2020.06.26.txt

==========New Papers==========
1, TITLE:       Diffusion-Weighted Magnetic Resonance Brain Images Generation with Generative Adversarial Networks and Variational Autoencoders: A Comparison Study
http://arxiv.org/abs/2006.13944
AUTHORS:        Alejandro Ungría Hirte ; Moritz Platscher ; Thomas Joyce ; Jeremy J. Heit ; Eric Tranvinh ; Christian Federau
COMMENTS:       20 pages, 5 figures, 2 tables
HIGHLIGHT:      Based on professional neuroradiologists' evaluations and diverse metrics with respect to quality and diversity of the generated synthetic brain images, we present two networks, the Introspective Variational Autoencoder and the Style-Based GAN, that qualify for data augmentation in the medical field, where information is saved in a dispatched and inhomogeneous way and access to it is in many aspects restricted.

2, TITLE:       Blacklight: Defending Black-Box Adversarial Attacks on Deep Neural Networks
http://arxiv.org/abs/2006.14042
AUTHORS:        Huiying Li ; Shawn Shan ; Emily Wenger ; Jiayun Zhang ; Haitao Zheng ; Ben Y. Zhao
HIGHLIGHT:      This paper proposes and evaluates Blacklight, a new defense against black-box adversarial attacks.

3, TITLE:       On the Nature of Programming Exercises
http://arxiv.org/abs/2006.14476
AUTHORS:        Alberto Simões ; Ricardo Queirós
HIGHLIGHT:      This paper explores different approaches on the creation of a programming exercise, starting with realizing how it is currently formalized, presented and evaluated.

4, TITLE:       Scalable Spectral Clustering with Nystrom Approximation: Practical and Theoretical Aspects
http://arxiv.org/abs/2006.14470
AUTHORS:        Farhad Pourkamali-Anaraki
COMMENTS:       14 pages, 6 figures
HIGHLIGHT:      To address the limitations, this work presents a principled spectral clustering algorithm that makes full use of the information obtained from the Nystrom method.

5, TITLE:       The variation of the sum of edge lengths in linear arrangements of trees
http://arxiv.org/abs/2006.14069
AUTHORS:        Ramon Ferrer-i-Cancho ; Carlos Gómez-Rodríguez ; Juan Luis Esteban
HIGHLIGHT:      In particular, we investigate various problems on the sum of edge lengths in trees of a fixed size: the minimum and the maximum value of the sum for specific trees, the minimum and the maximum in classes of trees (bistar trees and caterpillar trees) and finally the minimum and the maximum for any tree.

6, TITLE:       The flag manifold as a tool for analyzing and comparing data sets
http://arxiv.org/abs/2006.14086
AUTHORS:        Xiaofeng Ma ; Michael Kirby ; Chris Peterson
COMMENTS:       15 pages, 8 figures
HIGHLIGHT:      To make practical comparisons on a flag manifold, algorithms are proposed for determining the distances between points $[A], [B]$ on a flag manifold, where $A$ and $B$ are arbitrary orthogonal matrix representatives for $[A]$ and $[B]$, and for determining the initial direction of these minimal length geodesics.

7, TITLE:       Machine learning the real discriminant locus
http://arxiv.org/abs/2006.14078
AUTHORS:        Edgar A. Bernal ; Jonathan D. Hauenstein ; Dhagash Mehta ; Margaret H. Regan ; Tingting Tang
COMMENTS:       22 pages, 14 figures
HIGHLIGHT:      For multidimensional parameter spaces, this article presents a novel sampling method which carefully samples the parameter space.

8, TITLE:       Time for a Background Check! Uncovering the impact of Background Features on Deep Neural Networks
http://arxiv.org/abs/2006.14077
AUTHORS:        Vikash Sehwag ; Rajvardhan Oak ; Mung Chiang ; Prateek Mittal
COMMENTS:       6 pages, 5 figures
HIGHLIGHT:      In this paper, we investigate to what extent the increasing performance of deep neural networks is impacted by background features?

9, TITLE:       Neural Architecture Design for GPU-Efficient Networks
http://arxiv.org/abs/2006.14090
AUTHORS:        Ming Lin ; Hesen Chen ; Xiuyu Sun ; Qi Qian ; Hao Li ; Rong Jin
HIGHLIGHT:      To address this issue, we propose a general principle for designing GPU-efficient networks based on extensive empirical studies.

10, TITLE:      Smooth Adversarial Training
http://arxiv.org/abs/2006.14536
AUTHORS:        Cihang Xie ; Mingxing Tan ; Boqing Gong ; Alan Yuille ; Quoc V. Le
COMMENTS:       tech report
HIGHLIGHT:      Here we present evidence to challenge these common beliefs by a careful study about adversarial training.

11, TITLE:      Learning Reward Functions from Diverse Sources of Human Feedback: Optimally Integrating Demonstrations and Preferences
http://arxiv.org/abs/2006.14091
AUTHORS:        Erdem Bıyık ; Dylan P. Losey ; Malayandi Palan ; Nicholas C. Landolfi ; Gleb Shevchuk ; Dorsa Sadigh
COMMENTS:       19 pages, 17 figures. Submitted to The International Journal of Robotics Research (IJRR)
HIGHLIGHT:      Motivated by this general problem, we present a framework to integrate multiple sources of information, which are either passively or actively collected from human users.

12, TITLE:      Learning Source Phrase Representations for Neural Machine Translation
http://arxiv.org/abs/2006.14405
AUTHORS:        Hongfei Xu ; Josef van Genabith ; Deyi Xiong ; Qiuhui Liu ; Jingyi Zhang
HIGHLIGHT:      The Transformer translation model (Vaswani et al., 2017) based on a multi-head attention mechanism can be computed effectively in parallel and has significantly pushed forward the performance of Neural Machine Translation (NMT).

13, TITLE:      PropagationNet: Propagate Points to Curve to Learn Structure Information
http://arxiv.org/abs/2006.14308
AUTHORS:        Xiehe Huang ; Weihong Deng ; Haifeng Shen ; Xiubao Zhang ; Jieping Ye
COMMENTS:       10 pages, 8 figures, 8 tables, CVPR2020
HIGHLIGHT:      In this paper, we explore the instincts and reasons behind our two proposals, \emph{i.e}\onedot Propagation Module and Focal Wing Loss, to tackle the problem.

14, TITLE:      Tensor Programs II: Neural Tangent Kernel for Any Architecture
http://arxiv.org/abs/2006.14548
AUTHORS:        Greg Yang
COMMENTS:       11 pages of main text. 61 pages total
HIGHLIGHT:      We show that a randomly initialized neural network of *any architecture* has its Tangent Kernel converge to a deterministic limit, as the network widths tend to infinity.

15, TITLE:      Estimating Displaced Populations from Overhead
http://arxiv.org/abs/2006.14547
AUTHORS:        Armin Hadzic ; Gordon Christie ; Jeffrey Freeman ; Amber Dismer ; Stevan Bullard ; Ashley Greiner ; Nathan Jacobs ; Ryan Mukherjee
HIGHLIGHT:      We introduce a deep learning approach to perform fine-grained population estimation for displacement camps using high-resolution overhead imagery.

16, TITLE:      Lifted Disjoint Paths with Application in Multiple Object Tracking
http://arxiv.org/abs/2006.14550
AUTHORS:        Andrea Hornakova ; Roberto Henschel ; Bodo Rosenhahn ; Paul Swoboda
COMMENTS:       ICML 2020, Codebase available at https://github.com/AndreaHor/LifT_Solver
HIGHLIGHT:      We present an extension to the disjoint paths problem in which additional \emph{lifted} edges are introduced to provide path connectivity priors.

17, TITLE:      The Surprising Simplicity of the Early-Time Learning Dynamics of Neural Networks
http://arxiv.org/abs/2006.14599
AUTHORS:        Wei Hu ; Lechao Xiao ; Ben Adlam ; Jeffrey Pennington
HIGHLIGHT:      In this work, we show that these common perceptions can be completely false in the early phase of learning.

18, TITLE:      Anomaly Detection using Deep Reconstruction and Forecasting for Autonomous Systems
http://arxiv.org/abs/2006.14556
AUTHORS:        Nadarasar Bahavan ; Navaratnarajah Suman ; Sulhi Cader ; Ruwinda Ranganayake ; Damitha Seneviratne ; Vinu Maddumage ; Gershom Seneviratne ; Yasinha Supun ; Isuru Wijesiri ; Suchitha Dehigaspitiya ; Dumindu Tissera ; Chamira Edussooriya
COMMENTS:       Runners Up - IEEE Signal Processing Cup 2020
HIGHLIGHT:      We propose self-supervised deep algorithms to detect anomalies in heterogeneous autonomous systems using frontal camera video and IMU readings.

19, TITLE:      Deep Learning for Cornea Microscopy Blind Deblurring
http://arxiv.org/abs/2006.14319
AUTHORS:        Toussain Cardot ; Pilar Marxer ; Ivan Snozzi
HIGHLIGHT:      The goal of this project is to build a deep-learning solution that deblurs cornea scans, used for medical examination.

20, TITLE:      Learning compositional functions via multiplicative weight updates
http://arxiv.org/abs/2006.14560
AUTHORS:        Jeremy Bernstein ; Jiawei Zhao ; Markus Meister ; Ming-Yu Liu ; Anima Anandkumar ; Yisong Yue
HIGHLIGHT:      This paper proves that multiplicative weight updates satisfy a descent lemma tailored to compositional functions.

21, TITLE:      Dynamically Mitigating Data Discrepancy with Balanced Focal Loss for Replay Attack Detection
http://arxiv.org/abs/2006.14563
AUTHORS:        Yongqiang Dou ; Haocheng Yang ; Maolin Yang ; Yanyan Xu ; Dengfeng Ke
COMMENTS:       This work has been accepted by the 25th International Conference on Pattern Recognition (ICPR2020)
HIGHLIGHT:      In this paper, we argue that for anti-spoofing, it needs more attention for indistinguishable samples over easily-classified ones in the modeling process, to make correct discrimination a top priority.

22, TITLE:      Perfusion Quantification from Endoscopic Videos: Learning to Read Tumor Signatures
http://arxiv.org/abs/2006.14321
AUTHORS:        Sergiy Zhuk ; Jonathan P. Epperlein ; Rahul Nair ; Seshu Thirupati ; Pol Mac Aonghusa ; Ronan Cahill ; Donal O'Shea
COMMENTS:       To be published in 23rd International Conference on Medical Image Computing & Computer Assisted Intervention (MICCAI 2020)
HIGHLIGHT:      We propose a perfusion quantification method for computer-aided interpretation of subtle differences in dynamic perfusion patterns which can be used to distinguish between normal tissue and benign or malignant tumors intra-operatively in real-time by using multispectral endoscopic videos.

23, TITLE:      SOAC: The Soft Option Actor-Critic Architecture
http://arxiv.org/abs/2006.14363
AUTHORS:        Chenghao Li ; Xiaoteng Ma ; Chongjie Zhang ; Jun Yang ; Li Xia ; Qianchuan Zhao
HIGHLIGHT:      In this paper, we present a novel and stable off-policy approach that builds on the maximum entropy model to address these challenges.

24, TITLE:      Analyzing Effect of Repeated Reading on Oral Fluency and Narrative Production for Computer-Assisted Language Learning
http://arxiv.org/abs/2006.14320
AUTHORS:        Santosh Kumar Barnwal ; Uma Shanker Tiwary
COMMENTS:       5 pages, 1 figure
HIGHLIGHT:      Therefore, in this paper, we present our dataset, discuss its properties, and propose a method to assess oral fluency and narrative production for learners of English using acoustic, prosodic, lexical and syntactical characteristics.

25, TITLE:      Multimarginal Wasserstein Barycenter for Stain Normalization and Augmentation
http://arxiv.org/abs/2006.14566
AUTHORS:        Saad Nadeem ; Travis Hollmann ; Allen Tannenbaum
COMMENTS:       To appear in MICCAI 2020
HIGHLIGHT:      In this work, we present a new approach based on the multimarginal Wasserstein barycenter to normalize and augment H&E stained images given one or more references.

26, TITLE:      Explainable CNN-attention Networks (C-Attention Network) for Automated Detection of Alzheimer's Disease
http://arxiv.org/abs/2006.14135
AUTHORS:        Ning Wang ; Mingxuan Chen ; K. P. Subbalakshmi
HIGHLIGHT:      In this work, we propose three explainable deep learning architectures to automatically detect patients with Alzheimer`s disease based on their language abilities.

27, TITLE:      Discontinuous and Smooth Depth Completion with Binary Anisotropic Diffusion Tensor
http://arxiv.org/abs/2006.14374
AUTHORS:        Yasuhiro Yao ; Menandro Roxas ; Ryoichi Ishikawa ; Shingo Ando ; Jun Shimamura ; Takeshi Oishi
COMMENTS:       8 pages 6 figures
HIGHLIGHT:      We propose an unsupervised real-time dense depth completion from a sparse depth map guided by a single image.

28, TITLE:      Training Variational Networks with Multi-Domain Simulations: Speed-of-Sound Image Reconstruction
http://arxiv.org/abs/2006.14395
AUTHORS:        Melanie Bernhardt ; Valery Vishnevskiy ; Richard Rau ; Orcun Goksel
HIGHLIGHT:      In this work, we present for the first time a VN solution for a pulse-echo SoS image reconstruction problem using diverging waves with conventional transducers and single-sided tissue access.

29, TITLE:      Q-NET: A Formula for Numerical Integration of a Shallow Feed-forward Neural Network
http://arxiv.org/abs/2006.14396
AUTHORS:        Kartic Subr
COMMENTS:       11 pages (including appendix and references)
HIGHLIGHT:      We derive a formula in closed form to calculate the multidimensional integral of functions fw that are representable using a shallow feed-forward neural network with weights w and a sigmoid activation function.

30, TITLE:      Empirical Study on the Benefits of Multiobjectivization for Solving Single-Objective Problems
http://arxiv.org/abs/2006.14423
AUTHORS:        Vera Steinhoff ; Pascal Kerschke ; Christian Grimme
HIGHLIGHT:      In this paper we analyze how single-objective optimization can benefit from multiobjectivization by considering an additional objective.

31, TITLE:      DanHAR: Dual Attention Network For Multimodal Human Activity Recognition Using Wearable Sensors
http://arxiv.org/abs/2006.14435
AUTHORS:        Wenbin Gao ; Lei Zhang ; Qi Teng ; Hao Wu ; Fuhong Min ; Jun He
HIGHLIGHT:      In the paper, we propose a novel dual attention method called DanHAR, which introduces the framework of blending channel attention and temporal attention on a CNN, demonstrating superiority in improving the comprehensibility for multimodal HAR.

32, TITLE:      Plausible Reasoning about EL-Ontologies using Concept Interpolation
http://arxiv.org/abs/2006.14437
AUTHORS:        Yazmín Ibáñez-García ; Víctor Gutiérrez-Basulto ; Steven Schockaert
COMMENTS:       16 pages, 3 figures, accepted at KR 2020
HIGHLIGHT:      In this paper, we instead propose an inductive inference mechanism which is based on a clear model-theoretic semantics, and can thus be tightly integrated with standard deductive reasoning.

33, TITLE:      Fine granularity access in interactive compression of 360-degree images based on rate adaptive channel codes
http://arxiv.org/abs/2006.14239
AUTHORS:        Navid Mahmoudian Bidgoli ; Thomas Maugey ; Aline Roumy
HIGHLIGHT:      In this paper, we propose a new interactive compression scheme for omnidirectional images.

34, TITLE:      SRFlow: Learning the Super-Resolution Space with Normalizing Flow
http://arxiv.org/abs/2006.14200
AUTHORS:        Andreas Lugmayr ; Martin Danelljan ; Luc Van Gool ; Radu Timofte
HIGHLIGHT:      In this work, we therefore propose SRFlow: a normalizing flow based super-resolution method capable of learning the conditional distribution of the output given the low-resolution input.

35, TITLE:      One Thousand and One Hours: Self-driving Motion Prediction Dataset
http://arxiv.org/abs/2006.14480
AUTHORS:        John Houston ; Guido Zuidhof ; Luca Bergamini ; Yawei Ye ; Ashesh Jain ; Sammy Omari ; Vladimir Iglovikov ; Peter Ondruska
COMMENTS:       The full dataset is available at http://level5.lyft.com/
HIGHLIGHT:      We present the largest self-driving dataset for motion prediction to date, with over 1,000 hours of data.

36, TITLE:      Automatic Domain Adaptation Outperforms Manual Domain Adaptation for Predicting Financial Outcomes
http://arxiv.org/abs/2006.14209
AUTHORS:        Marina Sedinkina ; Nikolas Breitkopf ; Hinrich Schütze
COMMENTS:       Accepted at ACL2019
HIGHLIGHT:      In this paper, we automatically create sentiment dictionaries for predicting financial outcomes.

37, TITLE:      Searching towards Class-Aware Generators for Conditional Generative Adversarial Networks
http://arxiv.org/abs/2006.14208
AUTHORS:        Peng Zhou ; Lingxi Xie ; Xiaopeng Zhang ; Bingbing Ni ; Qi Tian
COMMENTS:       20 pages. Code is available at \url{https://github.com/PeterouZh/NAS_cGAN}
HIGHLIGHT:      This paper presents a novel idea that adopts NAS to find a distinct architecture for each class.

38, TITLE:      Learning Task-General Representations with Generative Neuro-Symbolic Modeling
http://arxiv.org/abs/2006.14448
AUTHORS:        Reuben Feinman ; Brenden M. Lake
HIGHLIGHT:      To help bridge this gap, we propose Generative Neuro-Symbolic (GNS) Modeling, a framework for learning task-general representations by combining the structure of symbolic models with the expressivity of neural networks.

39, TITLE:      Multilingual Jointly Trained Acoustic and Written Word Embeddings
http://arxiv.org/abs/2006.14007
AUTHORS:        Yushi Hu ; Shane Settle ; Karen Livescu
HIGHLIGHT:      In this work, we extend this idea to multiple low-resource languages.

40, TITLE:      Riccati-based feedback stabilization for unstable Power system models
http://arxiv.org/abs/2006.14210
AUTHORS:        Mahtab Uddin ; M. Monir Uddin ; Md. Abdul Hakim Khan
COMMENTS:       28 pages, 19 figures
HIGHLIGHT:      In this article, the objective is mainly focused on finding optimal control for the large-scale sparse unstable power system models using optimal feedback matrix achieved by the Riccati-based feedback stabilization process.

41, TITLE:      Deep Residual 3D U-Net for Joint Segmentation and Texture Classification of Nodules in Lung
http://arxiv.org/abs/2006.14215
AUTHORS:        Alexandr G. Rassadin
COMMENTS:       10 pages, 5 figures, 2 tables, accepted for publication at ICIAR 2020(LNDb Grand Challenge)
HIGHLIGHT:      In this work we present a method for lung nodules segmentation, their texture classification and subsequent follow-up recommendation from the CT image of lung.

42, TITLE:      Raising Expectations: Automating Expected Cost Analysis with Types
http://arxiv.org/abs/2006.14010
AUTHORS:        Di Wang ; David M Kahn ; Jan Hoffmann
HIGHLIGHT:      This article presents a type-based analysis for deriving upper bounds on the expected execution cost of probabilistic programs.

43, TITLE:      SmallBigNet: Integrating Core and Contextual Views for Video Classification
http://arxiv.org/abs/2006.14582
AUTHORS:        Xianhang Li ; Yali Wang ; Zhipeng Zhou ; Yu Qiao
COMMENTS:       CVPR2020
HIGHLIGHT:      To alleviate this problem, we propose a concise and novel SmallBig network, with the cooperation of small and big views.

44, TITLE:      SS-CAM: Smoothed Score-CAM for Sharper Visual Feature Localization
http://arxiv.org/abs/2006.14255
AUTHORS:        Rakshit Naidu ; Joy Michael
COMMENTS:       6 pages and 4 figures and 2 tables
HIGHLIGHT:      In this paper, built on the top of Score-CAM, we introduce an enhanced visual explanation in terms of visual sharpness called SS-CAM, which produces sharper localizations of object features within an image by smoothing.

45, TITLE:      Backdoor Attacks on Facial Recognition in the Physical World
http://arxiv.org/abs/2006.14580
AUTHORS:        Emily Wenger ; Josephine Passananti ; Yuanshun Yao ; Haitao Zheng ; Ben Y. Zhao
HIGHLIGHT:      In this paper, we present results of a detailed study on DNN backdoor attacks in the physical world, specifically focused on the task of facial recognition.

46, TITLE:      IIT Gandhinagar at SemEval-2020 Task 9: Code-Mixed Sentiment Classification Using Candidate Sentence Generation and Selection
http://arxiv.org/abs/2006.14465
AUTHORS:        Vivek Srivastava ; Mayank Singh
HIGHLIGHT:      We present a candidate sentence generation and selection based approach on top of the Bi-LSTM based neural classifier to classify the Hinglish code-mixed text into one of the three sentiment classes positive, negative, or neutral.

47, TITLE:      Neural Machine Translation For Paraphrase Generation
http://arxiv.org/abs/2006.14223
AUTHORS:        Alex Sokolov ; Denis Filimonov
COMMENTS:       Published in NIPS 2018: 2nd Conversational AI workshop
HIGHLIGHT:      In this work, we present an automatic natural language generation system, capable of generating both human-like interactions and annotations by the means of paraphrasing.

48, TITLE:      Road obstacles positional and dynamic features extraction combining object detection, stereo disparity maps and optical flow data
http://arxiv.org/abs/2006.14011
AUTHORS:        Thiago Rateke ; Aldo von Wangenheim
COMMENTS:       11 pages
HIGHLIGHT:      In this paper we present an approach for the identification of obstacles and extraction of class, position, depth and motion information from these objects that employs data gained exclusively from passive vision.

49, TITLE:      Fast and stable MAP-Elites in noisy domains using deep grids
http://arxiv.org/abs/2006.14253
AUTHORS:        Manon Flageat ; Antoine Cully
COMMENTS:       10 pages, 4 figures, to be published in the Proceedings of the 2020 Conference on Artificial Life
HIGHLIGHT:      In this work, we propose Deep-Grid MAP-Elites, a variant of the MAP-Elites algorithm that uses an archive of similar previously encountered solutions to approximate the performance of a solution.

50, TITLE:      XREF: Entity Linking for Chinese News Comments with Supplementary Article Reference
http://arxiv.org/abs/2006.14017
AUTHORS:        Xinyu Hua ; Lei Li ; Lifeng Hua ; Lu Wang
COMMENTS:       Accepted to AKBC2020, link to openreview: https://openreview.net/forum?id=1hLH6CKIjN
HIGHLIGHT:      In this paper, we study the problem of entity linking for Chinese news comments given mentions' spans.

51, TITLE:      Vector-Matrix-Vector Queries for Solving Linear Algebra, Statistics, and Graph Problems
http://arxiv.org/abs/2006.14015
AUTHORS:        Cyrus Rashtchian ; David P. Woodruff ; Hanlin Zhu
COMMENTS:       26 pages, to be published in RANDOM 2020
HIGHLIGHT:      We consider the general problem of learning about a matrix through vector-matrix-vector queries.

52, TITLE:      Epoch-evolving Gaussian Process Guided Learning
http://arxiv.org/abs/2006.14347
AUTHORS:        Jiabao Cui ; Xuewei Li ; Bin Li ; Hanbin Zhao ; Bourahla Omar ; Xi Li
HIGHLIGHT:      In this paper, we propose a novel learning scheme called epoch-evolving Gaussian Process Guided Learning (GPGL), which aims at characterizing the correlation information between the batch-level distribution and the global data distribution.

53, TITLE:      Block-matching in FPGA
http://arxiv.org/abs/2006.14105
AUTHORS:        Rafael Pizarro Solar ; Michal Pleskowicz
COMMENTS:       19 pages, 15 figures, paper submitted in "CS413 - Computational Photography" at EPFL, for project repository see $\href{https://github.com/UlisesLuzius/ImageProcessingPipeline/}{\text{link}}$
HIGHLIGHT:      Our goal is to enable other researchers to use our solution in the future for real-time video denoising in video cameras that use FPGAs (such as the AXIOM Beta).

54, TITLE:      SACT: Self-Aware Multi-Space Feature Composition Transformer for Multinomial Attention for Video Captioning
http://arxiv.org/abs/2006.14262
AUTHORS:        Chiranjib Sur
HIGHLIGHT:      In this work, we have introduced a new concept of Self-Aware Composition Transformer (SACT) that is capable of generating Multinomial Attention (MultAtt) which is a way of generating distributions of various combinations of frames.

55, TITLE:      Collaborative Boundary-aware Context Encoding Networks for Error Map Prediction
http://arxiv.org/abs/2006.14345
AUTHORS:        Zhenxi Zhang ; Chunna Tian ; Jie Li ; Zhusi Zhong ; Zhicheng Jiao ; Xinbo Gao
HIGHLIGHT:      In this paper, we propose collaborative boundaryaware context encoding networks called AEP-Net for error prediction task.

56, TITLE:      Privacy at Facebook Scale
http://arxiv.org/abs/2006.14109
AUTHORS:        Paulo Tanaka ; Sameet Sapra ; Nikolay Laptev
HIGHLIGHT:      Privacy at Facebook Scale

57, TITLE:      Empirical Analysis of Overfitting and Mode Drop in GAN Training
http://arxiv.org/abs/2006.14265
AUTHORS:        Yasin Yazici ; Chuan-Sheng Foo ; Stefan Winkler ; Kim-Hui Yap ; Vijay Chandrasekhar
COMMENTS:       To appear in ICIP2020
HIGHLIGHT:      We examine two key questions in GAN training, namely overfitting and mode drop, from an empirical perspective.

58, TITLE:      Kinematic-Structure-Preserved Representation for Unsupervised 3D Human Pose Estimation
http://arxiv.org/abs/2006.14107
AUTHORS:        Jogendra Nath Kundu ; Siddharth Seth ; Rahul M V ; Mugalodi Rakesh ; R. Venkatesh Babu ; Anirban Chakraborty
COMMENTS:       AAAI 2020 (Oral)
HIGHLIGHT:      Though weakly-supervised models have been proposed to address this shortcoming, performance of such models relies on availability of paired supervision on some related tasks, such as 2D pose or multi-view image pairs.

59, TITLE:      Self-Segregating and Coordinated-Segregating Transformer for Focused Deep Multi-Modular Network for Visual Question Answering
http://arxiv.org/abs/2006.14264
AUTHORS:        Chiranjib Sur
HIGHLIGHT:      We defined two strategies: Self-Segregating Transformer (SST) and Coordinated-Segregating Transformer (CST) and used it to solve visual question answering application.

60, TITLE:      Ultra-Low-Power FDSOI Neural Circuits for Extreme-Edge Neuromorphic Intelligence
http://arxiv.org/abs/2006.14270
AUTHORS:        Arianna Rubino ; Can Livanelioglu ; Ning Qiao ; Melika Payvand ; Giacomo Indiveri
COMMENTS:       11 pages, 9 figures, TCAS submission
HIGHLIGHT:      To reduce power consumption even further, we present a set of mixed-signal analog/digital circuits that exploit the features of advanced Fully-Depleted Silicon on Insulator (FDSOI) integration processes.

61, TITLE:      A Closer Look at Invalid Action Masking in Policy Gradient Algorithms
http://arxiv.org/abs/2006.14171
AUTHORS:        Shengyi Huang ; Santiago Ontañón
COMMENTS:       Preprint. Corrected a major issue of the withdrawn version submitted to NeurIPS 2020
HIGHLIGHT:      In this paper, we show that the standard working mechanism of invalid action masking corresponds to valid policy gradient updates.

62, TITLE:      Compositional Explanations of Neurons
http://arxiv.org/abs/2006.14032
AUTHORS:        Jesse Mu ; Jacob Andreas
HIGHLIGHT:      We describe a procedure for explaining neurons in deep representations by identifying compositional logical concepts that closely approximate neuron behavior.

63, TITLE:      Towards Differentially Private Text Representations
http://arxiv.org/abs/2006.14170
AUTHORS:        Lingjuan Lyu ; Yitong Li ; Xuanli He ; Tong Xiao
COMMENTS:       Accepted to SIGIR'20
HIGHLIGHT:      For the randomization module, we propose a novel local differentially private (LDP) protocol to reduce the impact of privacy parameter $\epsilon$ on accuracy, and provide enhanced flexibility in choosing randomization probabilities for LDP.

64, TITLE:      Learning to simulate complex scenes
http://arxiv.org/abs/2006.14611
AUTHORS:        Zhenfeng Xue ; Weijie Mao ; Liang Zheng
COMMENTS:       13 pages, 13 figures
HIGHLIGHT:      To optimize the attribute values and obtain a training set of similar content to real-world data, we propose a scalable discretization-and-relaxation (SDR) approach.

65, TITLE:      A causal view of compositional zero-shot recognition
http://arxiv.org/abs/2006.14610
AUTHORS:        Yuval Atzmon ; Felix Kreuk ; Uri Shalit ; Gal Chechik
HIGHLIGHT:      Here we describe an approach for compositional generalization that builds on causal ideas.

66, TITLE:      An Analysis of SVD for Deep Rotation Estimation
http://arxiv.org/abs/2006.14616
AUTHORS:        Jake Levinson ; Carlos Esteves ; Kefan Chen ; Noah Snavely ; Angjoo Kanazawa ; Afshin Rostamizadeh ; Ameesh Makadia
HIGHLIGHT:      We present a theoretical analysis that shows SVD is the natural choice for projecting onto the rotation group.

67, TITLE:      Layout Generation and Completion with Self-attention
http://arxiv.org/abs/2006.14615
AUTHORS:        Kamal Gupta ; Alessandro Achille ; Justin Lazarow ; Larry Davis ; Vijay Mahadevan ; Abhinav Shrivastava
HIGHLIGHT:      To do this, we propose a novel framework, LayoutTransformer, that leverages a self-attention based approach to learn contextual relationships between layout elements and generate layouts in a given domain.

68, TITLE:      Neural Machine Translation for Multilingual Grapheme-to-Phoneme Conversion
http://arxiv.org/abs/2006.14194
AUTHORS:        Alex Sokolov ; Tracy Rohlin ; Ariya Rastrow
COMMENTS:       Published in INTERSPEECH (2019)
HIGHLIGHT:      As an alternative, we present a single end-to-end trained neural G2P model that shares same encoder and decoder across multiple languages.

69, TITLE:      Space-Time Correspondence as a Contrastive Random Walk
http://arxiv.org/abs/2006.14613
AUTHORS:        Allan Jabri ; Andrew Owens ; Alexei A. Efros
HIGHLIGHT:      This paper proposes a simple self-supervised approach for learning representations for visual correspondence from raw video.

70, TITLE:      A Simple Approach to Case-Based Reasoning in Knowledge Bases
http://arxiv.org/abs/2006.14198
AUTHORS:        Rajarshi Das ; Ameya Godbole ; Shehzaad Dhuliawala ; Manzil Zaheer ; Andrew McCallum
HIGHLIGHT:      We present a surprisingly simple yet accurate approach to reasoning in knowledge graphs (KGs) that requires \emph{no training}, and is reminiscent of case-based reasoning in classical artificial intelligence (AI).

71, TITLE:      Parametric Instance Classification for Unsupervised Visual Feature Learning
http://arxiv.org/abs/2006.14618
AUTHORS:        Yue Cao ; Zhenda Xie ; Bin Liu ; Yutong Lin ; Zheng Zhang ; Han Hu
HIGHLIGHT:      This paper presents parametric instance classification (PIC) for unsupervised visual feature learning.

72, TITLE:      Does Adversarial Transferability Indicate Knowledge Transferability?
http://arxiv.org/abs/2006.14512
AUTHORS:        Kaizhao Liang ; Jacky Y. Zhang ; Oluwasanmi Koyejo ; Bo Li
COMMENTS:       First two authors contributed equally. Code https://github.com/AI-secure/Does-Adversairal-Transferability-Indicate-Knowledge-Transferability
HIGHLIGHT:      In this paper, we aim to turn the existence and pervasiveness of adversarial examples into an advantage.

73, TITLE:      On Mitigating Random and Adversarial Bit Errors
http://arxiv.org/abs/2006.13977
AUTHORS:        David Stutz ; Nandhini Chandramoorthy ; Matthias Hein ; Bernt Schiele
HIGHLIGHT:      Besides describing these error models in detail, we make first steps towards DNNs robust to random and adversarial bit errors by explicitly taking bit errors into account during training.

74, TITLE:      Unsupervised Cross-lingual Representation Learning for Speech Recognition
http://arxiv.org/abs/2006.13979
AUTHORS:        Alexis Conneau ; Alexei Baevski ; Ronan Collobert ; Abdelrahman Mohamed ; Michael Auli
HIGHLIGHT:      This paper presents XLSR which learns cross-lingual speech representations by pretraining a single model from the raw waveform of speech in multiple languages.

75, TITLE:      Extended Labeled Faces in-the-Wild (ELFW): Augmenting Classes for Face Segmentation
http://arxiv.org/abs/2006.13980
AUTHORS:        Rafael Redondo ; Jaume Gibert
COMMENTS:       14 pages, 12 figures
HIGHLIGHT:      In this work, we introduce Extended Labeled Faces in-the-Wild (ELFW), a dataset supplementing with additional face-related categories -- and also additional faces -- the originally released semantic labels in the vastly used Labeled Faces in-the-Wild (LFW) dataset.

76, TITLE:      Minimum Cost Active Labeling
http://arxiv.org/abs/2006.13999
AUTHORS:        Hang Qiu ; Krishna Chintalapudi ; Ramesh Govindan
HIGHLIGHT:      In this paper, we consider the problem of minimum-cost labeling: classifying all images in a large data set with a target accuracy bound at minimum dollar cost.

77, TITLE:      Normalizing Text using Language Modelling based on Phonetics and String Similarity
http://arxiv.org/abs/2006.14116
AUTHORS:        Fenil Doshi ; Jimit Gandhi ; Deep Gosalia ; Sudhir Bagul
COMMENTS:       Author 1, 2 and 3 are equal contributors; Number of pages: 9; Number of figures: 3
HIGHLIGHT:      We propose a new robust model to perform text normalization.



==========Updates to Previous Papers==========
1, TITLE:       Facing the Hard Problems in FGVC
http://arxiv.org/abs/2006.13190
AUTHORS:        Connor Anderson ; Matt Gwilliam ; Adam Teuscher ; Andrew Merrill ; Ryan Farrell
COMMENTS:       17 pages, 6 figures, 2 tables; fixed typo, minor adjustment to format, added equations
HIGHLIGHT:      We underscore the importance of such analysis, and demonstrate that combining complementary models can improve accuracy on the popular CUB-200 dataset by over 5%. In addition to detailed analysis and characterization of the errors made by these SOTA methods, we provide a clear set of recommended directions for future FGVC researchers.

2, TITLE:       ERA: A Dataset and Deep Learning Benchmark for Event Recognition in Aerial Videos
http://arxiv.org/abs/2001.11394
AUTHORS:        Lichao Mou ; Yuansheng Hua ; Pu Jin ; Xiao Xiang Zhu
COMMENTS:       IEEE Geoscience and Remote Sensing Magazine. Project page: https://lcmou.github.io/ERA_Dataset/
HIGHLIGHT:      In this paper, we introduce a novel problem of event recognition in unconstrained aerial videos in the remote sensing community and present a large-scale, human-annotated dataset, named ERA (Event Recognition in Aerial videos), consisting of 2,864 videos each with a label from 25 different classes corresponding to an event unfolding 5 seconds.

3, TITLE:       ReZero is All You Need: Fast Convergence at Large Depth
http://arxiv.org/abs/2003.04887
AUTHORS:        Thomas Bachlechner ; Bodhisattwa Prasad Majumder ; Huanru Henry Mao ; Garrison W. Cottrell ; Julian McAuley
HIGHLIGHT:      We apply this technique to language modeling and find that we can easily train 120-layer Transformers.

4, TITLE:       Universal Equivariant Multilayer Perceptrons
http://arxiv.org/abs/2002.02912
AUTHORS:        Siamak Ravanbakhsh
HIGHLIGHT:      Using tools from group theory, this paper proves the universality of a broad class of equivariant MLPs with a single hidden layer.

5, TITLE:       Deep Attentive Study Session Dropout Prediction in Mobile Learning Environment
http://arxiv.org/abs/2002.11624
AUTHORS:        Youngnam Lee ; Dongmin Shin ; HyunBin Loh ; Jaemin Lee ; Piljae Chae ; Junghyun Cho ; Seoyon Park ; Jinhwan Lee ; Jineon Baek ; Byungsoo Kim ; Youngduck Choi
COMMENTS:       CSEDU 2020
HIGHLIGHT:      In this paper, we investigate the study session dropout prediction problem in a mobile learning environment.

6, TITLE:       WaveFlow: A Compact Flow-based Model for Raw Audio
http://arxiv.org/abs/1912.01219
AUTHORS:        Wei Ping ; Kainan Peng ; Kexin Zhao ; Zhao Song
COMMENTS:       Published at ICML 2020. Code and pre-trained models: https://github.com/PaddlePaddle/Parakeet
HIGHLIGHT:      In this work, we propose WaveFlow, a small-footprint generative flow for raw audio, which is directly trained with maximum likelihood.

7, TITLE:       The Costs and Benefits of Goal-Directed Attention in Deep Convolutional Neural Networks
http://arxiv.org/abs/2002.02342
AUTHORS:        Xiaoliang Luo ; Brett D. Roads ; Bradley C. Love
HIGHLIGHT:      Akin to how attention weights in categorisation models warp representational spaces, we introduce a layer of attention weights to a DCNN that amplify or attenuate activity to further a goal.

8, TITLE:       Multimodal grid features and cell pointers for Scene Text Visual Question Answering
http://arxiv.org/abs/2006.00923
AUTHORS:        Lluís Gómez ; Ali Furkan Biten ; Rubèn Tito ; Andrés Mafla ; Marçal Rusiñol ; Ernest Valveny ; Dimosthenis Karatzas
COMMENTS:       This paper is under consideration at Pattern Recognition Letters
HIGHLIGHT:      This paper presents a new model for the task of scene text visual question answering, in which questions about a given image can only be answered by reading and understanding scene text that is present in it.

9, TITLE:       Semantic Understanding of Foggy Scenes with Purely Synthetic Data
http://arxiv.org/abs/1910.03997
AUTHORS:        Martin Hahner ; Dengxin Dai ; Christos Sakaridis ; Jan-Nico Zaech ; Luc Van Gool
COMMENTS:       independent class IoU scores corrected for BiSiNet architecture
HIGHLIGHT:      In this paper, we propose a novel method, which uses purely synthetic data to improve the performance on unseen real-world foggy scenes captured in the streets of Zurich and its surroundings. Our contributions are threefold, 1) we created a purely synthetic, high-quality foggy dataset of 25,000 unique outdoor scenes, that we call Foggy Synscapes and plan to release publicly 2) we show that with this data we outperform previous approaches on real-world foggy test data 3) we show that a combination of our data and previously used data can even further improve the performance on real-world foggy data.

10, TITLE:      Intelligent Coordination among Multiple Traffic Intersections Using Multi-Agent Reinforcement Learning
http://arxiv.org/abs/1912.03851
AUTHORS:        Ujwal Padam Tewari ; Vishal Bidawatka ; Varsha Raveendran ; Vinay Sudhakaran ; Shreedhar Kodate Shreeshai ; Jayanth V. Kulkarni
COMMENTS:       Accepted in the NeurIPS 2019 Deep RL Workshop : https://sites.google.com/view/deep-rl-workshop-neurips-2019/home
HIGHLIGHT:      We use Asynchronous Advantage Actor Critic (A3C) for implementing an AI agent in the controllers that optimize flow of traffic across a single intersection and then extend it to multiple intersections by considering a multi-agent setting.

11, TITLE:      Extract with Order for Coherent Multi-Document Summarization
http://arxiv.org/abs/1706.06542
AUTHORS:        Mir Tafseer Nayeem ; Yllias Chali
COMMENTS:       TextGraphs-11 at ACL 2017
HIGHLIGHT:      In this work, we aim at developing an extractive summarizer in the multi-document setting.

12, TITLE:      Towards a Metric for Automated Conversational Dialogue System Evaluation and Improvement
http://arxiv.org/abs/1909.12066
AUTHORS:        Jan Deriu ; Mark Cieliebak
COMMENTS:       8 Pages, To be published at the INLG 2019 converence
HIGHLIGHT:      We present "AutoJudge", an automated evaluation method for conversational dialogue systems.

13, TITLE:      Learning Nonlinear Loop Invariants with Gated Continuous Logic Networks (Extended Version)
http://arxiv.org/abs/2003.07959
AUTHORS:        Jianan Yao ; Gabriel Ryan ; Justin Wong ; Suman Jana ; Ronghui Gu
HIGHLIGHT:      In this paper, we introduce a new neural architecture for general SMT learning, the Gated Continuous Logic Network (G-CLN), and apply it to nonlinear loop invariant learning.

14, TITLE:      Deep-CAPTCHA: a deep learning based CAPTCHA solver for vulnerability assessment
http://arxiv.org/abs/2006.08296
AUTHORS:        Zahra Noury ; Mahdi Rezaei
COMMENTS:       Version 2.0
HIGHLIGHT:      In this research, we investigate a way to crack visual CAPTCHA tests by an automated deep learning based solution. To train and develop an efficient model, we have generated a dataset of 500,000 CAPTCHAs to train our model.

15, TITLE:      PENNI: Pruned Kernel Sharing for Efficient CNN Inference
http://arxiv.org/abs/2005.07133
AUTHORS:        Shiyu Li ; Edward Hanson ; Hai Li ; Yiran Chen
COMMENTS:       9 pages, 5 figures, to appear on ICML2020
HIGHLIGHT:      Based on this observation, we propose PENNI, a CNN model compression framework that is able to achieve model compactness and hardware efficiency simultaneously by (1) implementing kernel sharing in convolution layers via a small number of basis kernels and (2) alternately adjusting bases and coefficients with sparse constraints.

16, TITLE:      NODIS: Neural Ordinary Differential Scene Understanding
http://arxiv.org/abs/2001.04735
AUTHORS:        Cong Yuren ; Hanno Ackermann ; Wentong Liao ; Michael Ying Yang ; Bodo Rosenhahn
HIGHLIGHT:      In this work, we interpret that formulation as Ordinary Differential Equation (ODE).

17, TITLE:      KernelNet: A Data-Dependent Kernel Parameterization for Deep Generative Modeling
http://arxiv.org/abs/1912.00979
AUTHORS:        Yufan Zhou ; Changyou Chen ; Jinhui Xu
HIGHLIGHT:      To mitigate this burden, we propose in this paper a framework to construct and learn a data-dependent kernel based on random features and implicit spectral distributions that are parameterized by deep neural networks.

18, TITLE:      Two Routes to Scalable Credit Assignment without Weight Symmetry
http://arxiv.org/abs/2003.01513
AUTHORS:        Daniel Kunin ; Aran Nayebi ; Javier Sagastuy-Brena ; Surya Ganguli ; Jonathan M. Bloom ; Daniel L. K. Yamins
COMMENTS:       ICML 2020 Camera Ready Version, 19 pages including supplementary information, 10 figures
HIGHLIGHT:      Nonetheless, we find a performance and stability gap between this local rule and backpropagation that widens with increasing model depth.

19, TITLE:      The DeepFake Detection Challenge Dataset
http://arxiv.org/abs/2006.07397
AUTHORS:        Brian Dolhansky ; Joanna Bitton ; Ben Pflaum ; Jikuo Lu ; Russ Howes ; Menglin Wang ; Cristian Canton Ferrer
HIGHLIGHT:      To counter this emerging threat, we have constructed an extremely large face swap video dataset to enable the training of detection models, and organized the accompanying DeepFake Detection Challenge (DFDC) Kaggle competition.

20, TITLE:      MetalGAN: Multi-Domain Label-Less Image Synthesis Using cGANs and Meta-Learning
http://arxiv.org/abs/1912.02494
AUTHORS:        Tomaso Fontanini ; Eleonora Iotti ; Luca Donati ; Andrea Prati
HIGHLIGHT:      This paper proposes a novel architecture and a training algorithm, which are able to produce multi-domain outputs using a single network.

21, TITLE:      Multi-factorial Optimization for Large-scale Virtual Machine Placement in Cloud Computing
http://arxiv.org/abs/2001.06585
AUTHORS:        Zhengping Liang ; Jian Zhang ; Liang Feng ; Zexuan Zhu
HIGHLIGHT:      This paper aims to apply the MFO technology to the LVMP problem in heterogeneous environment.

22, TITLE:      A Methodology for Creating Question Answering Corpora Using Inverse Data Annotation
http://arxiv.org/abs/2004.07633
AUTHORS:        Jan Deriu ; Katsiaryna Mlynchyk ; Philippe Schläpfer ; Alvaro Rodrigo ; Dirk von Grünigen ; Nicolas Kaiser ; Kurt Stockinger ; Eneko Agirre ; Mark Cieliebak
HIGHLIGHT:      In this paper, we introduce a novel methodology to efficiently construct a corpus for question answering over structured data.

23, TITLE:      Bridging Anaphora Resolution as Question Answering
http://arxiv.org/abs/2004.07898
AUTHORS:        Yufang Hou
COMMENTS:       accepted at ACL2020. This version is slightly different than the ACL2020 camera-ready version. Thanks for Massimo Poesio's comments, I've made two small changes to describe GNOME and the work in Poesio et al. (2004) more accurately
HIGHLIGHT:      In this paper, we cast bridging anaphora resolution as question answering based on context.

24, TITLE:      SurfelGAN: Synthesizing Realistic Sensor Data for Autonomous Driving
http://arxiv.org/abs/2005.03844
AUTHORS:        Zhenpei Yang ; Yuning Chai ; Dragomir Anguelov ; Yin Zhou ; Pei Sun ; Dumitru Erhan ; Sean Rafferty ; Henrik Kretzschmar
HIGHLIGHT:      In this paper, we present a simple yet effective approach to generate realistic scenario sensor data, based only on a limited amount of lidar and camera data collected by an autonomous vehicle. We also create a novel dataset that contains cases in which two self-driving vehicles observe the same scene at the same time.

25, TITLE:      Modular Termination for Second-Order Computation Rules and Application to Algebraic Effect Handlers
http://arxiv.org/abs/1912.03434
AUTHORS:        Makoto Hamana
COMMENTS:       26 pages
HIGHLIGHT:      We present a new modular proof method of termination for second-order computation, and report its implementation SOL.

26, TITLE:      Diversity can be Transferred: Output Diversification for White- and Black-box Attacks
http://arxiv.org/abs/2003.06878
AUTHORS:        Yusuke Tashiro ; Yang Song ; Stefano Ermon
HIGHLIGHT:      To improve the efficiency of these attacks, we propose Output Diversified Sampling (ODS), a novel sampling strategy that attempts to maximize diversity in the target model's outputs among the generated samples.

27, TITLE:      FaceShifter: Towards High Fidelity And Occlusion Aware Face Swapping
http://arxiv.org/abs/1912.13457
AUTHORS:        Lingzhi Li ; Jianmin Bao ; Hao Yang ; Dong Chen ; Fang Wen
COMMENTS:       Accepted to CVPR 2020 (Oral), project webpage: lingzhili.com/FaceShifterPage/
HIGHLIGHT:      In this work, we propose a novel two-stage framework, called FaceShifter, for high fidelity and occlusion aware face swapping.

28, TITLE:      ACEnet: Anatomical Context-Encoding Network for Neuroanatomy Segmentation
http://arxiv.org/abs/2002.05773
AUTHORS:        Yuemeng Li ; Hongming Li ; Yong Fan
HIGHLIGHT:      In order to overcome this limitation, we develop an Anatomical Context-Encoding Network (ACEnet) to incorporate 3D spatial and anatomical contexts in 2D convolutional neural networks (CNNs) for efficient and accurate segmentation of brain structures from MR scans, consisting of 1) an anatomical context encoding module to incorporate anatomical information in 2D CNNs and 2) a spatial context encoding module to integrate 3D image information in 2D CNNs.

29, TITLE:      Softmax GAN
http://arxiv.org/abs/1704.06191
AUTHORS:        Min Lin
COMMENTS:       NIPS 2017 submission
HIGHLIGHT:      We futher demonstrate with experiments that this simple change stabilizes GAN training.

30, TITLE:      Inference with Artificial Neural Networks on Analog Neuromorphic Hardware
http://arxiv.org/abs/2006.13177
AUTHORS:        Johannes Weis ; Philipp Spilger ; Sebastian Billaudelle ; Yannik Stradmann ; Arne Emmel ; Eric Müller ; Oliver Breitwieser ; Andreas Grübl ; Joscha Ilmberger ; Vitali Karasenko ; Mitja Kleider ; Christian Mauch ; Korbinian Schreiber ; Johannes Schemmel
HIGHLIGHT:      In this paper, we discuss BrainScaleS-2 as an analog inference accelerator and present calibration as well as optimization strategies, highlighting the advantages of training with hardware in the loop.

31, TITLE:      Randomized Smoothing of All Shapes and Sizes
http://arxiv.org/abs/2002.08118
AUTHORS:        Greg Yang ; Tony Duan ; J. Edward Hu ; Hadi Salman ; Ilya Razenshteyn ; Jerry Li
COMMENTS:       9 pages main text, 48 pages total
HIGHLIGHT:      We propose a novel framework for devising and analyzing randomized smoothing schemes, and validate its effectiveness in practice.

32, TITLE:      Self-supervised Learning for Astronomical Image Classification
http://arxiv.org/abs/2004.11336
AUTHORS:        Ana Martinazzo ; Mateus Espadoto ; Nina S. T. Hirata
COMMENTS:       Accepted for ICPR 2020
HIGHLIGHT:      In this paper, we propose a technique to leverage unlabeled astronomical images to pre-train deep convolutional neural networks, in order to learn a domain-specific feature extractor which improves the results of machine learning techniques in setups with small amounts of labeled data available.

33, TITLE:      Deep Direct Visual Odometry
http://arxiv.org/abs/1912.05101
AUTHORS:        Chaoqiang Zhao ; Yang Tang ; Qiyu Sun ; Athanasios V. Vasilakos
COMMENTS:       10 pages,8 figures
HIGHLIGHT:      With the outstanding performance of deep learning, like image analysis and processing, previous works have shown that deep neural networks can effectively learn the 6-DOF pose between frames from large volumes of image sequences in an unsupervised manner.

34, TITLE:      Domain Decluttering: Simplifying Images to Mitigate Synthetic-Real Domain Shift and Improve Depth Estimation
http://arxiv.org/abs/2002.12114
AUTHORS:        Yunhan Zhao ; Shu Kong ; Daeyun Shin ; Charless Fowlkes
COMMENTS:       camera-ready version, CVPR2020
HIGHLIGHT:      We carry out extensive experiments to validate our attend-remove-complete approach (ARC) and find that it significantly outperforms state-of-the-art domain adaptation methods for depth prediction.

35, TITLE:      AdvectiveNet: An Eulerian-Lagrangian Fluidic reservoir for Point Cloud Processing
http://arxiv.org/abs/2002.00118
AUTHORS:        Xingzhe He ; Helen Lu Cao ; Bo Zhu
COMMENTS:       ICLR 2020
HIGHLIGHT:      This paper presents a novel physics-inspired deep learning approach for point cloud processing motivated by the natural flow phenomena in fluid mechanics.

36, TITLE:      Learning Adaptive Regularization for Image Labeling Using Geometric Assignment
http://arxiv.org/abs/1910.09976
AUTHORS:        Ruben Hühnerbein ; Fabrizio Savarino ; Stefania Petra ; Christoph Schnörr
HIGHLIGHT:      We study the inverse problem of model parameter learning for pixelwise image labeling, using the linear assignment flow and training data with ground truth.

37, TITLE:      Discriminative Feature Alignment: Improving Transferability of Unsupervised Domain Adaptation by Gaussian-guided Latent Alignment
http://arxiv.org/abs/2006.12770
AUTHORS:        Jing Wang ; Jiahong Chen ; Jianzhe Lin ; Leonid Sigal ; Clarence W. de Silva
COMMENTS:       12 pages, 12 figures
HIGHLIGHT:      In this study, we focus on the unsupervised domain adaptation problem where an approximate inference model is to be learned from a labeled data domain and expected to generalize well to an unlabeled data domain.

38, TITLE:      Deep DIH : Statistically Inferred Reconstruction of Digital In-Line Holography by Deep Learning
http://arxiv.org/abs/2004.12231
AUTHORS:        Huayu Li ; Xiwen Chen ; Haiyu Wu ; Zaoyi Chi ; Christopher Mann ; Abolfazl Razi
HIGHLIGHT:      In this paper, we proposed a novel implementation of autoencoder-based deep learning architecture for single-shot hologram reconstruction solely based on the current sample without the need for massive datasets to train the model.

39, TITLE:      Measuring Compositional Generalization: A Comprehensive Method on Realistic Data
http://arxiv.org/abs/1912.09713
AUTHORS:        Daniel Keysers ; Nathanael Schärli ; Nathan Scales ; Hylke Buisman ; Daniel Furrer ; Sergii Kashubin ; Nikola Momchev ; Danila Sinopalnikov ; Lukasz Stafiniak ; Tibor Tihon ; Dmitry Tsarkov ; Xiao Wang ; Marc van Zee ; Olivier Bousquet
COMMENTS:       Accepted for publication at ICLR 2020
HIGHLIGHT:      We introduce a novel method to systematically construct such benchmarks by maximizing compound divergence while guaranteeing a small atom divergence between train and test sets, and we quantitatively compare this method to other approaches for creating compositional generalization benchmarks.

40, TITLE:      IQA: Interactive Query Construction in Semantic Question Answering Systems
http://arxiv.org/abs/2006.11534
AUTHORS:        Hamid Zafar ; Mohnish Dubey ; Jens Lehmann ; Elena Demidova
HIGHLIGHT:      In this article, we aim to empower users in guiding SQA systems towards the intended semantic queries through interaction.

41, TITLE:      A review of possible effects of cognitive biases on interpretation of rule-based machine learning models
http://arxiv.org/abs/1804.02969
AUTHORS:        Tomáš Kliegr ; Štěpán Bahník ; Johannes Fürnkranz
HIGHLIGHT:      In particular, the goal of this paper is to discuss to what extent cognitive biases may affect human understanding of interpretable machine learning models, in particular of logical rules discovered from data.

42, TITLE:      Artist-Guided Semiautomatic Animation Colorization
http://arxiv.org/abs/2006.13717
AUTHORS:        Harrish Thasarathan ; Mehran Ebrahimi
COMMENTS:       This article supersedes our previous work arXiv:1904.09527
HIGHLIGHT:      We present a method for automating line art colorization by keeping artists in the loop to successfully reduce this workload while staying true to an artist's vision.

43, TITLE:      Prescribing Deep Attentive Score Prediction Attracts Improved Student Engagement
http://arxiv.org/abs/2005.05021
AUTHORS:        Youngnam Lee ; Byungsoo Kim ; Dongmin Shin ; JungHoon Kim ; Jineon Baek ; Jinhwan Lee ; Youngduck Choi
COMMENTS:       EDM 2020
HIGHLIGHT:      In this paper, we demonstrate that the accuracy of the score prediction model deployed in a real-world setting significantly impacts user engagement by providing empirical evidence.

44, TITLE:      Unsupervised Data Augmentation for Consistency Training
http://arxiv.org/abs/1904.12848
AUTHORS:        Qizhe Xie ; Zihang Dai ; Eduard Hovy ; Minh-Thang Luong ; Quoc V. Le
HIGHLIGHT:      In this work, we present a new perspective on how to effectively noise unlabeled examples and argue that the quality of noising, specifically those produced by advanced data augmentation methods, plays a crucial role in semi-supervised learning.

45, TITLE:      Derivative-free global minimization for a class of multiple minima problems
http://arxiv.org/abs/2006.08181
AUTHORS:        Xiaopeng Luo ; Xin Xu ; Daoyi Dong
COMMENTS:       14 pages, 3 figures
HIGHLIGHT:      We prove that the finite-difference based derivative-free descent (FD-DFD) methods have a capability to find the global minima for a class of multiple minima problems.

46, TITLE:      Predicting Goal-directed Human Attention Using Inverse Reinforcement Learning
http://arxiv.org/abs/2005.14310
AUTHORS:        Zhibo Yang ; Lihan Huang ; Yupei Chen ; Zijun Wei ; Seoyoung Ahn ; Gregory Zelinsky ; Dimitris Samaras ; Minh Hoai
COMMENTS:       16 pages, 13 figures, CVPR 2020
HIGHLIGHT:      We propose the first inverse reinforcement learning (IRL) model to learn the internal reward function and policy used by humans during visual search.

47, TITLE:      Understanding Anomaly Detection with Deep Invertible Networks through Hierarchies of Distributions and Features
http://arxiv.org/abs/2006.10848
AUTHORS:        Robin Tibor Schirrmeister ; Yuxuan Zhou ; Tonio Ball ; Dan Zhang
HIGHLIGHT:      To remove the negative impact of model bias and domain prior on detecting high-level differences, we propose two methods, first, using the log likelihood ratios of two identical models, one trained on the in-distribution data (e.g., CIFAR10) and the other one on a more general distribution of images (e.g., 80 Million Tiny Images).

48, TITLE:      Towards an Appropriate Query, Key, and Value Computation for Knowledge Tracing
http://arxiv.org/abs/2002.07033
AUTHORS:        Youngduck Choi ; Youngnam Lee ; Junghyun Cho ; Jineon Baek ; Byungsoo Kim ; Yeongmin Cha ; Dongmin Shin ; Chan Bae ; Jaewe Heo
COMMENTS:       L@S 2020
HIGHLIGHT:      In this paper, we propose a novel Transformer based model for knowledge tracing, SAINT: Separated Self-AttentIve Neural Knowledge Tracing.

49, TITLE:      Temporal Constraint Satisfaction Problems in Fixed-Point Logic
http://arxiv.org/abs/2002.09451
AUTHORS:        Manuel Bodirsky ; Wied Pakusa ; Jakub Rydval
COMMENTS:       76 pages
HIGHLIGHT:      We prove that there is no Maltsev condition that characterizes Datalog already for the CSPs of first-order reducts of (Q;<); such CSPs are called temporal CSPs and are of fundamental importance in infinite-domain constraint satisfaction.

50, TITLE:      Remote Sensing Image Scene Classification Meets Deep Learning: Challenges, Methods, Benchmarks, and Opportunities
http://arxiv.org/abs/2005.01094
AUTHORS:        Gong Cheng ; Xingxing Xie ; Junwei Han ; Lei Guo ; Gui-Song Xia
COMMENTS:       This manuscript is the accepted version for IEEE JSTARS
HIGHLIGHT:      To be specific, we discuss the main challenges of remote sensing image scene classification and survey (1) Autoencoder-based remote sensing image scene classification methods, (2) Convolutional Neural Network-based remote sensing image scene classification methods, and (3) Generative Adversarial Network-based remote sensing image scene classification methods. In addition, we introduce the benchmarks used for remote sensing image scene classification and summarize the performance of more than two dozen of representative algorithms on three commonly-used benchmark data sets.

51, TITLE:      Deep Learning on Point Clouds for False Positive Reduction at Nodule Detection in Chest CT Scans
http://arxiv.org/abs/2005.03654
AUTHORS:        Ivan Drokin ; Elena Ericheva
HIGHLIGHT:      We propose an algorithm for transforming 3D CT scan data to point cloud.

52, TITLE:      Neural Point Cloud Rendering via Multi-Plane Projection
http://arxiv.org/abs/1912.04645
AUTHORS:        Peng Dai ; Yinda Zhang ; Zhuwen Li ; Shuaicheng Liu ; Bing Zeng
COMMENTS:       17 pages
HIGHLIGHT:      We present a new deep point cloud rendering pipeline through multi-plane projections.

53, TITLE:      Kindly Bent to Free Us
http://arxiv.org/abs/1908.09681
AUTHORS:        Gabriel Radanne ; Hannes Saffrich ; Peter Thiemann
COMMENTS:       ICFP 2020
HIGHLIGHT:      We present Affe, an extension of ML that manages linearity and affinity properties using kinds and constrained types.

54, TITLE:      Brain tumor segmentation with missing modalities via latent multi-source correlation representation
http://arxiv.org/abs/2003.08870
AUTHORS:        Tongxue Zhou ; Stéphane Canu ; Pierre Vera ; Su Ruan
COMMENTS:       9 pages, 6 figures, accepted by MICCAI 2020
HIGHLIGHT:      We evaluate our model on BraTS 2018 datasets, it outperforms the current state-of-the-art method and produces robust results when one or more modalities are missing.

55, TITLE:      A Two stage Adaptive Knowledge Transfer Evolutionary Multi-tasking Based on Population Distribution for Multi/Many-Objective Optimization
http://arxiv.org/abs/2001.00810
AUTHORS:        Zhengping Liang ; Weiqi Liang ; Xiuju Xu ; Zexuan Zhu
COMMENTS:       13 pages, 8 figures, 7 tables, 56 references
HIGHLIGHT:      A Two stage Adaptive Knowledge Transfer Evolutionary Multi-tasking Based on Population Distribution for Multi/Many-Objective Optimization

56, TITLE:      Distributionally Robust Deep Learning using Hardness Weighted Sampling
http://arxiv.org/abs/2001.02658
AUTHORS:        Lucas Fidon ; Sebastien Ourselin ; Tom Vercauteren
HIGHLIGHT:      We propose SGD with hardness weighted sampling, a principled and efficient optimization method for DRO in machine learning that is particularly suited in the context of deep learning.

57, TITLE:      PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models
http://arxiv.org/abs/2003.03808
AUTHORS:        Sachit Menon ; Alexandru Damian ; Shijia Hu ; Nikhil Ravi ; Cynthia Rudin
COMMENTS:       Sachit Menon and Alexandru Damian contributed equally. Computer Vision and Pattern Recognition (CVPR) 2020
HIGHLIGHT:      We present a novel super-resolution algorithm addressing this problem, PULSE (Photo Upsampling via Latent Space Exploration), which generates high-resolution, realistic images at resolutions previously unseen in the literature.

58, TITLE:      Structural Language Models of Code
http://arxiv.org/abs/1910.00577
AUTHORS:        Uri Alon ; Roy Sadaka ; Omer Levy ; Eran Yahav
HIGHLIGHT:      We introduce a new approach to any-code completion that leverages the strict syntax of programming languages to model a code snippet as a tree - structural language modeling (SLM).

59, TITLE:      Assessment Modeling: Fundamental Pre-training Tasks for Interactive Educational Systems
http://arxiv.org/abs/2002.05505
AUTHORS:        Youngduck Choi ; Youngnam Lee ; Junghyun Cho ; Jineon Baek ; Dongmin Shin ; Seewoo Lee ; Youngmin Cha ; Byungsoo Kim ; Jaewe Heo
HIGHLIGHT:      To this end, we propose assessment modeling, fundamental pre-training tasks for IESs.

60, TITLE:      DeepNC: Deep Generative Network Completion
http://arxiv.org/abs/1907.07381
AUTHORS:        Cong Tran ; Won-Yong Shin ; Andreas Spitz ; Michael Gertz
COMMENTS:       15 pages, 10 figures, 5 tables
HIGHLIGHT:      In this paper, we present DeepNC, a novel method for inferring the missing parts of a network that is based on a deep generative model of graphs.