2020.02.20.txt

==========New Papers==========
1, TITLE:       Globally optimal point set registration by joint symmetry plane fitting
http://arxiv.org/abs/2002.07988
AUTHORS:        Lan Hu ; Haomin Shi ; Laurent Kneip
HIGHLIGHT:      The present work proposes a solution to the challenging problem of registering two partial point sets of the same object with very limited overlap.

2, TITLE:       U-Bubble Model for Mixed Unit Interval Graphs and its Applications: The MaxCut Problem Revisited
http://arxiv.org/abs/2002.08311
AUTHORS:        Jan Kratochvíl ; Tomáš Masařík ; Jana Novotná
COMMENTS:       26 pages, 4 figures
HIGHLIGHT:      Interval graphs, intersection graphs of segments on a real line (intervals), play a key role in the study of algorithms and special structural properties.

3, TITLE:       Towards Query-Efficient Black-Box Adversary with Zeroth-Order Natural Gradient Descent
http://arxiv.org/abs/2002.07891
AUTHORS:        Pu Zhao ; Pin-Yu Chen ; Siyue Wang ; Xue Lin
COMMENTS:       accepted by AAAI 2020
HIGHLIGHT:      In this paper, we propose a zeroth-order natural gradient descent (ZO-NGD) method to design the adversarial attacks, which incorporates the zeroth-order gradient estimation technique catering to the black-box attack scenario and the second-order natural gradient descent to achieve higher query efficiency.

4, TITLE:       Generating Automatic Curricula via Self-Supervised Active Domain Randomization
http://arxiv.org/abs/2002.07911
AUTHORS:        Sharath Chandra Raparthy ; Bhairav Mehta ; Florian Golemo ; Liam Paull
HIGHLIGHT:      In this work, we build on the framework of self-play, allowing an agent to interact with itself in order to make progress on some unknown task.

5, TITLE:       ITeM: Independent Temporal Motifs to Summarize and Compare Temporal Networks
http://arxiv.org/abs/2002.08312
AUTHORS:        Sumit Purohit ; Lawrence B. Holder ; George Chin
HIGHLIGHT:      We present the Independent Temporal Motif (ITeM) to characterize temporal graphs from different domains.

6, TITLE:       Taxonomy of bio-inspired algorithms
http://arxiv.org/abs/2002.08136
AUTHORS:        Daniel Molina ; Javier Poyatos ; Javier Del Ser ; Salvador García ; Amir Hussain ; Francisco Herrera
COMMENTS:       63 pages, 6 figures
HIGHLIGHT:      In this work, we have reviewed more than two hundred nature-inspired and bio-inspired algorithms, and proposed two taxonomies that group them in categories ans subcategories, considering two different criteria.

7, TITLE:       TIES: Temporal Interaction Embeddings For Enhancing Social Media Integrity At Facebook
http://arxiv.org/abs/2002.07917
AUTHORS:        Nima Noorshams ; Saurabh Verma ; Aude Hofleitner
COMMENTS:       Submitted to KDD 2020 applied DS track
HIGHLIGHT:      In this paper, we present our efforts to protect various social media entities at Facebook from people who try to abuse our platform.

8, TITLE:       CodeBERT: A Pre-Trained Model for Programming and Natural Languages
http://arxiv.org/abs/2002.08155
AUTHORS:        Zhangyin Feng ; Daya Guo ; Duyu Tang ; Nan Duan ; Xiaocheng Feng ; Ming Gong ; Linjun Shou ; Bing Qin ; Ting Liu ; Daxin Jiang ; Ming Zhou
COMMENTS:       10 pages
HIGHLIGHT:      We present CodeBERT, a bimodal pre-trained model for programming language (PL) and nat-ural language (NL). Furthermore, to investigate what type of knowledge is learned in CodeBERT, we construct a dataset for NL-PL probing, and evaluate in a zero-shot setting where parameters of pre-trained models are fixed.

9, TITLE:       Multilogue-Net: A Context Aware RNN for Multi-modal Emotion Detection and Sentiment Analysis in Conversation
http://arxiv.org/abs/2002.08267
AUTHORS:        Aman Shenoy ; Ashish Sardana
COMMENTS:       10 pages, 4 figures, 6 tables
HIGHLIGHT:      In this paper, we propose a recurrent neural network architecture that attempts to take into account all the mentioned drawbacks, and keeps track of the context of the conversation, interlocutor states, and the emotions conveyed by the speakers in the conversation.

10, TITLE:      Compressing BERT: Studying the Effects of Weight Pruning on Transfer Learning
http://arxiv.org/abs/2002.08307
AUTHORS:        Mitchell A. Gordon ; Kevin Duh ; Nicholas Andrews
HIGHLIGHT:      Universal feature extractors, such as BERT for natural language processing and VGG for computer vision, have become effective methods for improving deep learning models without requiring more labeled data.

11, TITLE:      Interpreting Interpretations: Organizing Attribution Methods by Criteria
http://arxiv.org/abs/2002.07985
AUTHORS:        Zifan Wang ; PiotrPiotr Mardziel ; Anupam Datta ; Matt Fredrikson
HIGHLIGHT:      This paper introduces a new way to decompose the evaluation for attribution methods into two criteria: ordering and proportionality.

12, TITLE:      Knowledge Reconciliation of $n$-ary Relations
http://arxiv.org/abs/2002.08103
AUTHORS:        Pierre Monnin ; Miguel Couceiro ; Amedeo Napoli ; Adrien Coulet
HIGHLIGHT:      In this paper, we propose a rule-based methodology for the reconciliation of $n$-ary relations.

13, TITLE:      Lake Ice Monitoring with Webcams and Crowd-Sourced Images
http://arxiv.org/abs/2002.07875
AUTHORS:        Rajanie Prabha ; Manu Tom ; Mathias Rothermel ; Emmanuel Baltsavias ; Laura Leal-Taixe ; Konrad Schindler
HIGHLIGHT:      As part of the work, we introduce a new benchmark dataset of webcam images, Photi-LakeIce, from multiple cameras and two different winters, along with pixel-wise ground truth annotations.

14, TITLE:      Fawkes: Protecting Personal Privacy against Unauthorized Deep Learning Models
http://arxiv.org/abs/2002.08327
AUTHORS:        Shawn Shan ; Emily Wenger ; Jiayun Zhang ; Huiying Li ; Haitao Zheng ; Ben Y. Zhao
HIGHLIGHT:      In this paper, we propose Fawkes, a system that allow individuals to inoculate themselves against unauthorized facial recognition models.

15, TITLE:      Best-item Learning in Random Utility Models with Subset Choices
http://arxiv.org/abs/2002.07994
AUTHORS:        Aadirupa Saha ; Aditya Gopalan
COMMENTS:       Accepted to 23rd International Conference on Artificial Intelligence and Statistics (AISTATS), 2020
HIGHLIGHT:      We consider the problem of PAC learning the most valuable item from a pool of $n$ items using sequential, adaptively chosen plays of subsets of $k$ items, when, upon playing a subset, the learner receives relative feedback sampled according to a general Random Utility Model (RUM) with independent noise perturbations to the latent item utilities.

16, TITLE:      Optimal DG allocation and sizing in power system networks using swarm-based algorithms
http://arxiv.org/abs/2002.08089
AUTHORS:        Kayode Adetunji ; Ivan Hofsajer ; Ling Cheng
HIGHLIGHT:      In this paper, two swarm-based meta-heuristic algorithms, particle swarm optimization (PSO) and whale optimization algorithm (WOA) were developed to solve optimal placement and sizing of DG units in the quest for transmission network planning.

17, TITLE:      Holistic Specifications for Robust Programs
http://arxiv.org/abs/2002.08334
AUTHORS:        Sophia Drossopoulou ; James Noble ; Julian Mackay ; Susan Eisenbach
COMMENTS:       44 pages, 1 Table, 11 Figures
HIGHLIGHT:      In this paper we propose the language Chainmail for writing holistic specifications that focus on necessary conditions (as well as sufficient conditions).

18, TITLE:      A Differential-form Pullback Programming Language for Higher-order Reverse-mode Automatic Differentiation
http://arxiv.org/abs/2002.08241
AUTHORS:        Carol Mak ; Luke Ong
HIGHLIGHT:      Building on the observation that reverse-mode automatic differentiation (AD) -- a generalisation of backpropagation -- can naturally be expressed as pullbacks of differential 1-forms, we design a simple higher-order programming language with a first-class differential operator, and present a reduction strategy which exactly simulates reverse-mode AD.

19, TITLE:      MEM_GE: a new maximum entropy method for image reconstruction from solar X-ray visibilities
http://arxiv.org/abs/2002.07921
AUTHORS:        Paolo Massa ; Richard Schwartz ; A Kim Tolbert ; Anna Maria Massone ; Brian R Dennis ; Michele Piana ; Federico Benvenuto
HIGHLIGHT:      This paper introduces a new approach to Maximum Entropy based on the constrained minimization of a convex functional.

20, TITLE:      Quantum statistical query learning
http://arxiv.org/abs/2002.08240
AUTHORS:        Srinivasan Arunachalam ; Alex B. Grilo ; Henry Yuen
COMMENTS:       24 Pages
HIGHLIGHT:      We propose a learning model called the quantum statistical learning QSQ model, which extends the SQ learning model introduced by Kearns to the quantum setting.

21, TITLE:      A Structured Approach to Trustworthy Autonomous/Cognitive Systems
http://arxiv.org/abs/2002.08210
AUTHORS:        Henrik J. Putzer ; Ernest Wozniak
HIGHLIGHT:      This paper presents a framework to exactly fill this gap.

22, TITLE:      Unsupervised Temporal Feature Aggregation for Event Detection in Unstructured Sports Videos
http://arxiv.org/abs/2002.08097
AUTHORS:        Subhajit Chaudhury ; Daiki Kimura ; Phongtharin Vinayavekhin ; Asim Munawar ; Ryuki Tachibana ; Koji Ito ; Yuki Inaba ; Minoru Matsumoto ; Shuji Kidokoro ; Hiroki Ozaki
COMMENTS:       Accepted to IEEE International Symposium on Multimedia, 2019
HIGHLIGHT:      In this paper, we study the case of event detection in sports videos for unstructured environments with arbitrary camera angles.

23, TITLE:      Weakly-Supervised Semantic Segmentation by Iterative Affinity Learning
http://arxiv.org/abs/2002.08098
AUTHORS:        Xiang Wang ; Sifei Liu ; Huimin Ma ; Ming-Hsuan Yang
COMMENTS:       IJCV 2020
HIGHLIGHT:      In this paper, we propose an iterative algorithm to learn such pairwise relations, which consists of two branches, a unary segmentation network which learns the label probabilities for each pixel, and a pairwise affinity network which learns affinity matrix and refines the probability map generated from the unary network.

24, TITLE:      Feasibility of Video-based Sub-meter Localization on Resource-constrained Platforms
http://arxiv.org/abs/2002.08039
AUTHORS:        Abm Musa ; Jakob Eriksson
HIGHLIGHT:      In this paper, we study the feasibility of real-time video-based localization on resource-constrained platforms.

25, TITLE:      Meta Segmentation Network for Ultra-Resolution Medical Images
http://arxiv.org/abs/2002.08043
AUTHORS:        Tong Wu ; Yuan Xie ; Yanyun Qu ; Bicheng Dai ; Shuxin Chen
HIGHLIGHT:      In this paper, we propose Meta Segmentation Network (MSN) to solve this challenging problem.

26, TITLE:      Universal Domain Adaptation through Self Supervision
http://arxiv.org/abs/2002.07953
AUTHORS:        Kuniaki Saito ; Donghyun Kim ; Stan Sclaroff ; Kate Saenko
HIGHLIGHT:      We propose a more universally applicable domain adaptation approach that can handle arbitrary category shift, called Domain Adaptative Neighborhood Clustering via Entropy optimization (DANCE).

27, TITLE:      On-line non-overlapping camera calibration net
http://arxiv.org/abs/2002.08005
AUTHORS:        Zhao Fangda ; Toru Tamaki ; Takio Kurita ; Bisser Raytchev ; Kazufumi Kaneda
COMMENTS:       7 pages
HIGHLIGHT:      We propose an easy-to-use non-overlapping camera calibration method.

28, TITLE:      Supporting OpenMP 5.0 Tasks in hpxMP -- A study of an OpenMP implementation within Task Based Runtime Systems
http://arxiv.org/abs/2002.07970
AUTHORS:        Tianyi Zhang ; Shahrzad Shirzad ; Bibek Wagle ; Adrian S. Lemoine ; Patrick Diehl ; Hartmut Kaiser
HIGHLIGHT:      In this paper, we present the implementation of task features, e.g. taskgroup, task depend, and task_reduction, of the OpenMP 5.0 standard and optimization of the #pragma omp parallel for pragma.

29, TITLE:      Truly Tight-in-$Δ$ Bounds for Bipartite Maximal Matching and Variants
http://arxiv.org/abs/2002.08216
AUTHORS:        Sebastian Brandt ; Dennis Olivetti
HIGHLIGHT:      We provide truly tight bounds in $\Delta$ for the complexity of bipartite maximal matching and many natural variants, up to and including the additive constant.

30, TITLE:      Randomized Smoothing of All Shapes and Sizes
http://arxiv.org/abs/2002.08118
AUTHORS:        Greg Yang ; Tony Duan ; Edward Hu ; Hadi Salman ; Ilya Razenshteyn ; Jerry Li
COMMENTS:       9 pages main text, 40 pages total
HIGHLIGHT:      In this work we propose a novel framework for devising and analyzing randomized smoothing schemes, and validate its effectiveness in practice.

31, TITLE:      Hierarchical Quantized Autoencoders
http://arxiv.org/abs/2002.08111
AUTHORS:        Will Williams ; Sam Ringer ; Tom Ash ; John Hughes ; David MacLeod ; Jamie Dougherty
HIGHLIGHT:      We show that the combination of quantization and hierarchical latent structure aids likelihood-based image compression.

32, TITLE:      Neural Networks on Random Graphs
http://arxiv.org/abs/2002.08104
AUTHORS:        Romuald A. Janik ; Aleksandra Nowak
HIGHLIGHT:      Apart from the classical random graph families including random, scale-free and small world graphs, we introduced a novel and flexible algorithm for directly generating random directed acyclic graphs (DAG) and studied a class of graphs derived from functional resting state fMRI networks.

33, TITLE:      Transfer Learning for Abstractive Summarization at Controllable Budgets
http://arxiv.org/abs/2002.07845
AUTHORS:        Ritesh Sarkhel ; Moniba Keymanesh ; Arnab Nandi ; Srinivasan Parthasarathy
COMMENTS:       9 pages, 5 figures
HIGHLIGHT:      We propose MLS, an end-to-end framework to generate abstractive summaries with limited training data at arbitrary compression budgets.

34, TITLE:      The Microsoft Toolkit of Multi-Task Deep Neural Networks for Natural Language Understanding
http://arxiv.org/abs/2002.07972
AUTHORS:        Xiaodong Liu ; Yu Wang ; Jianshu Ji ; Hao Cheng ; Xueyun Zhu ; Emmanuel Awa ; Pengcheng He ; Weizhu Chen ; Hoifung Poon ; Guihong Cao ; Jianfeng Gao
COMMENTS:       9 pages, 3 figures and 3 tables
HIGHLIGHT:      We present MT-DNN, an open-source natural language understanding (NLU) toolkit that makes it easy for researchers and developers to train customized deep learning models.

35, TITLE:      Toward Making the Most of Context in Neural Machine Translation
http://arxiv.org/abs/2002.07982
AUTHORS:        Zaixiang Zheng ; Xiang Yue ; Shujian Huang ; Jiajun Chen ; Alexandra Birch
COMMENTS:       Submitted to a conference
HIGHLIGHT:      We argue that previous research did not make a clear use of the global context, and propose a new document-level NMT framework that deliberately models the local context of each sentence with the awareness of the global context of the document in both source and target languages.

36, TITLE:      Studying the Effects of Cognitive Biases in Evaluation of Conversational Agents
http://arxiv.org/abs/2002.07927
AUTHORS:        Sashank Santhanam ; Alireza Karduni ; Samira Shaikh
COMMENTS:       Accepted at CHI 2020
HIGHLIGHT:      Researchers typically evaluate the output of their models through crowdsourced judgments, but there are no established best practices for conducting such studies.

37, TITLE:      A note on the explicit constructions of tree codes over polylogarithmic-sized alphabet
http://arxiv.org/abs/2002.08231
AUTHORS:        Siddharth Bhandari ; Prahladh Harsha
HIGHLIGHT:      In this short note, we give a unified and simpler presentation of Pudl\'{a}k and Cohen-Haeupler-Schulman's constructions.

38, TITLE:      Curriculum in Gradient-Based Meta-Reinforcement Learning
http://arxiv.org/abs/2002.07956
AUTHORS:        Bhairav Mehta ; Tristan Deleu ; Sharath Chandra Raparthy ; Chris J. Pal ; Liam Paull
COMMENTS:       11 pages, 10 figures
HIGHLIGHT:      In this work, we begin by highlighting intriguing failure cases of gradient-based meta-RL and show that task distributions can wildly affect algorithmic outputs, stability, and performance.

39, TITLE:      Block Switching: A Stochastic Approach for Deep Learning Security
http://arxiv.org/abs/2002.07920
AUTHORS:        Xiao Wang ; Siyue Wang ; Pin-Yu Chen ; Xue Lin ; Peter Chin
COMMENTS:       Accepted by AdvML19: Workshop on Adversarial Learning Methods for Machine Learning and Data Mining at KDD, Anchorage, Alaska, USA, August 5th, 2019, 5 pages
HIGHLIGHT:      In this paper, we introduce Block Switching (BS), a defense strategy against adversarial attacks based on stochasticity.

40, TITLE:      Short-Term Traffic Flow Prediction Using Variational LSTM Networks
http://arxiv.org/abs/2002.07922
AUTHORS:        Mehrdad Farahani ; Marzieh Farahani ; Mohammad Manthouri ; Okyay Kaynak
COMMENTS:       18 pages, 13 figures
HIGHLIGHT:      The purpose of this research is to suggest a forecasting model for traffic flow by using deep learning techniques based on historical data in the Intelligent Transportation Systems area.

41, TITLE:      Using AI for Mitigating the Impact of Network Delay in Cloud-based Intelligent Traffic Signal Control
http://arxiv.org/abs/2002.08303
AUTHORS:        Rusheng Zhang ; Xinze Zhou ; Ozan K. Tonguz
COMMENTS:       6 pages, 3 figures, submitted to IEEE BlackseaComm 2020
HIGHLIGHT:      In this paper, we introduce a new traffic signal control algorithm based on reinforcement learning, which performs well even under severe network delay.

42, TITLE:      LocoGAN -- Locally Convolutional GAN
http://arxiv.org/abs/2002.07897
AUTHORS:        Łukasz Struski ; Szymon Knop ; Jacek Tabor ; Wiktor Daniec ; Przemysław Spurek
HIGHLIGHT:      In the paper we construct a fully convolutional GAN model: LocoGAN, which latent space is given by noise-like images of possibly different resolutions.

43, TITLE:      The set of hyperbolic equilibria and of invertible zeros on the unit ball is computable
http://arxiv.org/abs/2002.08199
AUTHORS:        Daniel S. Graça ; Ning Zhong
HIGHLIGHT:      In this note, we construct an algorithm that, on input of a description of a structurally stable planar dynamical flow defined on the unit disk, outputs the exact number of the (hyperbolic) equilibrium points as well the locations of all equilibriums with arbitrary precision.

44, TITLE:      When Radiology Report Generation Meets Knowledge Graph
http://arxiv.org/abs/2002.08277
AUTHORS:        Yixiao Zhang ; Xiaosong Wang ; Ziyue Xu ; Qihang Yu ; Alan Yuille ; Daguang Xu
HIGHLIGHT:      Based on these concerns, we propose to utilize a pre-constructed graph embedding module (modeled with a graph convolutional neural network) on multiple disease findings to assist the generation of reports in this work.

45, TITLE:      VQA-LOL: Visual Question Answering under the Lens of Logic
http://arxiv.org/abs/2002.08325
AUTHORS:        Tejas Gokhale ; Pratyay Banerjee ; Chitta Baral ; Yezhou Yang
HIGHLIGHT:      In this paper, we investigate visual question answering (VQA) through the lens of logical transformation and posit that systems that seek to answer questions about images must be robust to these transformations of the question.

46, TITLE:      Extracting Semantic Indoor Maps from Occupancy Grids
http://arxiv.org/abs/2002.08348
AUTHORS:        Ziyuan Liu ; Georg von Wichert
HIGHLIGHT:      In this paper we focus on the semantic mapping of indoor environments.

47, TITLE:      Realtime Index-Free Single Source SimRank Processing on Web-Scale Graphs
http://arxiv.org/abs/2002.08082
AUTHORS:        Jieming Shi ; Tianyuan Jin ; Renchi Yang ; Xiaokui Xiao ; Yin Yang
COMMENTS:       To appear in PVLDB 2020
HIGHLIGHT:      Motivated by this, we propose SimPush, a novel algorithm that answers single source SimRank queries without any pre-computation, and at the same time achieves significantly higher query processing speed than even the fastest known index-based solutions.

48, TITLE:      Variational Encoder-based Reliable Classification
http://arxiv.org/abs/2002.08289
AUTHORS:        Chitresh Bhushan ; Zhaoyuan Yang ; Nurali Virani ; Naresh Iyer
COMMENTS:       7 pages, 6 figures
HIGHLIGHT:      To provide reliability, we propose an Epistemic Classifier (EC) that can provide justification of its belief using support from the training dataset as well as quality of reconstruction.

49, TITLE:      Non-Autoregressive Dialog State Tracking
http://arxiv.org/abs/2002.08024
AUTHORS:        Hung Le ; Richard Socher ; Steven C. H. Hoi
COMMENTS:       Accepted at ICLR 2020
HIGHLIGHT:      In this paper, we propose a novel framework of Non-Autoregressive Dialog State Tracking (NADST) which can factor in potential dependencies among domains and slots to optimize the models towards better prediction of dialogue states as a complete set rather than separate slots.

50, TITLE:      Rnn-transducer with language bias for end-to-end Mandarin-English code-switching speech recognition
http://arxiv.org/abs/2002.08126
AUTHORS:        Shuai Zhang ; Jiangyan Yi ; Zhengkun Tian ; Jianhua Tao ; Ye Bai
HIGHLIGHT:      In this work, we propose an improved recurrent neural network transducer (RNN-T) model with language bias to alleviate the problem.

51, TITLE:      Hierarchical models vs. transfer learning for document-level sentiment classification
http://arxiv.org/abs/2002.08131
AUTHORS:        Jeremy Barnes ; Vinit Ravishankar ; Lilja Øvrelid ; Erik Velldal
COMMENTS:       4 pages, 2 figures
HIGHLIGHT:      In this work we empirically compare hierarchical models and transfer learning for document-level sentiment classification.

52, TITLE:      LAMBERT: Layout-Aware language Modeling using BERT for information extraction
http://arxiv.org/abs/2002.08087
AUTHORS:        Łukasz Garncarek ; Rafał Powalski ; Tomasz Stanisławek ; Bartosz Topolski ; Piotr Halama ; Filip Graliński
COMMENTS:       v1: 9 pages; work in progress; this version of the paper was submitted to review on Dec 10, 2019, and subsequently withdrawn on Feb 17, 2020
HIGHLIGHT:      In this paper we introduce a novel approach to the problem of understanding documents where the local semantics is influenced by non-trivial layout.

53, TITLE:      Efficient Deep Reinforcement Learning through Policy Transfer
http://arxiv.org/abs/2002.08037
AUTHORS:        Tianpei Yang ; Jianye Hao ; Zhaopeng Meng ; Zongzhang Zhang ; Weixun Wang ; Yujing Hu ; Yingfeng Cheng ; Changjie Fan ; Zhaodong Wang ; Jiajie Peng
COMMENTS:       Accepted by AAMAS'2020 as an EXTENDED ABSTRACT
HIGHLIGHT:      In this paper, we propose a novel Policy Transfer Framework (PTF) to accelerate RL by taking advantage of this idea.

54, TITLE:      Enlarging Discriminative Power by Adding an Extra Class in Unsupervised Domain Adaptation
http://arxiv.org/abs/2002.08041
AUTHORS:        Hai H. Tran ; Sumyeong Ahn ; Taeyoung Lee ; Yung Yi
HIGHLIGHT:      In this paper, we study the problem of unsupervised domain adaptation that aims at obtaining a prediction model for the target domain using labeled data from the source domain and unlabeled data from the target domain.

55, TITLE:      Dissecting Neural ODEs
http://arxiv.org/abs/2002.08071
AUTHORS:        Stefano Massaroli ; Michael Poli ; Jinkyoo Park ; Atsushi Yamashita ; Hajime Asama
HIGHLIGHT:      In this work, we "open the box" and offer a system-theoretic perspective, including state augmentation strategies and robustness, with the aim of clarifying the influence of several design choices on the underlying dynamics.

56, TITLE:      A Fixed point view: A Model-Based Clustering Framework
http://arxiv.org/abs/2002.08032
AUTHORS:        Jianhao Ding ; Lansheng Han
COMMENTS:       10 pages, 2 figures
HIGHLIGHT:      Based on the view of fixed point, this paper restates the model-based clustering and proposes a unified clustering framework.

57, TITLE:      The complexity of knapsack problems in wreath products
http://arxiv.org/abs/2002.08086
AUTHORS:        Michael Figelius ; Moses Ganardi ; Markus Lohrey ; Georg Zetzsche
HIGHLIGHT:      For the knapsack problem we show $\mathsf{NP}$-completeness for iterated wreath products of free abelian groups and hence free solvable groups.

58, TITLE:      Weakly Supervised Semantic Segmentation of Satellite Images for Land Cover Mapping -- Challenges and Opportunities
http://arxiv.org/abs/2002.08254
AUTHORS:        Michael Schmitt ; Jonathan Prexl ; Patrick Ebel ; Lukas Liebel ; Xiao Xiang Zhu
COMMENTS:       8 pages, 6 figures
HIGHLIGHT:      Therefore, this paper seeks to make a case for the application of weakly supervised learning strategies to get the most out of available data sources and achieve progress in high-resolution large-scale land cover mapping.

59, TITLE:      DeFraudNet:End2End Fingerprint Spoof Detection using Patch Level Attention
http://arxiv.org/abs/2002.08214
AUTHORS:        B. V. S Anusha ; Sayan Banerjee ; Subhasis Chaudhuri
COMMENTS:       WACV 2020
HIGHLIGHT:      This paper proposes a novel method for fingerprint spoof detection using both global and local fingerprint feature descriptors.

60, TITLE:      Three-Stream Fusion Network for First-Person Interaction Recognition
http://arxiv.org/abs/2002.08219
AUTHORS:        Ye-Ji Kim ; Dong-Gyu Lee ; Seong-Whan Lee
COMMENTS:       30 pages, 9 figures
HIGHLIGHT:      For human interaction recognition from a first-person viewpoint, this paper proposes a three-stream fusion network with two main parts: three-stream architecture and three-stream correlation fusion.

61, TITLE:      Model-Agnostic Structured Sparsification with Learnable Channel Shuffle
http://arxiv.org/abs/2002.08127
AUTHORS:        Xin-Yu Zhang ; Kai Zhao ; Taihong Xiao ; Ming-Ming Cheng ; Ming-Hsuan Yang
HIGHLIGHT:      To this end, we propose a model-agnostic structured sparsification method for efficient network compression.

62, TITLE:      siaNMS: Non-Maximum Suppression with Siamese Networks for Multi-Camera 3D Object Detection
http://arxiv.org/abs/2002.08239
AUTHORS:        Irene Cortes ; Jorge Beltran ; Arturo de la Escalera ; Fernando Garcia
COMMENTS:       Submitted to IEEE Intelligent Vehicles Symposium 2020 (IV2020)
HIGHLIGHT:      In this work, a siamese network is integrated into the pipeline of a well-known 3D object detector approach to suppress duplicate proposals coming from different cameras via re-identification.

63, TITLE:      SPORES: Sum-Product Optimization via Relational Equality Saturation for Large Scale Linear Algebra
http://arxiv.org/abs/2002.07951
AUTHORS:        Yisu Remy Wang ; Shana Hutchison ; Jonathan Leang ; Bill Howe ; Dan Suciu
HIGHLIGHT:      We introduce a general optimization technique for LA expressions, by converting the LA expressions into Relational Algebra (RA) expressions, optimizing the latter, then converting the result back to (optimized) LA expressions.

64, TITLE:      BB_Evac: Fast Location-Sensitive Behavior-Based Building Evacuation
http://arxiv.org/abs/2002.08114
AUTHORS:        Subhra Mazumdar ; Arindam Pal ; Francesco Parisi ; V. S. Subrahmanian
HIGHLIGHT:      In this paper, we present a formal definition of a behavior-based evacuation problem (BBEP) in which a human behavior model is taken into account when planning an evacuation.

65, TITLE:      Learning Global Transparent Models from Local Contrastive Explanations
http://arxiv.org/abs/2002.08247
AUTHORS:        Tejaswini Pedapati ; Avinash Balakrishnan ; Karthikeyan Shanmugam ; Amit Dhurandhar
HIGHLIGHT:      In this work, we explore the question: Can we produce a transparent global model that is consistent with/derivable from local explanations?

66, TITLE:      SYMOG: learning symmetric mixture of Gaussian modes for improved fixed-point quantization
http://arxiv.org/abs/2002.08204
AUTHORS:        Lukas Enderich ; Fabian Timm ; Wolfram Burgard
COMMENTS:       Preprint submitted to Neurocomputing
HIGHLIGHT:      We propose SYMOG (symmetric mixture of Gaussian modes), which significantly decreases the complexity of DNNs through low-bit fixed-point quantization.



==========Updates to Previous Papers==========
1, TITLE:       Fine-grained complexity of the graph homomorphism problem for bounded-treewidth graphs
http://arxiv.org/abs/1906.08371
AUTHORS:        Karolina Okrasa ; Paweł Rzążewski
COMMENTS:       An extended abstract of this paper appeared on SODA 2020
HIGHLIGHT:      In this paper we are interested in the complexity of the problem, parameterized by the treewidth of the input graph $G$.

2, TITLE:       Inapproximability and parameterized results for the target set selection problem
http://arxiv.org/abs/1812.01482
AUTHORS:        Suman Banerjee ; Rogers Mathew ; Fahad Panolan
COMMENTS:       10 pages
HIGHLIGHT:      Under the non-progressive diffusion model, we have the following results on the TSS Problem: We show that the TSS Problem on bipartite graphs does not admit an approximation algorithm with a performance guarantee asymptotically better than $O(\log n_{min})$, where $n_{min}$ is the cardinality of the smaller bipartition, unless $P=NP$.

3, TITLE:       On the Degree of Boolean Functions as Polynomials over $\mathbb{Z}_m$
http://arxiv.org/abs/1910.12458
AUTHORS:        Xiaoming Sun ; Yuan Sun ; Jiaheng Wang ; Kewen Wu ; Zhiyu Xia ; Yufan Zheng
COMMENTS:       19 pages
HIGHLIGHT:      In this paper, we investigate the lower bound of modulo-$m$ degree of Boolean functions.

4, TITLE:       Entangled simultaneity versus classical interactivity in communication complexity
http://arxiv.org/abs/1602.05059
AUTHORS:        Dmitry Gavinsky
HIGHLIGHT:      In this work we answer the latter question affirmatively and present a partial function Shape, which can be computed by a protocol sending entangled simultaneous messages of poly-logarithmic size, and whose classical two-way complexity is lower bounded by a polynomial.

5, TITLE:       Robustness via Deep Low-Rank Representations
http://arxiv.org/abs/1804.07090
AUTHORS:        Amartya Sanyal ; Varun Kanade ; Philip H. S. Torr ; Puneet K. Dokania
HIGHLIGHT:      To achieve low dimensionality of learned representations, we propose an easy-to-use, end-to-end trainable, low-rank regularizer (LR) that can be applied to any intermediate layer representation of a DNN.

6, TITLE:       Isotropic Maximization Loss and Entropic Score: Fast, Accurate, Scalable, Unexposed, Turnkey, and Native Neural Networks Out-of-Distribution Detection
http://arxiv.org/abs/1908.05569
AUTHORS:        David Macêdo ; Tsang Ing Ren ; Cleber Zanchettin ; Adriano L. I. Oliveira ; Alain Tapp ; Teresa Ludermir
HIGHLIGHT:      In this paper, we argue that the uncertainty in neural networks is mainly due to SoftMax loss anisotropy.

7, TITLE:       Consistency Regularization for Generative Adversarial Networks
http://arxiv.org/abs/1910.12027
AUTHORS:        Han Zhang ; Zizhao Zhang ; Augustus Odena ; Honglak Lee
COMMENTS:       ICLR2020
HIGHLIGHT:      In this work, we propose a simple, effective training stabilizer based on the notion of consistency regularization---a popular technique in the semi-supervised learning literature.

8, TITLE:       Meta-Learning via Learned Loss
http://arxiv.org/abs/1906.05374
AUTHORS:        Sarah Bechtle ; Artem Molchanov ; Yevgen Chebotar ; Edward Grefenstette ; Ludovic Righetti ; Gaurav Sukhatme ; Franziska Meier
HIGHLIGHT:      Concretely, we present a meta-learning method for learning parametric loss functions that can generalize across different tasks and model architectures.

9, TITLE:       ALGAMES: A Fast Solver for Constrained Dynamic Games
http://arxiv.org/abs/1910.09713
AUTHORS:        Simon Le Cleac'h ; Mac Schwager ; Zachary Manchester
COMMENTS:       10 pages, 8 figures, submitted to Robotics: Science and Systems Conference (RSS) 2020
HIGHLIGHT:      This paper introduces ALGAMES (Augmented Lagrangian GAME-theoretic Solver), a solver that handles trajectory optimization problems with multiple actors and general nonlinear state and input constraints.

10, TITLE:      Speech Corpus of Ainu Folklore and End-to-end Speech Recognition for Ainu Language
http://arxiv.org/abs/2002.06675
AUTHORS:        Kohei Matsuura ; Sei Ueno ; Masato Mimura ; Shinsuke Sakai ; Tatsuya Kawahara
COMMENTS:       ver. 2
HIGHLIGHT:      In this paper, we report speech corpus development and the structure and performance of end-to-end ASR for Ainu.

11, TITLE:      Learn to Explain Efficiently via Neural Logic Inductive Learning
http://arxiv.org/abs/1910.02481
AUTHORS:        Yuan Yang ; Le Song
HIGHLIGHT:      In this work, we study the learning to explain problem in the scope of inductive logic programming (ILP).

12, TITLE:      Budgeted Training: Rethinking Deep Neural Network Training Under Resource Constraints
http://arxiv.org/abs/1905.04753
AUTHORS:        Mengtian Li ; Ersin Yumer ; Deva Ramanan
COMMENTS:       ICLR 2020. Project page with code is at http://www.cs.cmu.edu/~mengtial/proj/budgetnn/
HIGHLIGHT:      Therefore, we introduce a formal setting for studying training under the non-asymptotic, resource-constrained regime, i.e., budgeted training.

13, TITLE:      Representation Quality Explains Adversarial Attacks
http://arxiv.org/abs/1906.06627
AUTHORS:        Danilo Vasconcellos Vargas ; Shashank Kotyan ; Moe Matsuki
HIGHLIGHT:      Here, we propose a way to evaluate the representation quality of neural networks using a novel type of zero-shot test, entitled Raw Zero-Shot.

14, TITLE:      Principles alone cannot guarantee ethical AI
http://arxiv.org/abs/1906.06668
AUTHORS:        Brent Mittelstadt
COMMENTS:       A previous, pre-print version of this paper was entitled 'AI Ethics - Too Principled to Fail?'
HIGHLIGHT:      Principles alone cannot guarantee ethical AI

15, TITLE:      AI-based Pilgrim Detection using Convolutional Neural Networks
http://arxiv.org/abs/1911.07509
AUTHORS:        Marwa Ben Jabra ; Adel Ammar ; Anis Koubaa ; Omar Cheikhrouhou ; Habib Hamam
COMMENTS:       Accepted in ATSIP'2020
HIGHLIGHT:      To address this issue, we propose to use artificial intelligence technique based on deep learning and convolution neural networks to detect and identify Pilgrims and their features. For this purpose, we built a comprehensive dataset for the detection of pilgrims and their genders.

16, TITLE:      Non-local U-Net for Biomedical Image Segmentation
http://arxiv.org/abs/1812.04103
AUTHORS:        Zhengyang Wang ; Na Zou ; Dinggang Shen ; Shuiwang Ji
COMMENTS:       In Proceedings of the 34th AAAI Conference on Artificial Intelligence (AAAI), 2019
HIGHLIGHT:      In this work, we propose the non-local U-Nets, which are equipped with flexible global aggregation blocks, for biomedical image segmentation.

17, TITLE:      CNN 101: Interactive Visual Learning for Convolutional Neural Networks
http://arxiv.org/abs/2001.02004
AUTHORS:        Zijie J. Wang ; Robert Turko ; Omar Shaikh ; Haekyu Park ; Nilaksh Das ; Fred Hohman ; Minsuk Kahng ; Duen Horng Chau
COMMENTS:       CHI'20 Late-Breaking Work (April 25-30, 2020), 7 pages, 3 figures
HIGHLIGHT:      We present our ongoing work, CNN 101, an interactive visualization system for explaining and teaching convolutional neural networks.

18, TITLE:      Evolving Robust Neural Architectures to Defend from Adversarial Attacks
http://arxiv.org/abs/1906.11667
AUTHORS:        Danilo Vasconcellos Vargas ; Shashank Kotyan
HIGHLIGHT:      Here, we propose to use adversarial attacks as a function evaluation to automatically search for neural architectures that can resist such attacks.

19, TITLE:      Environmental drivers of systematicity and generalization in a situated agent
http://arxiv.org/abs/1910.00571
AUTHORS:        Felix Hill ; Andrew Lampinen ; Rosalia Schneider ; Stephen Clark ; Matthew Botvinick ; James L. McClelland ; Adam Santoro
HIGHLIGHT:      Here, we consider tests of out-of-sample generalisation that require an agent to respond to never-seen-before instructions by manipulating and positioning objects in a 3D Unity simulated room.

20, TITLE:      Human Action Performance using Deep Neuro-Fuzzy Recurrent Attention Model
http://arxiv.org/abs/2001.10953
AUTHORS:        Nihar Bendre ; Nima Ebadi ; Paul Rad
COMMENTS:       15 pages, 6 figures, Under review for IEEE Access Journal
HIGHLIGHT:      To remedy this uncertainty, in this paper, we coupled fuzzy logic rules with the neural-based action recognition model to index the intensity of the action as intense or mild.

21, TITLE:      Implementing Dynamic Answer Set Programming
http://arxiv.org/abs/2002.06916
AUTHORS:        Pedro Cabalar ; Martín Diéguez ; Torsten Schaub ; François Laferrière
HIGHLIGHT:      We introduce an implementation of an extension of Answer Set Programming (ASP) with language constructs from dynamic (and temporal) logic that provides an expressive computational framework for modeling dynamic applications.

22, TITLE:      OpenBioLink: A benchmarking framework for large-scale biomedical link prediction
http://arxiv.org/abs/1912.04616
AUTHORS:        Anna Breit ; Simon Ott ; Asan Agibetov ; Matthias Samwald
HIGHLIGHT:      With OpenBioLink, we introduce a large-scale, high-quality and highly challenging biomedical link prediction benchmark to transparently and reproducibly evaluate such algorithms.

23, TITLE:      The Next Decade in AI: Four Steps Towards Robust Artificial Intelligence
http://arxiv.org/abs/2002.06177
AUTHORS:        Gary Marcus
COMMENTS:       5 figures
HIGHLIGHT:      In contrast, I propose a hybrid, knowledge-driven, reasoning-based approach, centered around cognitive models, that could provide the substrate for a richer, more robust AI than is currently possible.

24, TITLE:      Assistive Relative Pose Estimation for On-orbit Assembly using Convolutional Neural Networks
http://arxiv.org/abs/2001.10673
AUTHORS:        Shubham Sonawani ; Ryan Alimo ; Renaud Detry ; Daniel Jeong ; Andrew Hess ; Heni Ben Amor
HIGHLIGHT:      In this paper, a convolutional neural network is leveraged to uniquely determine the translation and rotation of an object of interest relative to the camera.

25, TITLE:      Fast Efficient Object Detection Using Selective Attention
http://arxiv.org/abs/1811.07502
AUTHORS:        Shivanthan Yohanandan ; Andy Song ; Adrian G. Dyer ; Angela Faragasso ; Subhrajit Roy ; Dacheng Tao
COMMENTS:       Retraction due to significant oversight
HIGHLIGHT:      To that end, we leverage this knowledge to design a novel region proposal network and empirically show that it achieves high object detection performance on the COCO dataset.

26, TITLE:      Analysis Of Multi Field Of View Cnn And Attention Cnn On H&E Stained Whole-slide Images On Hepatocellular Carcinoma
http://arxiv.org/abs/2002.04836
AUTHORS:        Mehmet Burak Sayıcı ; Rikiya Yamashita ; Jeanne Shen
COMMENTS:       This paper has been withdrawn by the authors due to need for heavy revise
HIGHLIGHT:      In this work, the effect of tile size on performance for classification problem is analysed.

27, TITLE:      HyPar-Flow: Exploiting MPI and Keras for Scalable Hybrid-Parallel DNN Training using TensorFlow
http://arxiv.org/abs/1911.05146
AUTHORS:        Ammar Ahmad Awan ; Arpan Jain ; Quentin Anthony ; Hari Subramoni ; Dhabaleswar K. Panda
COMMENTS:       18 pages, 10 figures, Accepted, to be presented at ISC '20
HIGHLIGHT:      We create an internal distributed representation of the user-provided Keras model, utilize TF's Eager execution features for distributed forward/back-propagation across processes, exploit pipelining to improve performance and leverage efficient MPI primitives for scalable communication.

28, TITLE:      Regularization Matters in Policy Optimization
http://arxiv.org/abs/1910.09191
AUTHORS:        Zhuang Liu ; Xuanlin Li ; Bingyi Kang ; Trevor Darrell
COMMENTS:       More analytic experiments and evaluation metrics added on last version. Code link: https://github.com/xuanlinli17/po-rl-regularization
HIGHLIGHT:      In this work, we present the first comprehensive study of regularization techniques with multiple policy optimization algorithms on continuous control tasks.

29, TITLE:      Learning Bijective Feature Maps for Linear ICA
http://arxiv.org/abs/2002.07766
AUTHORS:        Alexander Camuto ; Matthew Willetts ; Brooks Paige ; Chris Holmes ; Stephen Roberts
COMMENTS:       8 pages
HIGHLIGHT:      Here we develop a method that jointly learns a linear independent component analysis (ICA) model with non-linear bijective feature maps.

30, TITLE:      Reconfigurable Interaction for MAS Modelling
http://arxiv.org/abs/1906.10793
AUTHORS:        Yehia Abd Alrahman ; Giuseppe Perelli ; Nir Piterman
COMMENTS:       This is a final and revised version. This research is funded by the ERC consolidator grant D-SynMA under the European Union's Horizon 2020 research and innovation programme (grant agreement No 772459)
HIGHLIGHT:      We propose a formalism to model and reason about multi-agent systems.

31, TITLE:      A generic imperative language for polynomial time
http://arxiv.org/abs/1911.04026
AUTHORS:        Daniel Leivant
COMMENTS:       18 pages, submitted to a conference
HIGHLIGHT:      We introduce a new approach to ramification which, among other benefits, adapts readily to fully general imperative programming.

32, TITLE:      FreeLB: Enhanced Adversarial Training for Natural Language Understanding
http://arxiv.org/abs/1909.11764
AUTHORS:        Chen Zhu ; Yu Cheng ; Zhe Gan ; Siqi Sun ; Tom Goldstein ; Jingjing Liu
COMMENTS:       ICLR 2020
HIGHLIGHT:      In this work, we propose a novel adversarial training algorithm, FreeLB, that promotes higher invariance in the embedding space, by adding adversarial perturbations to word embeddings and minimizing the resultant adversarial risk inside different regions around input samples.

33, TITLE:      Optimal Error Pseudodistributions for Read-Once Branching Programs
http://arxiv.org/abs/2002.07208
AUTHORS:        Eshan Chattopadhyay ; Jyun-Jie Liao
HIGHLIGHT:      In this work, we construct a PRPD with seed length $$O(\log n\cdot \log (nw)\cdot \log\log(nw)+\log(1/\varepsilon)).

34, TITLE:      A Simplified Fully Quantized Transformer for End-to-end Speech Recognition
http://arxiv.org/abs/1911.03604
AUTHORS:        Alex Bie ; Bharat Venkitesh ; Joao Monteiro ; Md. Akmal Haidar ; Mehdi Rezagholizadeh
COMMENTS:       Changed title Added references
HIGHLIGHT:      That being said, in this paper, we work on simplifying and compressing Transformer-based encoder-decoder architectures for the end-to-end ASR task.

35, TITLE:      Search problems in algebraic complexity, GCT, and hardness of generator for invariant rings
http://arxiv.org/abs/1910.01251
AUTHORS:        Ankit Garg ; Christian Ikenmeyer ; Visu Makam ; Rafael Oliveira ; Michael Walter ; Avi Wigderson
COMMENTS:       17 pages
HIGHLIGHT:      We consider the problem of computing succinct encodings of lists of generators for invariant rings for group actions.

36, TITLE:      Solving Vertex Cover in Polynomial Time on Hyperbolic Random Graphs
http://arxiv.org/abs/1904.12503
AUTHORS:        Thomas Bläsius ; Philipp Fischbeck ; Tobias Friedrich ; Maximilian Katzmann
HIGHLIGHT:      The VertexCover problem is proven to be computationally hard in different ways: It is NP-complete to find an optimal solution and even NP-hard to find an approximation with reasonable factors.

37, TITLE:      Generalization of Reinforcement Learners with Working and Episodic Memory
http://arxiv.org/abs/1910.13406
AUTHORS:        Meire Fortunato ; Melissa Tan ; Ryan Faulkner ; Steven Hansen ; Adrià Puigdomènech Badia ; Gavin Buttimore ; Charlie Deck ; Joel Z Leibo ; Charles Blundell
COMMENTS:       NeurIPS 2019. Equal contribution of first 4 authors
HIGHLIGHT:      In this paper, we aim to develop a comprehensive methodology to test different kinds of memory in an agent and assess how well the agent can apply what it learns in training to a holdout set that differs from the training set along dimensions that we suggest are relevant for evaluating memory-specific generalization. To that end, we first construct a diverse set of memory tasks that allow us to evaluate test-time generalization across multiple dimensions.

38, TITLE:      Constant Curvature Graph Convolutional Networks
http://arxiv.org/abs/1911.05076
AUTHORS:        Gregor Bachmann ; Gary Bécigneul ; Octavian-Eugen Ganea
HIGHLIGHT:      Here, we bridge this gap by proposing mathematically grounded generalizations of graph convolutional networks (GCN) to (products of) constant curvature spaces.

39, TITLE:      A Single RGB Camera Based Gait Analysis with a Mobile Tele-Robot for Healthcare
http://arxiv.org/abs/2002.04700
AUTHORS:        Ziyang Wang
HIGHLIGHT:      The purpose of this work is twofold, the software focuses on the analysis of gait, which is widely adopted for joint correction and assessing any lower limb or spinal problem.

40, TITLE:      Deep compositional robotic planners that follow natural language commands
http://arxiv.org/abs/2002.05201
AUTHORS:        Yen-Ling Kuo ; Boris Katz ; Andrei Barbu
COMMENTS:       Accepted in ICRA 2020
HIGHLIGHT:      We demonstrate how a sampling-based robotic planner can be augmented to learn to understand a sequence of natural language commands in a continuous configuration space to move and manipulate objects.

41, TITLE:      Machine Learning for Motor Learning: EEG-based Continuous Assessment of Cognitive Engagement for Adaptive Rehabilitation Robots
http://arxiv.org/abs/2002.07541
AUTHORS:        Neelesh Kumar ; Konstantinos P. Michmizos
COMMENTS:       6 pages, 6 figures, 1 table
HIGHLIGHT:      Here, we propose an end-to-end computational framework that assesses CE in real-time, using electroencephalography (EEG) signals as objective measurements.

42, TITLE:      Modular Inference of Linear Types for Multiplicity-Annotated Arrows
http://arxiv.org/abs/1911.00268
AUTHORS:        Kazutaka Matsuda
COMMENTS:       The full version of our paper to appear in ESOP 2020
HIGHLIGHT:      In this paper, based on OutsideIn(X) [Vytiniotis et al., 2011], we propose an inference system for a rank 1 qualified-typed variant of $\lambda^q_\to$, which infers principal types.

43, TITLE:      DeepTemporalSeg: Temporally Consistent Semantic Segmentation of 3D LiDAR Scans
http://arxiv.org/abs/1906.06962
AUTHORS:        Ayush Dewan ; Wolfram Burgard
COMMENTS:       Accepted for ICRA-2020. Code and dataset available at https://github.com/ayushais/DBLiDARNet
HIGHLIGHT:      In this paper, we propose a deep convolutional neural network (DCNN) for the semantic segmentation of a LiDAR scan into the classes car, pedestrian or bicyclist.

44, TITLE:      Learning End-to-End Lossy Image Compression: A Benchmark
http://arxiv.org/abs/2002.03711
AUTHORS:        Yueyu Hu ; Wenhan Yang ; Zhan Ma ; Jiaying Liu
COMMENTS:       https://huzi96.github.io/compression-bench.html
HIGHLIGHT:      We describe milestones in cutting-edge learned image-compression methods, review a broad range of existing works, and provide insights into their historical development routes.

45, TITLE:      MS-Net: Multi-Site Network for Improving Prostate Segmentation with Heterogeneous MRI Data
http://arxiv.org/abs/2002.03366
AUTHORS:        Quande Liu ; Qi Dou ; Lequan Yu ; Pheng Ann Heng
COMMENTS:       IEEE TMI, 2020
HIGHLIGHT:      In this paper, we propose a novel multi-site network (MS-Net) for improving prostate segmentation by learning robust representations, leveraging multiple sources of data.

46, TITLE:      The Complexity of Combinations of Qualitative Constraint Satisfaction Problems
http://arxiv.org/abs/1801.05965
AUTHORS:        Manuel Bodirsky ; Johannes Greiner
HIGHLIGHT:      We show that for a large class of $\omega$-categorical theories $T_1, T_2$ the Nelson-Oppen conditions are not only sufficient, but also necessary for polynomial-time tractability of $\mathrm{CSP}(T_1 \cup T_2)$ (unless P=NP).

47, TITLE:      Scalable Reinforcement Learning of Localized Policies for Multi-Agent Networked Systems
http://arxiv.org/abs/1912.02906
AUTHORS:        Guannan Qu ; Adam Wierman ; Na Li
COMMENTS:       Added experimental results
HIGHLIGHT:      In this paper, we propose a Scalable Actor-Critic (SAC) framework that exploits the network structure and finds a localized policy that is a $O(\rho^\kappa)$-approximation of a stationary point of the objective for some $\rho\in(0,1)$, with complexity that scales with the local state-action space size of the largest $\kappa$-hop neighborhood of the network.

48, TITLE:      Report on UG^2+ Challenge Track 1: Assessing Algorithms to Improve Video Object Detection and Classification from Unconstrained Mobility Platforms
http://arxiv.org/abs/1907.11529
AUTHORS:        Sreya Banerjee ; Rosaura G. VidalMata ; Zhangyang Wang ; Walter J. Scheirer
COMMENTS:       Supplemental material: http://bit.ly/UG2Supp
HIGHLIGHT:      16 algorithms were submitted by academic and corporate teams, and a detailed analysis of how they performed on each challenge problem is reported here.

49, TITLE:      Self-labelling via simultaneous clustering and representation learning
http://arxiv.org/abs/1911.05371
AUTHORS:        Yuki Markus Asano ; Christian Rupprecht ; Andrea Vedaldi
COMMENTS:       Accepted paper at the International Conference on Learning Representations (ICLR) 2020
HIGHLIGHT:      In this paper, we propose a novel and principled learning formulation that addresses these issues.

50, TITLE:      A critical analysis of self-supervision, or what we can learn from a single image
http://arxiv.org/abs/1904.13132
AUTHORS:        Yuki M. Asano ; Christian Rupprecht ; Andrea Vedaldi
COMMENTS:       Accepted paper at the International Conference on Learning Representations (ICLR) 2020
HIGHLIGHT:      We show that three different and representative methods, BiGAN, RotNet and DeepCluster, can learn the first few layers of a convolutional network from a single image as well as using millions of images and manual labels, provided that strong data augmentation is used.

51, TITLE:      Trust Your Model: Iterative Label Improvement and Robust Training by Confidence Based Filtering and Dataset Partitioning
http://arxiv.org/abs/2002.02705
AUTHORS:        Christian Haase-Schütz ; Rainer Stal ; Heinz Hertlein ; Bernhard Sick
HIGHLIGHT:      To alleviate this issue, we propose a novel meta training and labelling scheme that is able to use inexpensive unlabelled data by taking advantage of the generalization power of deep neural networks.

52, TITLE:      Prioritized Sequence Experience Replay
http://arxiv.org/abs/1905.12726
AUTHORS:        Marc Brittain ; Josh Bertram ; Xuxi Yang ; Peng Wei
COMMENTS:       18 pages
HIGHLIGHT:      In this paper, we propose Prioritized Sequence Experience Replay (PSER) a framework for prioritizing sequences of experience in an attempt to both learn more efficiently and to obtain better performance.

53, TITLE:      Network Deconvolution
http://arxiv.org/abs/1905.11926
AUTHORS:        Chengxi Ye ; Matthew Evanusa ; Hua He ; Anton Mitrokhin ; Thomas Goldstein ; James A. Yorke ; Cornelia Fermüller ; Yiannis Aloimonos
COMMENTS:       ICLR 2020
HIGHLIGHT:      In this work, we show that this redundancy has made neural network training challenging, and propose network deconvolution, a procedure which optimally removes pixel-wise and channel-wise correlations before the data is fed into each layer.

54, TITLE:      Making Logic Learnable With Neural Networks
http://arxiv.org/abs/2002.03847
AUTHORS:        Tobias Brudermueller ; Dennis L. Shung ; Loren Laine ; Adrian J. Stanley ; Stig B. Laursen ; Harry R. Dalton ; Jeffrey Ngu ; Michael Schultz ; Johannes Stegmaier ; Smita Krishnaswamy
HIGHLIGHT:      We propose a novel logic learning pipeline that combines the advantages of neural networks and logic circuits.

55, TITLE:      Retrospective and Prospective Mixture-of-Generators for Task-oriented Dialogue Response Generation
http://arxiv.org/abs/1911.08151
AUTHORS:        Jiahuan Pei ; Pengjie Ren ; Christof Monz ; Maarten de Rijke
COMMENTS:       The paper is accepted by 24th European Conference on Artificial Intelligence
HIGHLIGHT:      We propose a novel mixture-of-generators network (MoGNet) for DRG, where we assume that each token of a response is drawn from a mixture of distributions.

56, TITLE:      Joint Extraction of Entities and Relations Based on a Novel Decomposition Strategy
http://arxiv.org/abs/1909.04273
AUTHORS:        Bowen Yu ; Zhenyu Zhang ; Xiaobo Shu ; Yubin Wang ; Tingwen Liu ; Bin Wang ; Sujian Li
COMMENTS:       Accepted by ECAI 2020. Code and data are available at https://github.com/yubowen-ph/JointER
HIGHLIGHT:      To address these limitations, in this paper, we first decompose the joint extraction task into two interrelated subtasks, namely HE extraction and TER extraction.

57, TITLE:      KEPLER: A Unified Model for Knowledge Embedding and Pre-trained Language Representation
http://arxiv.org/abs/1911.06136
AUTHORS:        Xiaozhi Wang ; Tianyu Gao ; Zhaocheng Zhu ; Zhiyuan Liu ; Juanzi Li ; Jian Tang
COMMENTS:       work in progress
HIGHLIGHT:      In this paper, we propose a unified model for Knowledge Embedding and Pre-trained LanguagE Representation (KEPLER), which can not only better integrate factual knowledge into PLMs but also effectively learn KE through the abundant information in text.

58, TITLE:      LayoutLM: Pre-training of Text and Layout for Document Image Understanding
http://arxiv.org/abs/1912.13318
AUTHORS:        Yiheng Xu ; Minghao Li ; Lei Cui ; Shaohan Huang ; Furu Wei ; Ming Zhou
COMMENTS:       Work in progress
HIGHLIGHT:      In this paper, we propose the LayoutLM to jointly model the interaction between text and layout information across scanned document images, which is beneficial for a great number of real-world document image understanding tasks such as information extraction from scanned documents.

59, TITLE:      Tensorized Embedding Layers for Efficient Model Compression
http://arxiv.org/abs/1901.10787
AUTHORS:        Oleksii Hrinchuk ; Valentin Khrulkov ; Leyla Mirvakhabova ; Elena Orlova ; Ivan Oseledets
HIGHLIGHT:      We introduce a novel way of parametrizing embedding layers based on the Tensor Train (TT) decomposition, which allows compressing the model significantly at the cost of a negligible drop or even a slight gain in performance.

60, TITLE:      Open Knowledge Enrichment for Long-tail Entities
http://arxiv.org/abs/2002.06397
AUTHORS:        Ermei Cao ; Difeng Wang ; Jiacheng Huang ; Wei Hu
COMMENTS:       Accepted by the 29th International World Wide Web Conference (WWW 2020)
HIGHLIGHT:      In this paper, we propose a full-fledged approach to knowledge enrichment, which predicts missing properties and infers true facts of long-tail entities from the open Web.

61, TITLE:      BANANAS: Bayesian Optimization with Neural Architectures for Neural Architecture Search
http://arxiv.org/abs/1910.11858
AUTHORS:        Colin White ; Willie Neiswanger ; Yash Savani
HIGHLIGHT:      In this work, we develop a suite of techniques for high-performance BO applied to NAS that allows us to achieve state-of-the-art NAS results.

62, TITLE:      Quantum Query Complexity of Dyck Languages with Bounded Height
http://arxiv.org/abs/1912.02176
AUTHORS:        Kamil Khadiev ; Yixin Shen
HIGHLIGHT:      We consider the problem of determining if a sequence of parentheses is well parenthesized, with a depth of at most h.

63, TITLE:      Towards Bounding-Box Free Panoptic Segmentation
http://arxiv.org/abs/2002.07705
AUTHORS:        Ujwal Bonde ; Pablo F. Alcantarilla ; Stefan Leutenegger
COMMENTS:       13 pages, 6 figures
HIGHLIGHT:      In this work we introduce a new bounding-box free network (BBFNet) for panoptic segmentation.

64, TITLE:      Decoupling Representation and Classifier for Long-Tailed Recognition
http://arxiv.org/abs/1910.09217
AUTHORS:        Bingyi Kang ; Saining Xie ; Marcus Rohrbach ; Zhicheng Yan ; Albert Gordo ; Jiashi Feng ; Yannis Kalantidis
HIGHLIGHT:      In this work, we decouple the learning procedure into representation learning and classification, and systematically explore how different balancing strategies affect them for long-tailed recognition.

65, TITLE:      SpotTheFake: An Initial Report on a New CNN-Enhanced Platform for Counterfeit Goods Detection
http://arxiv.org/abs/2002.06735
AUTHORS:        Alexandru Şerban ; George Ilaş ; George-Cosmin Poruşniuc
COMMENTS:       7 pages, 13 figures
HIGHLIGHT:      This paper presents the design and early stage development of a novel counterfeit goods detection platform that makes use of the outstsanding learning capabilities of the classical VGG16 convolutional model trained through the process of "transfer learning" and a multi-stage fake detection procedure that proved to be not only reliable but also very robust in the experiments we have conducted so far using an image dataset of various goods which we gathered ourselves.

66, TITLE:      ConvPoint: Continuous Convolutions for Point Cloud Processing
http://arxiv.org/abs/1904.02375
AUTHORS:        Alexandre Boulch
COMMENTS:       12 pages
HIGHLIGHT:      In this paper, we propose a generalization of discrete convolutional neural networks (CNNs) in order to deal with point clouds by replacing discrete kernels by continuous ones.

67, TITLE:      The Computational Complexity of Finding Temporal Paths under Waiting Time Constraints
http://arxiv.org/abs/1909.06437
AUTHORS:        Arnaud Casteigts ; Anne-Sophie Himmel ; Hendrik Molter ; Philipp Zschoche
HIGHLIGHT:      In this paper, we investigate a basic constraint for temporal paths, where the time spent at each vertex must not exceed a given duration $\Delta$, referred to as $\Delta$-restless temporal paths.

68, TITLE:      Outside the Box: Abstraction-Based Monitoring of Neural Networks
http://arxiv.org/abs/1911.09032
AUTHORS:        Thomas A. Henzinger ; Anna Lukina ; Christian Schilling
COMMENTS:       accepted at ECAI 2020
HIGHLIGHT:      We propose a framework to monitor a neural network by observing the hidden layers.

69, TITLE:      Ada-LISTA: Learned Solvers Adaptive to Varying Models
http://arxiv.org/abs/2001.08456
AUTHORS:        Aviad Aberdam ; Alona Golts ; Michael Elad
HIGHLIGHT:      This work introduces an adaptive learned solver, termed Ada-LISTA, which receives pairs of signals and their corresponding dictionaries as inputs, and learns a universal architecture to serve them all.

70, TITLE:      SPACE: Unsupervised Object-Oriented Scene Representation via Spatial Attention and Decomposition
http://arxiv.org/abs/2001.02407
AUTHORS:        Zhixuan Lin ; Yi-Fu Wu ; Skand Vishwanath Peri ; Weihao Sun ; Gautam Singh ; Fei Deng ; Jindong Jiang ; Sungjin Ahn
COMMENTS:       Accepted in ICLR 2020
HIGHLIGHT:      In this paper, we propose a generative latent variable model, called SPACE, that provides a unified probabilistic modeling framework that combines the best of spatial-attention and scene-mixture approaches.

71, TITLE:      WaLDORf: Wasteless Language-model Distillation On Reading-comprehension
http://arxiv.org/abs/1912.06638
AUTHORS:        James Yi Tian ; Alexander P. Kreuzer ; Pai-Hung Chen ; Hans-Martin Will
COMMENTS:       Added Figure, minor edits for clarity
HIGHLIGHT:      Here, we propose a novel set of techniques which together produce a task-specific hybrid convolutional and transformer model, WaLDORf, that achieves state-of-the-art inference speed while still being more accurate than previous distilled models.

72, TITLE:      Coaxioms: flexible coinductive definitions by inference systems
http://arxiv.org/abs/1808.02943
AUTHORS:        Francesco Dagnino
COMMENTS:       This is a corrected version of the paper (arXiv:1808.02943v4) published originally on 12 March 2019
HIGHLIGHT:      We introduce a generalized notion of inference system to support more flexible interpretations of recursive definitions.

73, TITLE:      Adaptive spline fitting with particle swarm optimization
http://arxiv.org/abs/1907.12160
AUTHORS:        Soumya D. Mohanty ; Ethan Fahnestock
COMMENTS:       Expanded literature survey; performance comparison with WaveShrink and smoothing spline; new figures and a table added
HIGHLIGHT:      We present a method that uses particle swarm optimization (PSO) combined with model selection to address this challenge.