July 30, 2019

3085 words 15 mins read

Paper Group AWR 59

Tracking for Half an Hour. Enhancing Network Embedding with Auxiliary Information: An Explicit Matrix Factorization Perspective. SuperPoint: Self-Supervised Interest Point Detection and Description. A Low-Cost Ethics Shaping Approach for Designing Reinforcement Learning Agents. DropIn: Making Reservoir Computing Neural Networks Robust to Missing In …

Tracking for Half an Hour


Title	Tracking for Half an Hour
Authors	Ran Tao, Efstratios Gavves, Arnold W. M. Smeulders
Abstract	Long-term tracking requires extreme stability to the multitude of model updates and robustness to the disappearance and loss of the target as such will inevitably happen. For motivation, we have taken 10 randomly selected OTB-sequences, doubled each by attaching a reversed version and repeated each double sequence 20 times. On most of these repetitive videos, the best current tracker performs worse on each loop. This illustrates the difference between optimization for short-term versus long-term tracking. In a long-term tracker a combined global and local search strategy is beneficial, allowing for recovery from failures and disappearance. Most importantly, the proposed tracker also employs cautious updating, guided by self-quality assessment. The proposed tracker is still among the best on the 20-sec OTB-videos while achieving state-of-the-art on the 100-sec UAV20L benchmark. On 10 new half-an-hour videos with city bicycling, sport games etc, the proposed tracker outperforms others by a large margin where the 2010 TLD tracker comes second.
Tasks
Published	2017-11-28
URL	http://arxiv.org/abs/1711.10217v1
PDF	http://arxiv.org/pdf/1711.10217v1.pdf
PWC	https://paperswithcode.com/paper/tracking-for-half-an-hour
Repo	https://github.com/QUVA-Lab/Long-term-Siamese-Tracker
Framework	pytorch

Enhancing Network Embedding with Auxiliary Information: An Explicit Matrix Factorization Perspective


Title	Enhancing Network Embedding with Auxiliary Information: An Explicit Matrix Factorization Perspective
Authors	Junliang Guo, Linli Xu, Xunpeng Huang, Enhong Chen
Abstract	Recent advances in the field of network embedding have shown the low-dimensional network representation is playing a critical role in network analysis. However, most of the existing principles of network embedding do not incorporate auxiliary information such as content and labels of nodes flexibly. In this paper, we take a matrix factorization perspective of network embedding, and incorporate structure, content and label information of the network simultaneously. For structure, we validate that the matrix we construct preserves high-order proximities of the network. Label information can be further integrated into the matrix via the process of random walk sampling to enhance the quality of embedding in an unsupervised manner, i.e., without leveraging downstream classifiers. In addition, we generalize the Skip-Gram Negative Sampling model to integrate the content of the network in a matrix factorization framework. As a consequence, network embedding can be learned in a unified framework integrating network structure and node content as well as label information simultaneously. We demonstrate the efficacy of the proposed model with the tasks of semi-supervised node classification and link prediction on a variety of real-world benchmark network datasets.
Tasks	Link Prediction, Network Embedding, Node Classification
Published	2017-11-11
URL	http://arxiv.org/abs/1711.04094v2
PDF	http://arxiv.org/pdf/1711.04094v2.pdf
PWC	https://paperswithcode.com/paper/enhancing-network-embedding-with-auxiliary
Repo	https://github.com/lemmonation/APNE
Framework	none

SuperPoint: Self-Supervised Interest Point Detection and Description


Title	SuperPoint: Self-Supervised Interest Point Detection and Description
Authors	Daniel DeTone, Tomasz Malisiewicz, Andrew Rabinovich
Abstract	This paper presents a self-supervised framework for training interest point detectors and descriptors suitable for a large number of multiple-view geometry problems in computer vision. As opposed to patch-based neural networks, our fully-convolutional model operates on full-sized images and jointly computes pixel-level interest point locations and associated descriptors in one forward pass. We introduce Homographic Adaptation, a multi-scale, multi-homography approach for boosting interest point detection repeatability and performing cross-domain adaptation (e.g., synthetic-to-real). Our model, when trained on the MS-COCO generic image dataset using Homographic Adaptation, is able to repeatedly detect a much richer set of interest points than the initial pre-adapted deep model and any other traditional corner detector. The final system gives rise to state-of-the-art homography estimation results on HPatches when compared to LIFT, SIFT and ORB.
Tasks	Domain Adaptation, Homography Estimation, Interest Point Detection
Published	2017-12-20
URL	http://arxiv.org/abs/1712.07629v4
PDF	http://arxiv.org/pdf/1712.07629v4.pdf
PWC	https://paperswithcode.com/paper/superpoint-self-supervised-interest-point
Repo	https://github.com/magicleap/SuperGluePretrainedNetwork
Framework	pytorch

A Low-Cost Ethics Shaping Approach for Designing Reinforcement Learning Agents


Title	A Low-Cost Ethics Shaping Approach for Designing Reinforcement Learning Agents
Authors	Yueh-Hua Wu, Shou-De Lin
Abstract	This paper proposes a low-cost, easily realizable strategy to equip a reinforcement learning (RL) agent the capability of behaving ethically. Our model allows the designers of RL agents to solely focus on the task to achieve, without having to worry about the implementation of multiple trivial ethical patterns to follow. Based on the assumption that the majority of human behavior, regardless which goals they are achieving, is ethical, our design integrates human policy with the RL policy to achieve the target objective with less chance of violating the ethical code that human beings normally obey.
Tasks
Published	2017-12-12
URL	http://arxiv.org/abs/1712.04172v2
PDF	http://arxiv.org/pdf/1712.04172v2.pdf
PWC	https://paperswithcode.com/paper/a-low-cost-ethics-shaping-approach-for
Repo	https://github.com/kristery/EthicsShaping
Framework	none

DropIn: Making Reservoir Computing Neural Networks Robust to Missing Inputs by Dropout


Title	DropIn: Making Reservoir Computing Neural Networks Robust to Missing Inputs by Dropout
Authors	Davide Bacciu, Francesco Crecchi, Davide Morelli
Abstract	The paper presents a novel, principled approach to train recurrent neural networks from the Reservoir Computing family that are robust to missing part of the input features at prediction time. By building on the ensembling properties of Dropout regularization, we propose a methodology, named DropIn, which efficiently trains a neural model as a committee machine of subnetworks, each capable of predicting with a subset of the original input features. We discuss the application of the DropIn methodology in the context of Reservoir Computing models and targeting applications characterized by input sources that are unreliable or prone to be disconnected, such as in pervasive wireless sensor networks and ambient intelligence. We provide an experimental assessment using real-world data from such application domains, showing how the Dropin methodology allows to maintain predictive performances comparable to those of a model without missing features, even when 20%-50% of the inputs are not available.
Tasks
Published	2017-05-07
URL	http://arxiv.org/abs/1705.02643v1
PDF	http://arxiv.org/pdf/1705.02643v1.pdf
PWC	https://paperswithcode.com/paper/dropin-making-reservoir-computing-neural
Repo	https://github.com/FrancescoCrecchi/DropIn-ESN
Framework	none

Can Spatiotemporal 3D CNNs Retrace the History of 2D CNNs and ImageNet?


Title	Can Spatiotemporal 3D CNNs Retrace the History of 2D CNNs and ImageNet?
Authors	Kensho Hara, Hirokatsu Kataoka, Yutaka Satoh
Abstract	The purpose of this study is to determine whether current video datasets have sufficient data for training very deep convolutional neural networks (CNNs) with spatio-temporal three-dimensional (3D) kernels. Recently, the performance levels of 3D CNNs in the field of action recognition have improved significantly. However, to date, conventional research has only explored relatively shallow 3D architectures. We examine the architectures of various 3D CNNs from relatively shallow to very deep ones on current video datasets. Based on the results of those experiments, the following conclusions could be obtained: (i) ResNet-18 training resulted in significant overfitting for UCF-101, HMDB-51, and ActivityNet but not for Kinetics. (ii) The Kinetics dataset has sufficient data for training of deep 3D CNNs, and enables training of up to 152 ResNets layers, interestingly similar to 2D ResNets on ImageNet. ResNeXt-101 achieved 78.4% average accuracy on the Kinetics test set. (iii) Kinetics pretrained simple 3D architectures outperforms complex 2D architectures, and the pretrained ResNeXt-101 achieved 94.5% and 70.2% on UCF-101 and HMDB-51, respectively. The use of 2D CNNs trained on ImageNet has produced significant progress in various tasks in image. We believe that using deep 3D CNNs together with Kinetics will retrace the successful history of 2D CNNs and ImageNet, and stimulate advances in computer vision for videos. The codes and pretrained models used in this study are publicly available. https://github.com/kenshohara/3D-ResNets-PyTorch
Tasks	Action Recognition In Videos
Published	2017-11-27
URL	http://arxiv.org/abs/1711.09577v2
PDF	http://arxiv.org/pdf/1711.09577v2.pdf
PWC	https://paperswithcode.com/paper/can-spatiotemporal-3d-cnns-retrace-the
Repo	https://github.com/ptmcgrat/3D-ResNets-PyTorch
Framework	pytorch

An Adaptive Genetic Algorithm for Solving N-Queens Problem


Title	An Adaptive Genetic Algorithm for Solving N-Queens Problem
Authors	Uddalok Sarkar, Sayan Nag
Abstract	In this paper a Metaheuristic approach for solving the N-Queens Problem is introduced to find the best possible solution in a reasonable amount of time. Genetic Algorithm is used with a novel fitness function as the Metaheuristic. The aim of N-Queens Problem is to place N queens on an N x N chessboard, in a way so that no queen is in conflict with the others. Chromosome representation and genetic operations like Mutation and Crossover are described in detail. Results show that this approach yields promising and satisfactory results in less time compared to that obtained from the previous approaches for several large values of N.
Tasks
Published	2017-11-01
URL	http://arxiv.org/abs/1802.02006v1
PDF	http://arxiv.org/pdf/1802.02006v1.pdf
PWC	https://paperswithcode.com/paper/an-adaptive-genetic-algorithm-for-solving-n
Repo	https://github.com/depanker/ml-problems
Framework	none

Lasso Regularization Paths for NARMAX Models via Coordinate Descent


Title	Lasso Regularization Paths for NARMAX Models via Coordinate Descent
Authors	Antônio H. Ribeiro, Luis A. Aguirre
Abstract	We propose a new algorithm for estimating NARMAX models with $L_1$ regularization for models represented as a linear combination of basis functions. Due to the $L_1$-norm penalty the Lasso estimation tends to produce some coefficients that are exactly zero and hence gives interpretable models. The novelty of the contribution is the inclusion of error regressors in the Lasso estimation (which yields a nonlinear regression problem). The proposed algorithm uses cyclical coordinate descent to compute the parameters of the NARMAX models for the entire regularization path. It deals with the error terms by updating the regressor matrix along with the parameter vector. In comparative timings we find that the modification does not reduce the computational efficiency of the original algorithm and can provide the most important regressors in very few inexpensive iterations. The method is illustrated for linear and polynomial models by means of two examples.
Tasks
Published	2017-10-02
URL	http://arxiv.org/abs/1710.00598v2
PDF	http://arxiv.org/pdf/1710.00598v2.pdf
PWC	https://paperswithcode.com/paper/lasso-regularization-paths-for-narmax-models
Repo	https://github.com/antonior92/NarmaxLasso.jl
Framework	none

Parsing Universal Dependencies without training


Title	Parsing Universal Dependencies without training
Authors	Héctor Martínez Alonso, Željko Agić, Barbara Plank, Anders Søgaard
Abstract	We propose UDP, the first training-free parser for Universal Dependencies (UD). Our algorithm is based on PageRank and a small set of head attachment rules. It features two-step decoding to guarantee that function words are attached as leaf nodes. The parser requires no training, and it is competitive with a delexicalized transfer system. UDP offers a linguistically sound unsupervised alternative to cross-lingual parsing for UD, which can be used as a baseline for such systems. The parser has very few parameters and is distinctly robust to domain change across languages.
Tasks
Published	2017-01-11
URL	http://arxiv.org/abs/1701.03163v1
PDF	http://arxiv.org/pdf/1701.03163v1.pdf
PWC	https://paperswithcode.com/paper/parsing-universal-dependencies-without
Repo	https://github.com/hectormartinez/ud_unsup_parser
Framework	none

Code Attention: Translating Code to Comments by Exploiting Domain Features


Title	Code Attention: Translating Code to Comments by Exploiting Domain Features
Authors	Wenhao Zheng, Hong-Yu Zhou, Ming Li, Jianxin Wu
Abstract	Appropriate comments of code snippets provide insight for code functionality, which are helpful for program comprehension. However, due to the great cost of authoring with the comments, many code projects do not contain adequate comments. Automatic comment generation techniques have been proposed to generate comments from pieces of code in order to alleviate the human efforts in annotating the code. Most existing approaches attempt to exploit certain correlations (usually manually given) between code and generated comments, which could be easily violated if the coding patterns change and hence the performance of comment generation declines. In this paper, we first build C2CGit, a large dataset from open projects in GitHub, which is more than 20$\times$ larger than existing datasets. Then we propose a new attention module called Code Attention to translate code to comments, which is able to utilize the domain features of code snippets, such as symbols and identifiers. We make ablation studies to determine effects of different parts in Code Attention. Experimental results demonstrate that the proposed module has better performance over existing approaches in both BLEU and METEOR.
Tasks
Published	2017-09-22
URL	http://arxiv.org/abs/1709.07642v2
PDF	http://arxiv.org/pdf/1709.07642v2.pdf
PWC	https://paperswithcode.com/paper/code-attention-translating-code-to-comments
Repo	https://github.com/mf1832146/tree-transformer
Framework	none

Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset


Title	Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset
Authors	Joao Carreira, Andrew Zisserman
Abstract	The paucity of videos in current action classification datasets (UCF-101 and HMDB-51) has made it difficult to identify good video architectures, as most methods obtain similar performance on existing small-scale benchmarks. This paper re-evaluates state-of-the-art architectures in light of the new Kinetics Human Action Video dataset. Kinetics has two orders of magnitude more data, with 400 human action classes and over 400 clips per class, and is collected from realistic, challenging YouTube videos. We provide an analysis on how current architectures fare on the task of action classification on this dataset and how much performance improves on the smaller benchmark datasets after pre-training on Kinetics. We also introduce a new Two-Stream Inflated 3D ConvNet (I3D) that is based on 2D ConvNet inflation: filters and pooling kernels of very deep image classification ConvNets are expanded into 3D, making it possible to learn seamless spatio-temporal feature extractors from video while leveraging successful ImageNet architecture designs and even their parameters. We show that, after pre-training on Kinetics, I3D models considerably improve upon the state-of-the-art in action classification, reaching 80.9% on HMDB-51 and 98.0% on UCF-101.
Tasks	Action Classification, Action Recognition In Videos, Skeleton Based Action Recognition
Published	2017-05-22
URL	http://arxiv.org/abs/1705.07750v3
PDF	http://arxiv.org/pdf/1705.07750v3.pdf
PWC	https://paperswithcode.com/paper/quo-vadis-action-recognition-a-new-model-and
Repo	https://github.com/ahsaniqbal/Kinetics-FeatureExtractor
Framework	tf

Enhancing SDO/HMI images using deep learning


Title	Enhancing SDO/HMI images using deep learning
Authors	C. J. Diaz Baso, A. Asensio Ramos
Abstract	The Helioseismic and Magnetic Imager (HMI) provides continuum images and magnetograms with a cadence better than one per minute. It has been continuously observing the Sun 24 hours a day for the past 7 years. The obvious trade-off between full disk observations and spatial resolution makes HMI not enough to analyze the smallest-scale events in the solar atmosphere. Our aim is to develop a new method to enhance HMI data, simultaneously deconvolving and super-resolving images and magnetograms. The resulting images will mimic observations with a diffraction-limited telescope twice the diameter of HMI. Our method, which we call Enhance, is based on two deep fully convolutional neural networks that input patches of HMI observations and output deconvolved and super-resolved data. The neural networks are trained on synthetic data obtained from simulations of the emergence of solar active regions. We have obtained deconvolved and supper-resolved HMI images. To solve this ill-defined problem with infinite solutions we have used a neural network approach to add prior information from the simulations. We test Enhance against Hinode data that has been degraded to a 28 cm diameter telescope showing very good consistency. The code is open source.
Tasks
Published	2017-06-09
URL	http://arxiv.org/abs/1706.02933v2
PDF	http://arxiv.org/pdf/1706.02933v2.pdf
PWC	https://paperswithcode.com/paper/enhancing-sdohmi-images-using-deep-learning
Repo	https://github.com/cdiazbas/enhance
Framework	tf

Robust Synthetic Control


Title	Robust Synthetic Control
Authors	Muhammad Jehangir Amjad, Devavrat Shah, Dennis Shen
Abstract	We present a robust generalization of the synthetic control method for comparative case studies. Like the classical method, we present an algorithm to estimate the unobservable counterfactual of a treatment unit. A distinguishing feature of our algorithm is that of de-noising the data matrix via singular value thresholding, which renders our approach robust in multiple facets: it automatically identifies a good subset of donors, overcomes the challenges of missing data, and continues to work well in settings where covariate information may not be provided. To begin, we establish the condition under which the fundamental assumption in synthetic control-like approaches holds, i.e. when the linear relationship between the treatment unit and the donor pool prevails in both the pre- and post-intervention periods. We provide the first finite sample analysis for a broader class of models, the Latent Variable Model, in contrast to Factor Models previously considered in the literature. Further, we show that our de-noising procedure accurately imputes missing entries, producing a consistent estimator of the underlying signal matrix provided $p = \Omega( T^{-1 + \zeta})$ for some $\zeta > 0$; here, $p$ is the fraction of observed data and $T$ is the time interval of interest. Under the same setting, we prove that the mean-squared-error (MSE) in our prediction estimation scales as $O(\sigma^2/p + 1/\sqrt{T})$, where $\sigma^2$ is the noise variance. Using a data aggregation method, we show that the MSE can be made as small as $O(T^{-1/2+\gamma})$ for any $\gamma \in (0, 1/2)$, leading to a consistent estimator. We also introduce a Bayesian framework to quantify the model uncertainty through posterior probabilities. Our experiments, using both real-world and synthetic datasets, demonstrate that our robust generalization yields an improvement over the classical synthetic control method.
Tasks
Published	2017-11-18
URL	http://arxiv.org/abs/1711.06940v1
PDF	http://arxiv.org/pdf/1711.06940v1.pdf
PWC	https://paperswithcode.com/paper/robust-synthetic-control
Repo	https://github.com/jehangiramjad/tslib
Framework	none

Implementing the Deep Q-Network


Title	Implementing the Deep Q-Network
Authors	Melrose Roderick, James MacGlashan, Stefanie Tellex
Abstract	The Deep Q-Network proposed by Mnih et al. [2015] has become a benchmark and building point for much deep reinforcement learning research. However, replicating results for complex systems is often challenging since original scientific publications are not always able to describe in detail every important parameter setting and software engineering solution. In this paper, we present results from our work reproducing the results of the DQN paper. We highlight key areas in the implementation that were not covered in great detail in the original paper to make it easier for researchers to replicate these results, including termination conditions and gradient descent algorithms. Finally, we discuss methods for improving the computational performance and provide our own implementation that is designed to work with a range of domains, and not just the original Arcade Learning Environment [Bellemare et al., 2013].
Tasks	Atari Games
Published	2017-11-20
URL	http://arxiv.org/abs/1711.07478v1
PDF	http://arxiv.org/pdf/1711.07478v1.pdf
PWC	https://paperswithcode.com/paper/implementing-the-deep-q-network
Repo	https://github.com/xgfelicia/Reinforcement-Learning
Framework	pytorch

Fraternal Dropout


Title	Fraternal Dropout
Authors	Konrad Zolna, Devansh Arpit, Dendi Suhubdy, Yoshua Bengio
Abstract	Recurrent neural networks (RNNs) are important class of architectures among neural networks useful for language modeling and sequential prediction. However, optimizing RNNs is known to be harder compared to feed-forward neural networks. A number of techniques have been proposed in literature to address this problem. In this paper we propose a simple technique called fraternal dropout that takes advantage of dropout to achieve this goal. Specifically, we propose to train two identical copies of an RNN (that share parameters) with different dropout masks while minimizing the difference between their (pre-softmax) predictions. In this way our regularization encourages the representations of RNNs to be invariant to dropout mask, thus being robust. We show that our regularization term is upper bounded by the expectation-linear dropout objective which has been shown to address the gap due to the difference between the train and inference phases of dropout. We evaluate our model and achieve state-of-the-art results in sequence modeling tasks on two benchmark datasets - Penn Treebank and Wikitext-2. We also show that our approach leads to performance improvement by a significant margin in image captioning (Microsoft COCO) and semi-supervised (CIFAR-10) tasks.
Tasks	Image Captioning, Language Modelling
Published	2017-10-31
URL	http://arxiv.org/abs/1711.00066v4
PDF	http://arxiv.org/pdf/1711.00066v4.pdf
PWC	https://paperswithcode.com/paper/fraternal-dropout
Repo	https://github.com/kondiz/fraternal-dropout
Framework	pytorch