pivots. The experiments on Amazon
review dataset show the effectiveness of HAGAN. | [
"cs.LG",
"stat.ML"
] |
Recurrent graph convolutional neural networks are highly effective machine
learning techniques for spatiotemporal signal processing. Newly proposed graph
neural network architectures are repetitively evaluated on standard tasks such
as traffic or weather forecasting. In this paper, we propose the Chickenpox
Cases in Hungary dataset as a new dataset for comparing graph neural network
architectures. Our time series analysis and forecasting experiments demonstrate
that the Chickenpox Cases in Hungary dataset is adequate for comparing the
predictive performance and forecasting capabilities of novel recurrent graph
neural network architectures. | [
"cs.LG",
"cs.AI"
] |
Inverse reinforcement learning has proved its ability to explain state-action
trajectories of expert agents by recovering their underlying reward functions
in increasingly challenging environments. Recent advances in adversarial
learning have allowed extending inverse RL to applications with non-stationary
environment dynamics unknown to the agents, arbitrary structures of reward
functions and improved handling of the ambiguities inherent to the ill-posed
nature of inverse RL. This is particularly relevant in real time applications
on stochastic environments involving risk, like volatile financial markets.
Moreover, recent work on simulation of complex environments enable learning
algorithms to engage with real market data through simulations of its latent
space representations, avoiding a costly exploration of the original
environment. In this paper, we explore whether adversarial inverse RL
algorithms can be adapted and trained within such latent space simulations from
real market data, while maintaining their ability to recover agent rewards
robust to variations in the underlying dynamics, and transfer them to new
regimes of the original environment. | [
"cs.LG",
"q-fin.TR",
"stat.ML"
] |
To learn good joint policies for multi-agent collaboration with imperfect
information remains a fundamental challenge. While for two-player zero-sum
games, coordinate-ascent approaches (optimizing one agent's policy at a time,
e.g., self-play) work with guarantees, in multi-agent cooperative setting they
often converge to sub-optimal Nash equilibrium. On the other hand, directly
modeling joint policy changes in imperfect information game is nontrivial due
to complicated interplay of policies (e.g., upstream updates affect downstream
state reachability). In this paper, we show global changes of game values can
be decomposed to policy changes localized at each information set, with a novel
term named policy-change density. Based on this, we propose Joint Policy
Search(JPS) that iteratively improves joint policies of collaborative agents in
imperfect information games, without re-evaluating the entire game. On
multi-agent collaborative tabular games, JPS is proven to never worsen
performance and can improve solutions provided by unilateral approaches (e.g,
CFR), outperforming algorithms designed for collaborative policy learning (e.g.
BAD). Furthermore, for real-world games, JPS has an online form that naturally
links with gradient updates. We test it to Contract Bridge, a 4-player
imperfect-information game where a team of $2$ collaborates to compete against
the other. In its bidding phase, players bid in turn to find a good contract
through a limited information channel. Based on a strong baseline agent that
bids competitive bridge purely through domain-agnostic self-play, JPS improves
collaboration of team players and outperforms WBridge5, a championship-winning
software, by $+0.63$ IMPs (International Matching Points) per board over 1k
games, substantially better than previous SoTA ($+0.41$ IMPs/b) under
Double-Dummy evaluation. | [
"cs.LG",
"cs.AI",
"cs.GT",
"cs.MA",
"stat.ML"
] |
Despite their recent successes, GAN models for semantic image synthesis still
suffer from poor image quality when trained with only adversarial supervision.
Historically, additionally employing the VGG-based perceptual loss has helped
to overcome this issue, significantly improving the synthesis quality, but at
the same time limiting the progress of GAN models for semantic image synthesis.
In this work, we propose a novel, simplified GAN model, which needs only
adversarial supervision to achieve high quality results. We re-design the
discriminator as a semantic segmentation network, directly using the given
semantic label maps as the ground truth for training. By providing stronger
supervision to the discriminator as well as to the generator through spatially-
and semantically-aware discriminator feedback, we are able to synthesize images
of higher fidelity with better alignment to their input label maps, making the
use of the perceptual loss superfluous. Moreover, we enable high-quality
multi-modal image synthesis through global and local sampling of a 3D noise
tensor injected into the generator, which allows complete or partial image
change. We show that images synthesized by our model are more diverse and
follow the color and texture distributions of real images more closely. We
achieve an average improvement of $6$ FID and $5$ mIoU points over the state of
the art across different datasets using only adversarial supervision. | [
"cs.CV",
"cs.LG",
"eess.IV"
] |
Identify the cells' nuclei is the important point for most medical analyses.
To assist doctors finding the accurate cell' nuclei location automatically is
highly demanded in the clinical practice. Recently, fully convolutional neural
network (FCNs) serve as the back-bone in many image segmentation, like liver
and tumer segmentation in medical field, human body block in technical filed.
The cells' nuclei identification task is also kind of image segmentation. To
achieve this, we prefer to use deep learning algorithms. we construct three
general frameworks, one is Mask Region-based Convolutional Neural Network (Mask
RCNN), which has the high performance in many image segmentations, one is
U-net, which has the high generalization performance on small dataset and the
other is DenseUNet, which is mixture network architecture with Dense Net and
U-net. we compare the performance of these three frameworks. And we evaluated
our method on the dataset of data science bowl 2018 challenge. For single model
without any ensemble, they all have good performance. | [
"cs.CV",
"cs.LG",
"eess.IV"
] |
Label Propagation (LPA) and Graph Convolutional Neural Networks (GCN) are
both message passing algorithms on graphs. Both solve the task of node
classification but LPA propagates node label information across the edges of
the graph, while GCN propagates and transforms node feature information.
However, while conceptually similar, theoretical relation between LPA and GCN
has not yet been investigated. Here we study the relationship between LPA and
GCN in terms of two aspects: (1) feature/label smoothing where we analyze how
the feature/label of one node is spread over its neighbors; And, (2)
feature/label influence of how much the initial feature/label of one node
influences the final feature/label of another node. Based on our theoretical
analysis, we propose an end-to-end model that unifies GCN and LPA for node
classification. In our unified model, edge weights are learnable, and the LPA
serves as regularization to assist the GCN in learning proper edge weights that
lead to improved classification performance. Our model can also be seen as
learning attention weights based on node labels, which is more task-oriented
than existing feature-based attention models. In a number of experiments on
real-world graphs, our model shows superiority over state-of-the-art GCN-based
methods in terms of node classification accuracy. | [
"cs.LG",
"stat.ML"
] |
Many automotive applications, such as Advanced Driver Assistance Systems
(ADAS) for collision avoidance and warnings, require estimating the future
automotive risk of a driving scene. We present a low-cost system that predicts
the collision risk over an intermediate time horizon from a monocular video
source, such as a dashboard-mounted camera. The modular system includes
components for object detection, object tracking, and state estimation. We
introduce solutions to the object tracking and distance estimation problems.
Advanced approaches to the other tasks are used to produce real-time
predictions of the automotive risk for the next 10 s at over 5 Hz. The system
is designed such that alternative components can be substituted with minimal
effort. It is demonstrated on common physical hardware, specifically an
off-the-shelf gaming laptop and a webcam. We extend the framework to support
absolute speed estimation and more advanced risk estimation techniques. | [
"cs.CV",
"cs.RO"
] |
Classical optimal transport problem seeks a transportation map that preserves
the total mass betwenn two probability distributions, requiring their mass to
be the same. This may be too restrictive in certain applications such as color
or shape matching, since the distributions may have arbitrary masses and/or
that only a fraction of the total mass has to be transported. Several
algorithms have been devised for computing partial Wasserstein metrics that
rely on an entropic regularization, but when it comes with exact solutions,
almost no partial formulation of neither Wasserstein nor Gromov-Wasserstein are
available yet. This precludes from working with distributions that do not lie
in the same metric space or when invariance to rotation or translation is
needed. In this paper, we address the partial Wasserstein and
Gromov-Wasserstein problems and propose exact algorithms to solve them. We
showcase the new formulation in a positive-unlabeled (PU) learning application.
To the best of our knowledge, this is the first application of optimal
transport in this context and we first highlight that partial Wasserstein-based
metrics prove effective in usual PU learning settings. We then demonstrate that
partial Gromov-Wasserstein metrics is efficient in scenario where point clouds
come from different domains or have different features. | [
"stat.ML",
"cs.LG"
] |
Learning to produce efficient movement behaviour for humanoid robots from
scratch is a hard problem, as has been illustrated by the "Learning to run"
competition at NIPS 2017. The goal of this competition was to train a
two-legged model of a humanoid body to run in a simulated race course with
maximum speed. All submissions took a tabula rasa approach to reinforcement
learning (RL) and were able to produce relatively fast, but not optimal running
behaviour. In this paper, we demonstrate how data from videos of human running
(e.g. taken from YouTube) can be used to shape the reward of the humanoid
learning agent to speed up the learning and produce a better result.
Specifically, we are using the positions of key body parts at regular time
intervals to define a potential function for potential-based reward shaping
(PBRS). Since PBRS does not change the optimal policy, this approach allows the
RL agent to overcome sub-optimalities in the human movements that are shown in
the videos.
We present experiments in which we combine selected techniques from the top
ten approaches from the NIPS competition with further optimizations to create
an high-performing agent as a baseline. We then demonstrate how video-based
reward shaping improves the performance further, resulting in an RL agent that
runs twice as fast as the baseline in 12 hours of training. We furthermore show
that our approach can overcome sub-optimal running behaviour in videos, with
the learned policy significantly outperforming that of the running agent from
the video. | [
"cs.LG",
"cs.CV",
"cs.RO"
] |
We propose an efficient method for non-rigid surface tracking from monocular
RGB videos. Given a video and a template mesh, our algorithm sequentially
registers the template non-rigidly to each frame. We formulate the per-frame
registration as an optimization problem that includes a novel texture term
specifically tailored towards tracking objects with uniform texture but
fine-scale structure, such as the regular micro-structural patterns of fabric.
Our texture term exploits the orientation information in the micro-structures
of the objects, e.g., the yarn patterns of fabrics. This enables us to
accurately track uniformly colored materials that have these high frequency
micro-structures, for which traditional photometric terms are usually less
effective. The results demonstrate the effectiveness of our method on both
general textured non-rigid objects and monochromatic fabrics. | [
"cs.CV"
] |
Robust visual tracking is a challenging computer vision problem, with many
real-world applications. Most existing approaches employ hand-crafted
appearance features, such as HOG or Color Names. Recently, deep RGB features
extracted from convolutional neural networks have been successfully applied for
tracking. Despite their success, these features only capture appearance
information. On the other hand, motion cues provide discriminative and
complementary information that can improve tracking performance. Contrary to
visual tracking, deep motion features have been successfully applied for action
recognition and video classification tasks. Typically, the motion features are
learned by training a CNN on optical flow images extracted from large amounts
of labeled videos.
This paper presents an investigation of the impact of deep motion features in
a tracking-by-detection framework. We further show that hand-crafted, deep RGB,
and deep motion features contain complementary information. To the best of our
knowledge, we are the first to propose fusing appearance information with deep
motion features for visual tracking. Comprehensive experiments clearly suggest
that our fusion approach with deep motion features outperforms standard methods
relying on appearance information alone. | [
"cs.CV"
] |
Deep neural networks (DNNs) are vulnerable to adversarial examples with small
perturbations. Adversarial defense thus has been an important means which
improves the robustness of DNNs by defending against adversarial examples.
Existing defense methods focus on some specific types of adversarial examples
and may fail to defend well in real-world applications. In practice, we may
face many types of attacks where the exact type of adversarial examples in
real-world applications can be even unknown. In this paper, motivated by that
adversarial examples are more likely to appear near the classification
boundary, we study adversarial examples from a new perspective that whether we
can defend against adversarial examples by pulling them back to the original
clean distribution. We theoretically and empirically verify the existence of
defense affine transformations that restore adversarial examples. Relying on
this, we learn a defense transformer to counterattack the adversarial examples
by parameterizing the affine transformations and exploiting the boundary
information of DNNs. Extensive experiments on both toy and real-world datasets
demonstrate the effectiveness and generalization of our defense transformer. | [
"cs.LG",
"cs.CR"
] |
We present a new approach to modeling sequential data: the deep equilibrium
model (DEQ). Motivated by an observation that the hidden layers of many
existing deep sequence models converge towards some fixed point, we propose the
DEQ approach that directly finds these equilibrium points via root-finding.
Such a method is equivalent to running an infinite depth (weight-tied)
feedforward network, but has the notable advantage that we can analytically
backpropagate through the equilibrium point using implicit differentiation.
Using this approach, training and prediction in these networks require only
constant memory, regardless of the effective "depth" of the network. We
demonstrate how DEQs can be applied to two state-of-the-art deep sequence
models: self-attention transformers and trellis networks. On large-scale
language modeling tasks, such as the WikiText-103 benchmark, we show that DEQs
1) often improve performance over these state-of-the-art models (for similar
parameter counts); 2) have similar computational requirements to existing
models; and 3) vastly reduce memory consumption (often the bottleneck for
training large sequence models), demonstrating an up-to 88% memory reduction in
our experiments. The code is available at https://github.com/locuslab/deq . | [
"cs.LG",
"stat.ML"
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.