2020.04.30.txt

==========New Papers==========
1, TITLE:       Unmanned Aerial Systems for Wildland and Forest Fires: Sensing, Perception, Cooperation and Assistance
http://arxiv.org/abs/2004.13883
AUTHORS:        Moulay A. Akhloufi ; Nicolas A. Castro ; Andy Couturier
COMMENTS:       Extended and updated version of a conference paper (see DOI)
HIGHLIGHT:      In this paper we review previous work related to the use of UAS in wildfires.

2, TITLE:       A Stochastic Team Formation Approach for Collaborative Mobile Crowdsourcing
http://arxiv.org/abs/2004.13881
AUTHORS:        Aymen Hamrouni ; Hakim Ghazzai ; Turki Alelyani ; Yehia Massoud
COMMENTS:       This paper is accepted for publication in 2019 31st International Conference on Microelectronics (ICM)
HIGHLIGHT:      In this paper, we present a hybrid approach where requesters are able to hire a team that, not only has the required expertise, but also is socially connected and can accomplish tasks collaboratively.

3, TITLE:       Analyzing Political Parody in Social Media
http://arxiv.org/abs/2004.13878
AUTHORS:        Antonis Maronikolakis ; Danae Sanchez Villegas ; Daniel Preotiuc-Pietro ; Nikolaos Aletras
HIGHLIGHT:      In this paper, we present the first computational study of parody. We introduce a new publicly available data set of tweets from real politicians and their corresponding parody accounts.

4, TITLE:       Classifying Image Sequences of Astronomical Transients with Deep Neural Networks
http://arxiv.org/abs/2004.13877
AUTHORS:        Catalina Gómez ; Mauricio Neira ; Marcela Hernández Hoyos ; Pablo Arbeláez ; Jaime E. Forero-Romero
COMMENTS:       9 pages, 6 figures. Submitted. Comments are welcome
HIGHLIGHT:      We present a successful deep learning approach that learns directly from imaging data.

5, TITLE:       Towards Prediction Explainability through Sparse Communication
http://arxiv.org/abs/2004.13876
AUTHORS:        Marcos V. Treviso ; André F. T. Martins
HIGHLIGHT:      In this work, we provide a unified perspective of explainability as a communication problem between an explainer and a layperson about a classifier's decision.

6, TITLE:       Histogram-based Auto Segmentation: A Novel Approach to Segmenting Integrated Circuit Structures from SEM Images
http://arxiv.org/abs/2004.13874
AUTHORS:        Ronald Wilson ; Navid Asadizanjani ; Domenic Forte ; Damon L. Woodard
HIGHLIGHT:      In this paper, we introduce an algorithm to segment out Integrated Circuit (IC) structures from the SEM image.

7, TITLE:       Deflating Dataset Bias Using Synthetic Data Augmentation
http://arxiv.org/abs/2004.13866
AUTHORS:        Nikita Jaipuria ; Xianling Zhang ; Rohan Bhasin ; Mayar Arafa ; Punarjay Chakravarty ; Shubham Shrivastava ; Sagar Manglani ; Vidya N. Murali
HIGHLIGHT:      The goal of this paper is to investigate the use of targeted synthetic data augmentation - combining the benefits of gaming engine simulations and sim2real style transfer techniques - for filling gaps in real datasets for vision tasks.

8, TITLE:       Empower Entity Set Expansion via Language Model Probing
http://arxiv.org/abs/2004.13897
AUTHORS:        Yunyi Zhang ; Jiaming Shen ; Jingbo Shang ; Jiawei Han
COMMENTS:       ACL 2020
HIGHLIGHT:      In this study, we propose a novel iterative set expansion framework that leverages automatically generated class names to address the semantic drift issue.

9, TITLE:       LNMap: Departures from Isomorphic Assumption in Bilingual Lexicon Induction Through Non-Linear Mapping in Latent Space
http://arxiv.org/abs/2004.13889
AUTHORS:        Tasnim Mohiuddin ; M Saiful Bari ; Shafiq Joty
COMMENTS:       10 pages, 1 figure
HIGHLIGHT:      In this work, we propose a novel semi-supervised method to learn cross-lingual word embeddings for BLI.

10, TITLE:      Synonymy = Translational Equivalence
http://arxiv.org/abs/2004.13886
AUTHORS:        Bradley Hauer ; Grzegorz Kondrak
HIGHLIGHT:      This paper proposes a unifying treatment of these two relations, which is validated by experiments on existing resources.

11, TITLE:      A Cross-Genre Ensemble Approach to Robust Reddit Part of Speech Tagging
http://arxiv.org/abs/2004.14312
AUTHORS:        Shabnam Behzad ; Amir Zeldes
COMMENTS:       Proceedings of the 12th Web as Corpus Workshop (WAC-XII)
HIGHLIGHT:      In this work, we study how a state-of-the-art tagging model trained on different genres performs on Web content from unfiltered Reddit forum discussions.

12, TITLE:      How to Learn a Useful Critic? Model-based Action-Gradient-Estimator Policy Optimization
http://arxiv.org/abs/2004.14309
AUTHORS:        Pierluca D'Oro ; Wojciech Jaśkowski
HIGHLIGHT:      In this paper, we propose MAGE, a model-based actor-critic algorithm, grounded in the theory of policy gradients, which explicitly learns the action-value gradient.

13, TITLE:      UniConv: A Unified Conversational Neural Architecture for Multi-domain Task-oriented Dialogues
http://arxiv.org/abs/2004.14307
AUTHORS:        Hung Le ; Doyen Sahoo ; Chenghao Liu ; Nancy F. Chen ; Steven C. H. Hoi
HIGHLIGHT:      Unlike the existing approaches that are often designed to train each module separately, we propose "UniConv" -- a novel unified neural architecture for end-to-end conversational systems in multi-domain task-oriented dialogues, which is designed to jointly train (i) a Bi-level State Tracker which tracks dialogue states by learning signals at both slot and domain level independently, and (ii) a Joint Dialogue Act and Response Generator which incorporates information from various input components and models dialogue acts and target responses simultaneously.

14, TITLE:      TUNIZI: a Tunisian Arabizi sentiment analysis Dataset
http://arxiv.org/abs/2004.14303
AUTHORS:        Chayma Fourati ; Abir Messaoudi ; Hatem Haddad
HIGHLIGHT:      In this paper, we introduce TUNIZI a sentiment analysis Tunisian Arabizi Dataset, collected from social networks, preprocessed for analytical studies and annotated manually by Tunisian native speakers.

15, TITLE:      Evaluating Dialogue Generation Systems via Response Selection
http://arxiv.org/abs/2004.14302
AUTHORS:        Shiki Sato ; Reina Akama ; Hiroki Ouchi ; Jun Suzuki ; Kentaro Inui
COMMENTS:       accepted by ACL 2020
HIGHLIGHT:      Specifically, we propose to construct test sets filtering out some types of false candidates: (i) those unrelated to the ground-truth response and (ii) those acceptable as appropriate responses.

16, TITLE:      DomBERT: Domain-oriented Language Model for Aspect-based Sentiment Analysis
http://arxiv.org/abs/2004.13816
AUTHORS:        Hu Xu ; Bing Liu ; Lei Shu ; Philip S. Yu
HIGHLIGHT:      We propose DomBERT, an extension of BERT to learn from both in-domain corpus and relevant domain corpora.

17, TITLE:      How Chaotic Are Recurrent Neural Networks?
http://arxiv.org/abs/2004.13838
AUTHORS:        Pourya Vakilipourtakalou ; Lili Mou
COMMENTS:       ICLR 2020 Workshop DeepDiffEq
HIGHLIGHT:      Our findings suggest that future work in this direction should address the other side of non-linear dynamics for RNN.

18, TITLE:      Pyramid Attention Networks for Image Restoration
http://arxiv.org/abs/2004.13824
AUTHORS:        Yiqun Mei ; Yuchen Fan ; Yulun Zhang ; Jiahui Yu ; Yuqian Zhou ; Ding Liu ; Yun Fu ; Thomas S. Huang ; Honghui Shi
HIGHLIGHT:      To solve this problem, we present a novel Pyramid Attention module for image restoration, which captures long-range feature correspondences from a multi-scale feature pyramid.

19, TITLE:      Less is More: Sample Selection and Label Conditioning Improve Skin Lesion Segmentation
http://arxiv.org/abs/2004.13856
AUTHORS:        Vinicius Ribeiro ; Sandra Avila ; Eduardo Valle
COMMENTS:       Accepted to the ISIC Skin Image Analysis Workshop @ CVPR 2020
HIGHLIGHT:      In this work, we show that segmentation may improve with less data, by selecting the training samples with best inter-annotator agreement, and conditioning the ground-truth masks to remove excessive detail.

20, TITLE:      Assessing Car Damage using Mask R-CNN
http://arxiv.org/abs/2004.14173
AUTHORS:        Sarath P ; Soorya M ; Shaik Abdul Rahman A ; S Suresh Kumar ; K Devaki
HIGHLIGHT:      In this paper we consider the issue of vehicle harm characterization, where a portion of the classifications can be fine-granular.

21, TITLE:      Deepfake Video Forensics based on Transfer Learning
http://arxiv.org/abs/2004.14178
AUTHORS:        Rahul U ; Ragul M ; Raja Vignesh K ; Tejeswinee K
HIGHLIGHT:      Deepfake models can create fake images and videos that humans cannot differentiate them from the genuine ones.

22, TITLE:      Conversations with Search Engines
http://arxiv.org/abs/2004.14162
AUTHORS:        Pengjie Ren ; Zhumin Chen ; Zhaochun Ren ; Evangelos Kanoulas ; Christof Monz ; Maarten de Rijke
COMMENTS:       under review at TOIS
HIGHLIGHT:      In this paper, we address the problem of answering complex information needs by conversing conversations with search engines, in the sense that users can express their queries in natural language, and directly receivethe information they need from a short system response in a conversational manner. Finally, we release the SaaC dataset and the code for CaSE and all models used for comparison to facilitate future research on this topic.

23, TITLE:      Interpretable Multimodal Routing for Human Multimodal Language
http://arxiv.org/abs/2004.14198
AUTHORS:        Yao-Hung Hubert Tsai ; Martin Q. Ma ; Muqiao Yang ; Ruslan Salakhutdinov ; Louis-Philippe Morency
HIGHLIGHT:      In this paper, we propose Multimodal Routing to separate the contributions to the prediction from each modality and the interactions between modalities.

24, TITLE:      UDapter: Language Adaptation for Truly Universal Dependency Parsing
http://arxiv.org/abs/2004.14327
AUTHORS:        Ahmet Üstün ; Arianna Bisazza ; Gosse Bouma ; Gertjan van Noord
HIGHLIGHT:      To address these issues, we propose a novel multilingual task adaptation approach based on recent work in parameter-efficient transfer learning, which allows for an easy but effective integration of existing linguistic typology features into the parsing network.

25, TITLE:      Seeing voices and hearing voices: learning discriminative embeddings using cross-modal self-supervision
http://arxiv.org/abs/2004.14326
AUTHORS:        Soo-Whan Chung ; Hong Goo Kang ; Joon Son Chung
COMMENTS:       Under submission as a conference paper
HIGHLIGHT:      The goal of this work is to train discriminative cross-modal embeddings without access to manually annotated data.

26, TITLE:      Don't Neglect the Obvious: On the Role of Unambiguous Words in Word Sense Disambiguation
http://arxiv.org/abs/2004.14325
AUTHORS:        Daniel Loureiro ; Jose Camacho-Collados
COMMENTS:       Paper submitted to ACL 2020 (rejected). The three associated ACL reviews and author response can be accessed at http://danlou.github.io/files/papers/uwa_acl20_revs.pdf - overall recommendation scores (1-5 scale): 4-4-3.5
HIGHLIGHT:      In this paper we propose a simple method to provide annotations for most unambiguous words in a large corpus. We introduce the UWA (Unambiguous Word Annotations) dataset and show how a state-of-the-art propagation-based model can use it to extend the coverage and quality of its word sense embeddings by a significant margin, improving on its original results on WSD.

27, TITLE:      The Dual Polynomial of Bipartite Perfect Matching
http://arxiv.org/abs/2004.14318
AUTHORS:        Gal Beniamini
HIGHLIGHT:      We obtain a description of the Boolean dual function of the Bipartite Perfect Matching decision problem, as a multilinear polynomial over the Reals.

28, TITLE:      Detecting Domain Polarity-Changes of Words in a Sentiment Lexicon
http://arxiv.org/abs/2004.14357
AUTHORS:        Shuai Wang ; Guangyi Lv ; Sahisnu Mazumder ; Bing Liu
HIGHLIGHT:      In this paper, we propose a graph-based technique to tackle this problem.

29, TITLE:      AxCell: Automatic Extraction of Results from Machine Learning Papers
http://arxiv.org/abs/2004.14356
AUTHORS:        Marcin Kardas ; Piotr Czapla ; Pontus Stenetorp ; Sebastian Ruder ; Sebastian Riedel ; Ross Taylor ; Robert Stojnic
HIGHLIGHT:      In this paper, we present AxCell, an automatic machine learning pipeline for extracting results from papers. We also release a structured, annotated dataset for training models for results extraction, and a dataset for evaluating the performance of models on this task.

30, TITLE:      Learning to Learn to Disambiguate: Meta-Learning for Few-Shot Word Sense Disambiguation
http://arxiv.org/abs/2004.14355
AUTHORS:        Nithin Holla ; Pushkar Mishra ; Helen Yannakoudakis ; Ekaterina Shutova
HIGHLIGHT:      We propose a meta-learning framework for few-shot word sense disambiguation (WSD), where the goal is to disambiguate unseen words from only a few labeled instances.

31, TITLE:      End-to-End Slot Alignment and Recognition for Cross-Lingual NLU
http://arxiv.org/abs/2004.14353
AUTHORS:        Weijia Xu ; Batool Haider ; Saab Mansour
HIGHLIGHT:      In this work, we propose a novel end-to-end model that learns to align and predict slots. To this end, we construct a multilingual NLU corpus, MultiATIS++, by extending the Multilingual ATIS corpus to nine languages across various language families.

32, TITLE:      Adversarial Subword Regularization for Robust Neural Machine Translation
http://arxiv.org/abs/2004.14109
AUTHORS:        Jungsoo Park ; Mujeen Sung ; Jinhyuk Lee ; Jaewoo Kang
HIGHLIGHT:      In this paper, we present adversarial subword regularization (ADVSR) to study whether gradient signals during training can be a substitute criterion for choosing segmentation among candidates.

33, TITLE:      Informative Scene Decomposition for Crowd Analysis, Comparison and Simulation Guidance
http://arxiv.org/abs/2004.14107
AUTHORS:        Feixiang He ; Yuanhang Xiang ; Xi Zhao ; He Wang
COMMENTS:       accepted in SIGGRAPH 2020
HIGHLIGHT:      In this paper, we propose a new framework which comprehensively tackles this problem.

34, TITLE:      Beyond Instructional Videos: Probing for More Diverse Visual-Textual Grounding on YouTube
http://arxiv.org/abs/2004.14338
AUTHORS:        Jack Hessel ; Zhenhai Zhu ; Bo Pang ; Radu Soricut
COMMENTS:       8 pages including supplementary materials
HIGHLIGHT:      Because instructional videos make up only a fraction of the web's diverse video content, we ask: can similar models be trained on broader corpora?

35, TITLE:      Towards Faster Reasoners By Using Transparent Huge Pages
http://arxiv.org/abs/2004.14378
AUTHORS:        Johannes K. Fichte ; Norbert Manthey ; Julian Stecklina ; André Schidler
HIGHLIGHT:      In this work, we present an approach to reduce the runtime of AR tools by 10% on average and up to 20% for long running tasks.

36, TITLE:      ToTTo: A Controlled Table-To-Text Generation Dataset
http://arxiv.org/abs/2004.14373
AUTHORS:        Ankur P. Parikh ; Xuezhi Wang ; Sebastian Gehrmann ; Manaal Faruqui ; Bhuwan Dhingra ; Diyi Yang ; Dipanjan Das
HIGHLIGHT:      We present ToTTo, an open-domain English table-to-text dataset with over 120,000 training examples that proposes a controlled generation task: given a Wikipedia table and a set of highlighted table cells, produce a one-sentence description.

37, TITLE:      VGGSound: A Large-scale Audio-Visual Dataset
http://arxiv.org/abs/2004.14368
AUTHORS:        Honglie Chen ; Weidi Xie ; Andrea Vedaldi ; Andrew Zisserman
COMMENTS:       ICASSP2020
HIGHLIGHT:      Our goal is to collect a large-scale audio-visual dataset with low label noise from videos in the wild using computer vision techniques.

38, TITLE:      Editing in Style: Uncovering the Local Semantics of GANs
http://arxiv.org/abs/2004.14367
AUTHORS:        Edo Collins ; Raja Bala ; Bob Price ; Sabine Süsstrunk
HIGHLIGHT:      Focusing on StyleGAN, we introduce a simple and effective method for making local, semantically-aware edits to a target output image.

39, TITLE:      Avoiding catastrophic forgetting in mitigating model biases in sentence-pair classification with elastic weight consolidation
http://arxiv.org/abs/2004.14366
AUTHORS:        James Thorne ; Andreas Vlachos
HIGHLIGHT:      In this paper, we show that elastic weight consolidation (EWC) allows fine-tuning of models to mitigate biases for NLI and fact verification while being less susceptible to catastrophic forgetting.

40, TITLE:      Generating Safe Diversity in NLG via Imitation Learning
http://arxiv.org/abs/2004.14364
AUTHORS:        Giulio Zhou ; Gerasimos Lampouras
COMMENTS:       10 pages, 5 figures, 5 tables
HIGHLIGHT:      In this work, we propose to ameliorate this cost by using an Imitation Learning approach to explore the level of diversity that a language generation model can safely produce.

41, TITLE:      Learning Non-Monotonic Automatic Post-Editing of Translations from Human Orderings
http://arxiv.org/abs/2004.14120
AUTHORS:        António Góis ; Kyunghyun Cho ; André Martins
COMMENTS:       Accepted at EAMT 2020; dataset available here: https://github.com/antoniogois/keystrokes_ape
HIGHLIGHT:      In this paper, we analyze the orderings produced by human post-editors and use them to train an automatic post-editing system.

42, TITLE:      Analysing Lexical Semantic Change with Contextualised Word Representations
http://arxiv.org/abs/2004.14118
AUTHORS:        Mario Giulianelli ; Marco Del Tredici ; Raquel Fernández
COMMENTS:       To appear in Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL-2020)
HIGHLIGHT:      This paper presents the first unsupervised approach to lexical semantic change that makes use of contextualised word representations. We create a new evaluation dataset and show that the model representations and the detected semantic shifts are positively correlated with human judgements.

43, TITLE:      A Fast 3D CNN for Hyperspectral Image Classification
http://arxiv.org/abs/2004.14152
AUTHORS:        Muhammad Ahmad
COMMENTS:       5 pages, 8 figures, (IEEE GRSL)
HIGHLIGHT:      Therefore, this work proposed a 3D CNN model that utilizes both spatial-spectral feature maps to attain good performance.

44, TITLE:      On the Existence of Algebraically Natural Proofs
http://arxiv.org/abs/2004.14147
AUTHORS:        Prerona Chatterjee ; Mrinal Kumar ; C. Ramya ; Ramprasad Saptharishi ; Anamay Tengse
HIGHLIGHT:      Thus, in this setting of polynomials with small integer coefficients, this provides evidence \emph{against} a natural proof like barrier for proving algebraic circuit lower bounds, a framework for which was proposed in the works of Forbes, Shpilka and Volk (2018), and Grochow, Kumar, Saks and Saraf (2017).

45, TITLE:      Entity Candidate Network for Whole-Aware Named Entity Recognition
http://arxiv.org/abs/2004.14145
AUTHORS:        Wendong He ; Yizhen Shao ; Pingjian Zhang
COMMENTS:       10 pages, 4 figures
HIGHLIGHT:      Inspired by one-stage object detection models in computer vision (CV), this paper proposes a new no-tag scheme, the Whole-Aware Detection, which makes NER an object detection task.

46, TITLE:      Zero-Shot Learning and its Applications from Autonomous Vehicles to COVID-19 Diagnosis: A Review
http://arxiv.org/abs/2004.14143
AUTHORS:        Mahdi Rezaei ; Mahsa Shahidi
HIGHLIGHT:      In this paper, we present the definition of the problem, we review over fundamentals, and the challenging steps of Zero-shot learning, including recent categories of solutions, motivations behind each approach, and their advantages over other categories.

47, TITLE:      Retinal vessel segmentation by probing adaptive to lighting variations
http://arxiv.org/abs/2004.13992
AUTHORS:        Guillaume Noyel ; Christine Vartin ; Peter Boyle ; Laurent Kodjikian
COMMENTS:       Proceedings of 2020 IEEE 17th International Symposium on Biomedical Imaging (ISBI).To appear in https://ieeexplore.ieee.org
HIGHLIGHT:      We introduce a novel method to extract the vessels in eye fun-dus images which is adaptive to lighting variations.

48, TITLE:      Multi-choice Dialogue-Based Reading Comprehension with Knowledge and Key Turns
http://arxiv.org/abs/2004.13988
AUTHORS:        Junlong Li ; Zhuosheng Zhang ; Hai Zhao
HIGHLIGHT:      In this paper, the relevance of each turn to the question are calculated to choose key turns.

49, TITLE:      Motion Guided 3D Pose Estimation from Videos
http://arxiv.org/abs/2004.13985
AUTHORS:        Jingbo Wang ; Sijie Yan ; Yuanjun Xiong ; Dahua Lin
HIGHLIGHT:      We propose a new loss function, called motion loss, for the problem of monocular 3D Human pose estimation from 2D pose.

50, TITLE:      Conspiracy in the Time of Corona: Automatic detection of Covid-19 Conspiracy Theories in Social Media and the News
http://arxiv.org/abs/2004.13783
AUTHORS:        Shadi Shahsavari ; Pavan Holur ; Timothy R. Tangherlini ; Vwani Roychowdhury
COMMENTS:       Covid-19, Corona virus, conspiracy theories, 5G, Bill Gates, China, bio-weapons, rumor, narrative, machinelearning, social media, 4Chan, Reddit, networks, data visualization
HIGHLIGHT:      Inspired by narrative theory, we crawl social media sites and news reports and, through the application of automated machine-learning methods, discover the underlying narrative frameworks supporting the generation of these stories.

51, TITLE:      Cross-modal Speaker Verification and Recognition: A Multilingual Perspective
http://arxiv.org/abs/2004.13780
AUTHORS:        Muhammad Saad Saeed ; Shah Nawaz ; Pietro Morerio ; Arif Mahmood ; Ignazio Gallo ; Muhammad Haroon Yousaf ; Alessio Del Bue
HIGHLIGHT:      The aim of this paper is to answer two closely related questions: \textit{"Is face-voice association language independent?"} To answer them, we collected a Multilingual Audio-Visual dataset, containing human speech clips of $154$ identities with $3$ language annotations extracted from various videos uploaded online.

52, TITLE:      Minority Reports Defense: Defending Against Adversarial Patches
http://arxiv.org/abs/2004.13799
AUTHORS:        Michael McCoyd ; Won Park ; Steven Chen ; Neil Shah ; Ryan Roggenkemper ; Minjune Hwang ; Jason Xinyu Liu ; David Wagner
COMMENTS:       9 pages, 5 figures
HIGHLIGHT:      We propose a defense against patch attacks based on partially occluding the image around each candidate patch location, so that a few occlusions each completely hide the patch.

53, TITLE:      Mixing Probabilistic and non-Probabilistic Objectives in Markov Decision Processes
http://arxiv.org/abs/2004.13789
AUTHORS:        Raphaël Berthon ; Shibashis Guha ; Jean-François Raskin
COMMENTS:       Paper accepted to LICS 2020 - Full version
HIGHLIGHT:      In this paper, we consider algorithms to decide the existence of strategies in MDPs for Boolean combinations of objectives.

54, TITLE:      A Practical Framework for Relation Extraction with Noisy Labels Based on Doubly Transitional Loss
http://arxiv.org/abs/2004.13786
AUTHORS:        Shanchan Wu ; Kai Fan
COMMENTS:       10 pages
HIGHLIGHT:      To address this issue, we introduce a practical end-to-end deep learning framework, including a standard feature extractor and a novel noisy classifier with our proposed doubly transitional mechanism.

55, TITLE:      Neural Additive Models: Interpretable Machine Learning with Neural Nets
http://arxiv.org/abs/2004.13912
AUTHORS:        Rishabh Agarwal ; Nicholas Frosst ; Xuezhou Zhang ; Rich Caruana ; Geoffrey E. Hinton
HIGHLIGHT:      We propose Neural Additive Models (NAMs) which combine some of the expressivity of DNNs with the inherent intelligibility of generalized additive models.

56, TITLE:      An Auto-Encoder Strategy for Adaptive Image Segmentation
http://arxiv.org/abs/2004.13903
AUTHORS:        Evan M. Yu ; Juan Eugenio Iglesias ; Adrian V. Dalca ; Mert R. Sabuncu
COMMENTS:       MIDL 2020
HIGHLIGHT:      In this paper, we propose a novel perspective of segmentation as a discrete representation learning problem, and present a variational autoencoder segmentation strategy that is flexible and adaptive.

57, TITLE:      Synergistic CPU-FPGA Acceleration of Sparse Linear Algebra
http://arxiv.org/abs/2004.13907
AUTHORS:        Mohammadreza Soltaniyeh ; Richard P. Martin ; Santosh Nagarakatte
COMMENTS:       12 pages
HIGHLIGHT:      This paper describes REAP, a software-hardware approach that enables high performance sparse linear algebra computations on a cooperative CPU-FPGA platform.

58, TITLE:      Multi-layer local optima networks for the analysis of advanced local search-based algorithms
http://arxiv.org/abs/2004.13936
AUTHORS:        Marcella Scoczynski Ribeiro Martins ; Mohamed El Yafrani ; Myriam R. B. S. Delgado ; Ricardo Luders
COMMENTS:       Accepted in GECCO2020
HIGHLIGHT:      Therefore, in the present paper we investigate a twolayer LON obtained from instances of a combinatorial problem using bitflip and swap operators.

59, TITLE:      Span-based Localizing Network for Natural Language Video Localization
http://arxiv.org/abs/2004.13931
AUTHORS:        Hao Zhang ; Aixin Sun ; Wei Jing ; Joey Tianyi Zhou
COMMENTS:       To appear at ACL 2020
HIGHLIGHT:      In this work, we address NLVL task with a span-based QA approach by treating the input video as text passage.

60, TITLE:      Evaluating the Role of Language Typology in Transformer-Based Multilingual Text Classification
http://arxiv.org/abs/2004.13939
AUTHORS:        Sophie Groenwold ; Samhita Honnavalli ; Lily Ou ; Aesha Parekh ; Sharon Levy ; Diba Mirza ; William Yang Wang
COMMENTS:       Total of 15 pages (9 pages for paper, 2 pages for references, 4 pages for appendix)
HIGHLIGHT:      Through a detailed discussion of word order typology, morphological typology, and comparative linguistics, we identify which variables most affect language modeling efficacy; in addition, we calculate word order and morphological similarity indices to aid our empirical study.

61, TITLE:      Revisiting Round-Trip Translation for Quality Estimation
http://arxiv.org/abs/2004.13937
AUTHORS:        Jihyung Moon ; Hyunchang Cho ; Eunjeong L. Park
COMMENTS:       To be published in EAMT 2020
HIGHLIGHT:      In this paper, we employ semantic embeddings to RTT-based QE.

62, TITLE:      Hybrid Adaptive Evolutionary Algorithm for Multi-objective Optimization
http://arxiv.org/abs/2004.13925
AUTHORS:        Jeisson Prieto ; Jonatan Gomez
HIGHLIGHT:      This paper proposed a new Multi-objective Algorithm as an extension of the Hybrid Adaptive Evolutionary algorithm (HAEA) called MoHAEA.

63, TITLE:      Revisiting Pre-Trained Models for Chinese Natural Language Processing
http://arxiv.org/abs/2004.13922
AUTHORS:        Yiming Cui ; Wanxiang Che ; Ting Liu ; Bing Qin ; Shijin Wang ; Guoping Hu
COMMENTS:       11 pages, as an extension to arXiv:1906.08101
HIGHLIGHT:      In this paper, we target on revisiting Chinese pre-trained models to examine their effectiveness in a non-English language and release the Chinese pre-trained model series to the community.

64, TITLE:      Zero-shot topic generation
http://arxiv.org/abs/2004.13956
AUTHORS:        Oleg Vasilyev ; Kathryn Evans ; Anna Venancio-Marques ; John Bohannon
COMMENTS:       12 pages, 9 figures, 3 tables
HIGHLIGHT:      We present an approach to generating topics using a model trained only for document title generation, with zero examples of topics given during training.

65, TITLE:      Rethink the Connections among Generalization, Memorization and the Spectral Bias of DNNs
http://arxiv.org/abs/2004.13954
AUTHORS:        Xiao Zhang ; Dongrui Wu
COMMENTS:       14 pages,11 figures
HIGHLIGHT:      Recent studies claimed that DNNs first learn simple patterns and then memorize noise; some other works showed that DNNs have a spectral bias to learn target functions from low to high frequencies during training.

66, TITLE:      Data Augmentation for Spoken Language Understanding via Pretrained Models
http://arxiv.org/abs/2004.13952
AUTHORS:        Baolin Peng ; Chenguang Zhu ; Michael Zeng ; Jianfeng Gao
COMMENTS:       6 pages, 1 figure
HIGHLIGHT:      In this paper, we put forward a data augmentation method with pretrained language models to boost the variability and accuracy of generated utterances.

67, TITLE:      Video Contents Understanding using Deep Neural Networks
http://arxiv.org/abs/2004.13959
AUTHORS:        Mohammadhossein Toutiaee ; Abbas Keshavarzi ; Abolfazl Farahani ; John A. Miller
HIGHLIGHT:      We propose a novel application of Transfer Learning to classify video-frame sequences over multiple classes.

68, TITLE:      Learning Better Universal Representations from Pre-trained Contextualized Language Models
http://arxiv.org/abs/2004.13947
AUTHORS:        Yian Li ; Hai Zhao
HIGHLIGHT:      In this work, we present a novel framework on BERT that is capable of generating universal, fixed-size representations for input sequences of any lengths, i.e., words, phrases, and sentences, using a large scale of natural language inference and paraphrase data with multiple training objectives.

69, TITLE:      Basic Linguistic Resources and Baselines for Bhojpuri, Magahi and Maithili for Natural Language Processing
http://arxiv.org/abs/2004.13945
AUTHORS:        Rajesh Kumar Mundotiya ; Manish Kumar Singh ; Rahul Kapur ; Swasti Mishra ; Anil Kumar Singh
HIGHLIGHT:      They are closely related to Hindi, which is a relatively high-resource language, which is why we make our comparisons with Hindi. We collected corpora for these three languages from various sources and cleaned them to the extent possible, without changing the data in them.

70, TITLE:      Conditional Neural Generation using Sub-Aspect Functions for Extractive News Summarization
http://arxiv.org/abs/2004.13983
AUTHORS:        Zhengyuan Liu ; Ke Shi ; Nancy Chen
HIGHLIGHT:      We propose a neural framework that can flexibly control which sub-aspect functions (i.e. importance, diversity, position) to focus on during summary generation.

71, TITLE:      Measuring Information Propagation in Literary Social Networks
http://arxiv.org/abs/2004.13980
AUTHORS:        Matthew Sims ; David Bamman
HIGHLIGHT:      We describe a new pipeline for measuring information propagation in this domain and publish a new dataset for speaker attribution, enabling the evaluation of an important component of this pipeline on a wider range of literary texts than previously studied.

72, TITLE:      Skeleton Focused Human Activity Recognition in RGB Video
http://arxiv.org/abs/2004.13979
AUTHORS:        Bruce X. B. Yu ; Yan Liu ; Keith C. C. Chan
COMMENTS:       8 pages
HIGHLIGHT:      In this paper, we propose a multimodal feature fusion model that utilizes both skeleton and RGB modalities to infer human activity.

73, TITLE:      Effective Human Activity Recognition Based on Small Datasets
http://arxiv.org/abs/2004.13977
AUTHORS:        Bruce X. B. Yu ; Yan Liu ; Keith C. C. Chan
COMMENTS:       7 pages
HIGHLIGHT:      To do so, we propose a HAR method that consists of three steps: (i) data transformation involving the generation of new features based on transforming of raw data, (ii) feature extraction involving the learning of a classifier based on the AdaBoost algorithm and the use of training data consisting of the transformed features, and (iii) parameter determination and pattern recognition involving the determination of parameters based on the features generated in (ii) and the use of the parameters as training data for deep learning algorithms to be used to recognize human activities.

74, TITLE:      Deep Transfer Learning For Plant Center Localization
http://arxiv.org/abs/2004.13973
AUTHORS:        Enyu Cai ; Sriram Baireddy ; Changye Yang ; Melba Crawford ; Edward J. Delp
HIGHLIGHT:      The goal of this paper is to investigate methods that estimate plant locations for a field-based crop using RGB aerial images captured using Unmanned Aerial Vehicles (UAVs).

75, TITLE:      Reduced Bond Graph via machine learning for nonlinear multiphysics dynamic systems
http://arxiv.org/abs/2004.13971
AUTHORS:        Youssef Hammadi ; David Ryckelynck ; Amin El-Bakkali
HIGHLIGHT:      In this paper, a neural network is obtained by a linear calibration procedure.

76, TITLE:      Minimal Rolling Shutter Absolute Pose with Unknown Focal Length and Radial Distortion
http://arxiv.org/abs/2004.14052
AUTHORS:        Zuzana Kukelova ; Cenek Albl ; Akihiro Sugimoto ; Konrad Schindler ; Tomas Pajdla
HIGHLIGHT:      We present the first minimal solutions for the absolute pose of a rolling shutter camera with unknown rolling shutter parameters, focal length, and radial distortion.

77, TITLE:      Detecting Perceived Emotions in Hurricane Disasters
http://arxiv.org/abs/2004.14299
AUTHORS:        Shrey Desai ; Cornelia Caragea ; Junyi Jessy Li
COMMENTS:       Accepted to ACL 2020; code available at https://github.com/shreydesai/hurricane
HIGHLIGHT:      In this paper, we introduce HurricaneEmo, an emotion dataset of 15,000 English tweets spanning three hurricanes: Harvey, Irma, and Maria.

78, TITLE:      Topic Propagation in Conversational Search
http://arxiv.org/abs/2004.14054
AUTHORS:        I. Mele ; C. I. Muntean ; F. M. Nardini ; R. Perego ; N. Tonellotto ; O. Frieder
COMMENTS:       5 pages
HIGHLIGHT:      We present a comprehensive experimental evaluation of the architecture assessed in terms of traditional IR metrics at small cutoffs.

79, TITLE:      SubjQA: A Dataset for Subjectivity and Review Comprehension
http://arxiv.org/abs/2004.14283
AUTHORS:        Johannes Bjerva ; Nikita Bhutani ; Behzad Golshan ; Wang-Chiew Tan ; Isabelle Augenstein
HIGHLIGHT:      We therefore investigate the relationship between subjectivity and QA, while developing a new dataset. We release an English QA dataset (SubjQA) based on customer reviews, containing subjectivity annotations for questions and answer spans across 6 distinct domains.

80, TITLE:      Towards Character-Level Transformer NMT by Finetuning Subword Systems
http://arxiv.org/abs/2004.14280
AUTHORS:        Jindřich Libovický ; Alexander Fraser
COMMENTS:       6 pages, 1 figure
HIGHLIGHT:      A few approaches have been proposed that partially overcome this problem by using explicit segmentation into tokens.

81, TITLE:      General Purpose Text Embeddings from Pre-trained Language Models for Scalable Inference
http://arxiv.org/abs/2004.14287
AUTHORS:        Jingfei Du ; Myle Ott ; Haoran Li ; Xing Zhou ; Veselin Stoyanov
HIGHLIGHT:      We compare approaches for training such an encoder and show that encoders pre-trained over multiple tasks generalize well to unseen tasks.

82, TITLE:      Single-Side Domain Generalization for Face Anti-Spoofing
http://arxiv.org/abs/2004.14043
AUTHORS:        Yunpei Jia ; Jie Zhang ; Shiguang Shan ; Xilin Chen
HIGHLIGHT:      In this work, we propose an end-to-end single-side domain generalization framework (SSDG) to improve the generalization ability of face anti-spoofing.

83, TITLE:      Pre-training Is (Almost) All You Need: An Application to Commonsense Reasoning
http://arxiv.org/abs/2004.14074
AUTHORS:        Alexandre Tamborrino ; Nicola Pellicano ; Baptiste Pannier ; Pascal Voitot ; Louise Naudin
COMMENTS:       Accepted at ACL 2020
HIGHLIGHT:      In this paper, we introduce a new scoring method that casts a plausibility ranking task in a full-text format and leverages the masked language modeling head tuned during the pre-training phase.

84, TITLE:      Image Morphing with Perceptual Constraints and STN Alignment
http://arxiv.org/abs/2004.14071
AUTHORS:        Noa Fish ; Richard Zhang ; Lilach Perry ; Daniel Cohen-Or ; Eli Shechtman ; Connelly Barnes
HIGHLIGHT:      In this paper, we propose a conditional GAN morphing framework operating on a pair of input images.

85, TITLE:      DR-SPAAM: A Spatial-Attention and Auto-regressive Model for Person Detection in 2D Range Data
http://arxiv.org/abs/2004.14079
AUTHORS:        Dan Jia ; Alexander Hermans ; Bastian Leibe
HIGHLIGHT:      In this paper, we propose a person detection network which uses an alternative strategy to combine scans obtained at different times.

86, TITLE:      Morphological Disambiguation of South Sámi with FSTs and Neural Networks
http://arxiv.org/abs/2004.14062
AUTHORS:        Mika Hämäläinen ; Linda Wiechetek
COMMENTS:       1st Joint SLTU and CCURL Workshop (SLTU-CCURL 2020)
HIGHLIGHT:      We present a method for conducting morphological disambiguation for South S\'ami, which is an endangered language.

87, TITLE:      Enhancing Answer Boundary Detection for Multilingual Machine Reading Comprehension
http://arxiv.org/abs/2004.14069
AUTHORS:        Fei Yuan ; Linjun Shou ; Xuanyu Bai ; Ming Gong ; Yaobo Liang ; Nan Duan ; Yan Fu ; Daxin Jiang
COMMENTS:       ACL 2020
HIGHLIGHT:      In this paper, we propose two auxiliary tasks in the fine-tuning stage to create additional phrase boundary supervision: (1) A mixed MRC task, which translates the question or passage to other languages and builds cross-lingual question-passage pairs; (2) A language-agnostic knowledge masking task by leveraging knowledge phrases mined from web.

88, TITLE:      FitChat: Conversational Artificial Intelligence Interventions for Encouraging Physical Activity in Older Adults
http://arxiv.org/abs/2004.14067
AUTHORS:        Nirmalie Wiratunga ; Kay Cooper ; Anjana Wijekoon ; Chamath Palihawadana ; Vanessa Mendham ; Ehud Reiter ; Kyle Martin
HIGHLIGHT:      In this work, we explore the use of voice-based AI chatbots as a novel mode of intervention delivery, specifically targeting older adults to encourage physical activity.

89, TITLE:      Automatically Identifying Gender Issues in Machine Translation using Perturbations
http://arxiv.org/abs/2004.14065
AUTHORS:        Hila Gonen ; Kellie Webster
HIGHLIGHT:      We use our new method to compile an evaluation benchmark spanning examples relating to four languages from three language families, which we will publicly release to facilitate research.

90, TITLE:      Do Neural Language Models Show Preferences for Syntactic Formalisms?
http://arxiv.org/abs/2004.14096
AUTHORS:        Artur Kulmizev ; Vinit Ravishankar ; Mostafa Abdou ; Joakim Nivre
COMMENTS:       ACL 2020
HIGHLIGHT:      In this study, we aim to investigate the extent to which the semblance of syntactic structure captured by language models adheres to a surface-syntactic or deep syntactic style of analysis, and whether the patterns are consistent across different languages.

91, TITLE:      Compilation of Coordinated Choice
http://arxiv.org/abs/2004.14084
AUTHORS:        Yuki Nishida ; Atsushi Igarashi
HIGHLIGHT:      In this paper, we define two simply typed lambda calculi called $\lambda^\parallel$ equipped with standard choices and $\lambda^{\parallel\omega}$ equipped with coordinated choices, and give compilation rules from the former into the latter.

92, TITLE:      Modeling Long Context for Task-Oriented Dialogue State Generation
http://arxiv.org/abs/2004.14080
AUTHORS:        Jun Quan ; Deyi Xiong
COMMENTS:       ACL 2020
HIGHLIGHT:      Based on the recently proposed transferable dialogue state generator (TRADE) that predicts dialogue states from utterance-concatenated dialogue context, we propose a multi-task learning model with a simple yet effective utterance tagging technique and a bidirectional language model as an auxiliary task for task-oriented dialogue state generation.

93, TITLE:      Leveraging Declarative Knowledge in Text and First-Order Logic for Fine-Grained Propaganda Detection
http://arxiv.org/abs/2004.14201
AUTHORS:        Ruize Wang ; Duyu Tang ; Nan Duan ; Wanjun Zhong ; Zhongyu Wei ; Xuanjing Huang ; Daxin Jiang ; Ming Zhou
HIGHLIGHT:      Instead of merely learning from input-output datapoints in training data, we introduce an approach to inject declarative knowledge of fine-grained propaganda techniques.

94, TITLE:      Syntax-aware Data Augmentation for Neural Machine Translation
http://arxiv.org/abs/2004.14200
AUTHORS:        Sufeng Duan ; Hai Zhao ; Dongdong Zhang ; Rui Wang
HIGHLIGHT:      In this paper, we propose a novel data augmentation enhancement strategy for neural machine translation.

95, TITLE:      Normalizing Compositional Structures Across Graphbanks
http://arxiv.org/abs/2004.14236
AUTHORS:        Lucia Donatelli ; Jonas Groschwitz ; Alexander Koller ; Matthias Lindemann ; Pia Weißenhorn
COMMENTS:       16 pages, 6 figures
HIGHLIGHT:      We present a methodology for normalizing discrepancies between MRs at the compositional level (Lindemann et al., 2019), finding that we can normalize the majority of divergent phenomena using linguistically-grounded rules.

96, TITLE:      Image Captioning through Image Transformer
http://arxiv.org/abs/2004.14231
AUTHORS:        Sen He ; Wentong Liao ; Hamed R. Tavakoli ; Michael Yang ; Bodo Rosenhahn ; Nicolas Pugeault
HIGHLIGHT:      In this work, we introduce the \textbf{\textit{image transformer}}, which consists of a modified encoding transformer and an implicit decoding transformer, motivated by the relative spatial relationship between image regions.

97, TITLE:      Meta-Transfer Learning for Code-Switched Speech Recognition
http://arxiv.org/abs/2004.14228
AUTHORS:        Genta Indra Winata ; Samuel Cahyawijaya ; Zhaojiang Lin ; Zihan Liu ; Peng Xu ; Pascale Fung
COMMENTS:       Accepted in ACL 2020. The first two authors contributed equally to this work
HIGHLIGHT:      We therefore propose a new learning method, meta-transfer learning, to transfer learn on a code-switched speech recognition system in a low-resource setting by judiciously extracting information from high-resource monolingual datasets.

98, TITLE:      Exploiting Structured Knowledge in Text via Graph-Guided Representation Learning
http://arxiv.org/abs/2004.14224
AUTHORS:        Tao Shen ; Yi Mao ; Pengcheng He ; Guodong Long ; Adam Trischler ; Weizhu Chen
HIGHLIGHT:      In this work, we aim at equipping pre-trained language models with structured knowledge.

99, TITLE:      Exploring Fine-tuning Techniques for Pre-trained Cross-lingual Models via Continual Learning
http://arxiv.org/abs/2004.14218
AUTHORS:        Zihan Liu ; Genta Indra Winata ; Andrea Madotto ; Pascale Fung
HIGHLIGHT:      To alleviate this issue, we leverage the idea of continual learning to preserve the original cross-lingual ability of the pre-trained model when we fine-tune it to downstream cross-lingual tasks.

100, TITLE:     Politeness Transfer: A Tag and Generate Approach
http://arxiv.org/abs/2004.14257
AUTHORS:        Aman Madaan ; Amrith Setlur ; Tanmay Parekh ; Barnabas Poczos ; Graham Neubig ; Yiming Yang ; Ruslan Salakhutdinov ; Alan W Black ; Shrimai Prabhumoye
COMMENTS:       To appear at ACL 2020
HIGHLIGHT:      This paper introduces a new task of politeness transfer which involves converting non-polite sentences to polite sentences while preserving the meaning. We also provide a dataset of more than 1.39 instances automatically labeled for politeness to encourage benchmark evaluations on this new task.

101, TITLE:     Versatile Black-Box Optimization
http://arxiv.org/abs/2004.14014
AUTHORS:        Jialin Liu ; Antoine Moreau ; Mike Preuss ; Baptiste Roziere ; Jeremy Rapin ; Fabien Teytaud ; Olivier Teytaud
COMMENTS:       Accepted at GECCO 2020
HIGHLIGHT:      We present Shiwa, an algorithm good at both discrete and continuous, noisy and noise-free, sequential and parallel, black-box optimization.

102, TITLE:     Task-oriented Dialogue System for Automatic Disease Diagnosis via Hierarchical Reinforcement Learning
http://arxiv.org/abs/2004.14254
AUTHORS:        Kangenbei Liao ; Qianlong Liu ; Zhongyu Wei ; Baolin Peng ; Qin Chen ; Weijian Sun ; Xuanjing Huang
HIGHLIGHT:      In this paper, we focus on automatic disease diagnosis with reinforcement learning (RL) methods in task-oriented dialogues setting.

103, TITLE:     GePpeTto Carves Italian into a Language Model
http://arxiv.org/abs/2004.14253
AUTHORS:        Lorenzo De Mattei ; Michele Cafagna ; Felice Dell'Orletta ; Malvina Nissim ; Marco Guerini
HIGHLIGHT:      We provide a thorough analysis of GePpeTto's quality by means of both an automatic and a human-based evaluation.

104, TITLE:     Counting of Grapevine Berries in Images via Semantic Segmentation using Convolutional Neural Networks
http://arxiv.org/abs/2004.14010
AUTHORS:        Laura Zabawa ; Anna Kicherer ; Lasse Klingbeil ; Reinhard Töpfer ; Heiner Kuhlmann ; Ribana Roscher
HIGHLIGHT:      In this paper we present an objective framework based on automatic image analysis which works on two different training systems.

105, TITLE:     Action Sequence Predictions of Vehicles in Urban Environments using Map and Social Context
http://arxiv.org/abs/2004.14251
AUTHORS:        Jan-Nico Zaech ; Dengxin Dai ; Alexander Liniger ; Luc Van Gool
HIGHLIGHT:      Our second contribution lies in applying the method to the well-known traffic agent tracking and prediction dataset Argoverse, resulting in 228,000 action sequences.

106, TITLE:     Utterance Pair Scoring for Noisy Dialogue Data Filtering
http://arxiv.org/abs/2004.14008
AUTHORS:        Reina Akama ; Sho Yokoi ; Jun Suzuki ; Kentaro Inui
COMMENTS:       10 pages
HIGHLIGHT:      In this work, we propose a scoring function that is specifically designed to identify low-quality utterance--response pairs to filter noisy training data.

107, TITLE:     Benchmarking Robustness of Machine Reading Comprehension Models
http://arxiv.org/abs/2004.14004
AUTHORS:        Chenglei Si ; Ziqing Yang ; Yiming Cui ; Wentao Ma ; Ting Liu ; Shijin Wang
COMMENTS:       Work in progress
HIGHLIGHT:      Our benchmark is constructed automatically based on the existing RACE benchmark, and thus the construction pipeline can be easily adopted by other tasks and datasets. We will release the data and source codes to facilitate future work.

108, TITLE:     The International Workshop on Osteoarthritis Imaging Knee MRI Segmentation Challenge: A Multi-Institute Evaluation and Analysis Framework on a Standardized Dataset
http://arxiv.org/abs/2004.14003
AUTHORS:        Arjun D. Desai ; Francesco Caliva ; Claudia Iriondo ; Naji Khosravan ; Aliasghar Mortazi ; Sachin Jambawalikar ; Drew Torigian ; Jutta Ellerman ; Mehmet Akcakaya ; Ulas Bagci ; Radhika Tibrewala ; Io Flament ; Matthew O`Brien ; Sharmila Majumdar ; Mathias Perslev ; Akshay Pai ; Christian Igel ; Erik B. Dam ; Sibaji Gaj ; Mingrui Yang ; Kunio Nakamura ; Xiaojuan Li ; Cem M. Deniz ; Vladimir Juras ; Ravinder Regatte ; Garry E. Gold ; Brian A. Hargreaves ; Valentina Pedoia ; Akshay S. Chaudhari
COMMENTS:       Submitted to Radiology: Artificial Intelligence
HIGHLIGHT:      The International Workshop on Osteoarthritis Imaging Knee MRI Segmentation Challenge: A Multi-Institute Evaluation and Analysis Framework on a Standardized Dataset

109, TITLE:     Towards Transparent and Explainable Attention Models
http://arxiv.org/abs/2004.14243
AUTHORS:        Akash Kumar Mohankumar ; Preksha Nema ; Sharan Narasimhan ; Mitesh M. Khapra ; Balaji Vasan Srinivasan ; Balaraman Ravindran
COMMENTS:       Accepted at ACL 2020
HIGHLIGHT:      To make attention mechanisms more faithful and plausible, we propose a modified LSTM cell with a diversity-driven training objective that ensures that the hidden representations learned at different time steps are diverse.

110, TITLE:     The Holy Grail of Quantum Artificial Intelligence: Major Challenges in Accelerating the Machine Learning Pipeline
http://arxiv.org/abs/2004.14035
AUTHORS:        Thomas Gabor ; Leo Sünkel ; Fabian Ritz ; Thomy Phan ; Lenz Belzner ; Christoph Roch ; Sebastian Feld ; Claudia Linnhoff-Popien
COMMENTS:       6 pages, 4 figures, accepted at the 1st International Workshop on Quantum Software Engineering (Q-SE 2020) at ICSE 2020 and to be published in the corresponding proceedings
HIGHLIGHT:      We discuss the synergetic connection between quantum computing and artificial intelligence.

111, TITLE:     Emerging Relation Network and Task Embedding for Multi-Task Regression Problems
http://arxiv.org/abs/2004.14034
AUTHORS:        Jens Schreiber ; Bernhard Sick
COMMENTS:       8 pages;2 tables;5 figures
HIGHLIGHT:      Therefore, this article provides a comparative study of the following recent and important mtl architectures: Hard parameter sharing, cross-stitch network, sluice network (sn).

112, TITLE:     Tensor train rank minimization with nonlocal self-similarity for tensor completion
http://arxiv.org/abs/2004.14273
AUTHORS:        Meng Ding ; Ting-Zhu Huang ; Xi-Le Zhao ; Michael K. Ng ; Tian-Hui Ma
HIGHLIGHT:      We develop the alternating direction method of multipliers tailored for the specific structure to solve the proposed model.

113, TITLE:     Multi-View Attention Networks for Visual Dialog
http://arxiv.org/abs/2004.14025
AUTHORS:        Sungjin Park ; Taesun Whang ; Yeochan Yoon ; Hueiseok Lim
HIGHLIGHT:      In this paper, we propose Multi-View Attention Network (MVAN), which considers complementary views of multimodal inputs based on attention mechanisms.

114, TITLE:     Exploring the Suitability of Semantic Spaces as Word Association Models for the Extraction of Semantic Relationships
http://arxiv.org/abs/2004.14265
AUTHORS:        Epaminondas Kapetanios ; Vijayan Sugumaran ; Anastassia Angelopoulou
COMMENTS:       10 pages, 11 tables
HIGHLIGHT:      In this paper, we empirically study and explore the potential of a novel idea of using classical semantic spaces and models, e.g., Word Embedding, generated for extracting word association, in conjunction with relation extraction approaches.

115, TITLE:     Multiscale Collaborative Deep Models for Neural Machine Translation
http://arxiv.org/abs/2004.14021
AUTHORS:        Xiangpeng Wei ; Heng Yu ; Yue Hu ; Yue Zhang ; Rongxiang Weng ; Weihua Luo
COMMENTS:       Accpeted by ACL-2020
HIGHLIGHT:      In this paper, we present a MultiScale Collaborative (MSC) framework to ease the training of NMT models that are substantially deeper than those used previously.



==========Updates to Previous Papers==========
1, TITLE:       Benchmarking Deep Spiking Neural Networks on Neuromorphic Hardware
http://arxiv.org/abs/2004.01656
AUTHORS:        Christoph Ostrau ; Jonas Homburg ; Christian Klarhorst ; Michael Thies ; Ulrich Rückert
HIGHLIGHT:      In this work, we use the methodology of converting pre-trained non-spiking to spiking neural networks to evaluate the performance loss and measure the energy-per-inference for three neuromorphic hardware systems (BrainScaleS, Spikey, SpiNNaker) and common simulation frameworks for CPU (NEST) and CPU/GPU (GeNN).

2, TITLE:       Analysis of Evolutionary Algorithms on Fitness Function with Time-linkage Property
http://arxiv.org/abs/2004.12304
AUTHORS:        Weijie Zheng ; Huanhuan Chen ; Xin Yao
HIGHLIGHT:      Based on the basic OneMax function, we propose a time-linkage function where the first bit value of the last time step is integrated but has a different preference from the current first bit.

3, TITLE:       Attention Is All You Need for Chinese Word Segmentation
http://arxiv.org/abs/1910.14537
AUTHORS:        Sufeng Duan ; Hai Zhao
HIGHLIGHT:      Taking greedy decoding algorithm as it should be, this work focuses on further strengthening the model itself for Chinese word segmentation (CWS), which results in an even more fast and more accurate CWS model.

4, TITLE:       Contrastive Multi-document Question Generation
http://arxiv.org/abs/1911.03047
AUTHORS:        Woon Sang Cho ; Yizhe Zhang ; Sudha Rao ; Asli Celikyilmaz ; Chenyan Xiong ; Jianfeng Gao ; Mengdi Wang ; Bill Dolan
HIGHLIGHT:      To generate such specific questions, we propose Multi-Source Coordinated Question Generator (MSCQG), a novel framework that includes a supervised learning (SL) stage and a reinforcement learning (RL) stage.

5, TITLE:       TRADI: Tracking deep neural network weight distributions
http://arxiv.org/abs/1912.11316
AUTHORS:        Gianni Franchi ; Andrei Bursuc ; Emanuel Aldea ; Severine Dubuisson ; Isabelle Bloch
HIGHLIGHT:      In this work we propose to make use of this knowledge and leverage it for computing the distributions of the weights of the DNN.

6, TITLE:       Structure-Level Knowledge Distillation For Multilingual Sequence Labeling
http://arxiv.org/abs/2004.03846
AUTHORS:        Xinyu Wang ; Yong Jiang ; Nguyen Bach ; Tao Wang ; Fei Huang ; Kewei Tu
COMMENTS:       Accepted to ACL 2020, camera-ready. 14 pages
HIGHLIGHT:      In this paper, we propose to reduce the gap between monolingual models and the unified multilingual model by distilling the structural knowledge of several monolingual models (teachers) to the unified multilingual model (student).

7, TITLE:       A Global Benchmark of Algorithms for Segmenting Late Gadolinium-Enhanced Cardiac Magnetic Resonance Imaging
http://arxiv.org/abs/2004.12314
AUTHORS:        Zhaohan Xiong ; Qing Xia ; Zhiqiang Hu ; Ning Huang ; Cheng Bian ; Yefeng Zheng ; Sulaiman Vesal ; Nishant Ravikumar ; Andreas Maier ; Xin Yang ; Pheng-Ann Heng ; Dong Ni ; Caizi Li ; Qianqian Tong ; Weixin Si ; Elodie Puybareau ; Younes Khoudli ; Thierry Geraud ; Chen Chen ; Wenjia Bai ; Daniel Rueckert ; Lingchao Xu ; Xiahai Zhuang ; Xinzhe Luo ; Shuman Jia ; Maxime Sermesant ; Yashu Liu ; Kuanquan Wang ; Davide Borra ; Alessandro Masci ; Cristiana Corsi ; Coen de Vente ; Mitko Veta ; Rashed Karim ; Chandrakanth Jayachandran Preetha ; Sandy Engelhardt ; Menyun Qiao ; Yuanyuan Wang ; Qian Tao ; Marta Nunez-Garcia ; Oscar Camara ; Nicolo Savioli ; Pablo Lamata ; Jichao Zhao
HIGHLIGHT:      In this paper, extensive analysis of the submitted algorithms using technical and biological metrics was performed by undergoing subgroup analysis and conducting hyper-parameter analysis, offering an overall picture of the major design choices of convolutional neural networks (CNNs) and practical considerations for achieving state-of-the-art left atrium segmentation.

8, TITLE:       Dynamic Fusion Network for Multi-Domain End-to-end Task-Oriented Dialog
http://arxiv.org/abs/2004.11019
AUTHORS:        Libo Qin ; Xiao Xu ; Wanxiang Che ; Yue Zhang ; Ting Liu
COMMENTS:       ACL2020
HIGHLIGHT:      To this end, we investigate methods that can make explicit use of domain knowledge and introduce a shared-private network to learn shared and specific knowledge.

9, TITLE:       Active Learning for Coreference Resolution using Discrete Annotation
http://arxiv.org/abs/2004.13671
AUTHORS:        Belinda Z. Li ; Gabriel Stanovsky ; Luke Zettlemoyer
COMMENTS:       12 pages, 7 figures, ACL 2020; fixed ablations table
HIGHLIGHT:      We improve upon pairwise annotation for active learning in coreference resolution, by asking annotators to identify mention antecedents if a presented mention pair is deemed not coreferent.

10, TITLE:      DiVA: Diverse Visual Feature Aggregation for Deep Metric Learning
http://arxiv.org/abs/2004.13458
AUTHORS:        Timo Milbich ; Karsten Roth ; Homanga Bharadhwaj ; Samarth Sinha ; Yoshua Bengio ; Björn Ommer ; Joseph Paul Cohen
COMMENTS:       18 pages
HIGHLIGHT:      To this end, we propose and study multiple complementary learning tasks, targeting conceptually different data relationships by only resorting to the available training samples and labels of a standard DML setting.

11, TITLE:      Quantum Adiabatic Algorithm Design using Reinforcement Learning
http://arxiv.org/abs/1812.10797
AUTHORS:        Jian Lin ; Zhong Yuan Lai ; Xiaopeng Li
COMMENTS:       11 pages, 10 figures
HIGHLIGHT:      We benchmark this approach in Grover-search and 3-SAT problems, and find that the adiabatic-algorithm obtained by our RL approach leads to significant improvement in the resultant success probability.

12, TITLE:      The Responsibility Quantification (ResQu) Model of Human Interaction with Automation
http://arxiv.org/abs/1810.12644
AUTHORS:        Nir Douer ; Joachim Meyer
HIGHLIGHT:      Using Information Theory, we develop a responsibility quantification (ResQu) model of human involvement in intelligent automated systems and demonstrate its applications on decisions regarding AWS.

13, TITLE:      BERTRAM: Improved Word Embeddings Have Big Impact on Contextualized Model Performance
http://arxiv.org/abs/1910.07181
AUTHORS:        Timo Schick ; Hinrich Schütze
COMMENTS:       Accepted at ACL2020
HIGHLIGHT:      In this work, we transfer this idea to pretrained language models: We introduce BERTRAM, a powerful architecture based on BERT that is capable of inferring high-quality embeddings for rare words that are suitable as input representations for deep language models.

14, TITLE:      Deepening Hidden Representations from Pre-trained Language Models
http://arxiv.org/abs/1911.01940
AUTHORS:        Junjie Yang ; Hai Zhao
HIGHLIGHT:      Transformer-based pre-trained language models have proven to be effective for learning contextualized language representation.

15, TITLE:      BLEU Neighbors: A Reference-less Approach to Automatic Evaluation
http://arxiv.org/abs/2004.12726
AUTHORS:        Kawin Ethayarajh ; Dorsa Sadigh
HIGHLIGHT:      To this end, we propose BLEU Neighbors, a nearest neighbors model for estimating language quality by using the BLEU score as a kernel function.

16, TITLE:      Zero-Shot Cross-Lingual Transfer with Meta Learning
http://arxiv.org/abs/2003.02739
AUTHORS:        Farhad Nooralahzadeh ; Giannis Bekoulis ; Johannes Bjerva ; Isabelle Augenstein
HIGHLIGHT:      We show that this challenging setup can be approached using meta-learning, where, in addition to training a source language model, another model learns to select which training instances are the most beneficial to the first.

17, TITLE:      Parsing All: Syntax and Semantics, Dependencies and Spans
http://arxiv.org/abs/1908.11522
AUTHORS:        Junru Zhou ; Zuchao Li ; Hai Zhao
COMMENTS:       arXiv admin note: text overlap with arXiv:1907.02684
HIGHLIGHT:      In this paper, we propose a novel joint model of syntactic and semantic parsing on both span and dependency representations, which incorporates syntactic information effectively in the encoder of neural network and benefits from two representation formalisms in a uniform way.

18, TITLE:      Fixed Encoder Self-Attention Patterns in Transformer-Based Machine Translation
http://arxiv.org/abs/2002.10260
AUTHORS:        Alessandro Raganato ; Yves Scherrer ; Jörg Tiedemann
HIGHLIGHT:      In this paper, we propose to replace all but one attention head of each encoder layer with simple fixed -- non-learnable -- attentive patterns that are solely based on position and do not require any external knowledge.

19, TITLE:      Minimal Solvers for Rectifying from Radially-Distorted Conjugate Translations
http://arxiv.org/abs/1911.01507
AUTHORS:        James Pritts ; Zuzana Kukelova ; Viktor Larsson ; Yaroslava Lochman ; Ondřej Chum
HIGHLIGHT:      This paper introduces minimal solvers that jointly solve for affine-rectification and radial lens undistortion from the image of translated and reflected coplanar features.

20, TITLE:      Enhancing Pre-trained Chinese Character Representation with Word-aligned Attention
http://arxiv.org/abs/1911.02821
AUTHORS:        Yanzeng Li ; Bowen Yu ; Mengge Xue ; Tingwen Liu
COMMENTS:       Accepted to appear at ACL 2020
HIGHLIGHT:      Hence, we propose a novel word-aligned attention to exploit explicit word information, which is complementary to various character-based Chinese pre-trained language models.

21, TITLE:      On the complexity of zero gap MIP*
http://arxiv.org/abs/2002.10490
AUTHORS:        Hamoon Mousavi ; Seyed Sajjad Nezhadi ; Henry Yuen
COMMENTS:       Fixed typos and edited protocol to more smoothly follow from references
HIGHLIGHT:      In this paper we investigate the complexity of deciding whether the quantum value of a non-local game $G$ is exactly $1$.

22, TITLE:      Improving BERT with Self-Supervised Attention
http://arxiv.org/abs/2004.03808
AUTHORS:        Xiaoyu Kou ; Yaming Yang ; Yujing Wang ; Ce Zhang ; Yiren Chen ; Yunhai Tong ; Yan Zhang ; Jing Bai
HIGHLIGHT:      In this paper, we propose a novel technique, called Self-Supervised Attention (SSA) to help facilitate this generalization challenge.

23, TITLE:      Light Field Synthesis by Training Deep Network in the Refocused Image Domain
http://arxiv.org/abs/1910.06072
AUTHORS:        Chang-Le Liu ; Kuang-Tsu Shih ; Jiun-Woei Huang ; Homer H. Chen
COMMENTS:       Accepted to IEEE Transactions on Image Processing
HIGHLIGHT:      In this paper, we propose a new loss function called refocused image error (RIE) to address the issue.

24, TITLE:      Do We Really Need to Access the Source Data? Source Hypothesis Transfer for Unsupervised Domain Adaptation
http://arxiv.org/abs/2002.08546
AUTHORS:        Jian Liang ; Dapeng Hu ; Jiashi Feng
HIGHLIGHT:      In this work we tackle a novel setting where only a trained source model is available and investigate how we can effectively utilize such a model without source data to solve UDA problems.

25, TITLE:      Local Search is a Remarkably Strong Baseline for Neural Architecture Search
http://arxiv.org/abs/2004.08996
AUTHORS:        T. Den Ottelander ; A. Dushatskiy ; M. Virgolin ; P. A. N. Bosman
COMMENTS:       18 pages, 13 figures
HIGHLIGHT:      In this work we consider, for the first time, a simple Local Search (LS) algorithm for NAS.

26, TITLE:      Null It Out: Guarding Protected Attributes by Iterative Nullspace Projection
http://arxiv.org/abs/2004.07667
AUTHORS:        Shauli Ravfogel ; Yanai Elazar ; Hila Gonen ; Michael Twiton ; Yoav Goldberg
COMMENTS:       Accepted as a long paper in ACL 2020
HIGHLIGHT:      We present Iterative Null-space Projection (INLP), a novel method for removing information from neural representations.

27, TITLE:      Scene Text Recognition via Transformer
http://arxiv.org/abs/2003.08077
AUTHORS:        Xinjie Feng ; Hongxun Yao ; Yuankai Qi ; Jun Zhang ; Shengping Zhang
COMMENTS:       We found that there are some errors in the experiment code, and we are correcting the result temporarily, so we temporarily withdraw this paper
HIGHLIGHT:      In this paper, we find that the rectification is completely unnecessary.

28, TITLE:      The Dialogue Dodecathlon: Open-Domain Knowledge and Image Grounded Conversational Agents
http://arxiv.org/abs/1911.03768
AUTHORS:        Kurt Shuster ; Da Ju ; Stephen Roller ; Emily Dinan ; Y-Lan Boureau ; Jason Weston
COMMENTS:       ACL 2020
HIGHLIGHT:      We introduce dodecaDialogue: a set of 12 tasks that measures if a conversational agent can communicate engagingly with personality and empathy, ask questions, answer questions by utilizing knowledge resources, discuss topics and situations, and perceive and converse about images.

29, TITLE:      A Comprehensive Review of Shepherding as a Bio-inspired Swarm-Robotics Guidance Approach
http://arxiv.org/abs/1912.07796
AUTHORS:        Nathan K Long ; Karl Sammut ; Daniel Sgarioto ; Matthew Garratt ; Hussein Abbass
COMMENTS:       Copyright 2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works
HIGHLIGHT:      We present a comprehensive review of the literature on swarm shepherding to reveal the advantages and potential of the approach to be applied to a plethora of robotic systems in the future.

30, TITLE:      Generate, Delete and Rewrite: A Three-Stage Framework for Improving Persona Consistency of Dialogue Generation
http://arxiv.org/abs/2004.07672
AUTHORS:        Haoyu Song ; Yan Wang ; Wei-Nan Zhang ; Xiaojiang Liu ; Ting Liu
COMMENTS:       Accepted by ACL2020
HIGHLIGHT:      In this work, we introduce a three-stage framework that employs a generate-delete-rewrite mechanism to delete inconsistent words from a generated response prototype and further rewrite it to a personality-consistent one.

31, TITLE:      A Comprehensive Survey on Traffic Prediction
http://arxiv.org/abs/2004.08555
AUTHORS:        Xueyan Yin ; Genze Wu ; Jinze Wei ; Yanming Shen ; Heng Qi ; Baocai Yin
HIGHLIGHT:      The purpose of this paper is to provide a comprehensive survey for traffic prediction.

32, TITLE:      Practical Comparable Data Collection for Low-Resource Languages via Images
http://arxiv.org/abs/2004.11954
AUTHORS:        Aman Madaan ; Shruti Rijhwani ; Antonios Anastasopoulos ; Yiming Yang ; Graham Neubig
COMMENTS:       Accepted for poster presentation at the Practical Machine Learning for Developing Countries (PML4DC) workshop, ICLR 2020
HIGHLIGHT:      We propose a method of curating high-quality comparable training data for low-resource languages with monolingual annotators.

33, TITLE:      Deep SCNN-based Real-time Object Detection for Self-driving Vehicles Using LiDAR Temporal Data
http://arxiv.org/abs/1912.07906
AUTHORS:        Shibo Zhou ; Ying Chen ; Xiaohua Li ; Arindam Sanyal
HIGHLIGHT:      In this paper, we integrate spiking convolutional neural network (SCNN) with temporal coding into the YOLOv2 architecture for real-time object detection. To take the advantage of spiking signals, we develop a novel data preprocessing layer that translates 3D point-cloud data into spike time data.

34, TITLE:      Gradient Surgery for Multi-Task Learning
http://arxiv.org/abs/2001.06782
AUTHORS:        Tianhe Yu ; Saurabh Kumar ; Abhishek Gupta ; Sergey Levine ; Karol Hausman ; Chelsea Finn
COMMENTS:       Code is available at https://github.com/tianheyu927/PCGrad
HIGHLIGHT:      In this work, we identify a set of three conditions of the multi-task optimization landscape that cause detrimental gradient interference, and develop a simple yet general approach for avoiding such interference between task gradients.

35, TITLE:      Predicting Landslides Using Contour Aligning Convolutional Neural Networks
http://arxiv.org/abs/1911.04651
AUTHORS:        Ainaz Hajimoradlou ; Gioachino Roberti ; David Poole
HIGHLIGHT:      We propose a model called Locally Aligned Convolutional Neural Network, LACNN, that follows the ground surface at multiple scales to predict possible landslide occurrence for a single point. To validate our method, we created a standardized dataset of georeferenced images consisting of the heterogeneous features as inputs, and compared our method to several baselines, including linear regression, a neural network, and a convolutional network, using log-likelihood error and Receiver Operating Characteristic curves on the test set.

36, TITLE:      Attacks on Image Encryption Schemes for Privacy-Preserving Deep Neural Networks
http://arxiv.org/abs/2004.13263
AUTHORS:        Alex Habeen Chang ; Benjamin M. Case
COMMENTS:       For associated code, see https://github.com/ahchang98/image-encryption-scheme-attacks
HIGHLIGHT:      We present new chosen-plaintext and ciphertext-only attacks against both of these proposed image encryption schemes and demonstrate the attacks' effectiveness on several examples.

37, TITLE:      How Does NLP Benefit Legal System: A Summary of Legal Artificial Intelligence
http://arxiv.org/abs/2004.12158
AUTHORS:        Haoxi Zhong ; Chaojun Xiao ; Cunchao Tu ; Tianyang Zhang ; Zhiyuan Liu ; Maosong Sun
COMMENTS:       Accepted by ACL 2020
HIGHLIGHT:      In this paper, we introduce the history, the current state, and the future directions of research in LegalAI.

38, TITLE:      VD-BERT: A Unified Vision and Dialog Transformer with BERT
http://arxiv.org/abs/2004.13278
AUTHORS:        Yue Wang ; Shafiq Joty ; Michael R. Lyu ; Irwin King ; Caiming Xiong ; Steven C. H. Hoi
COMMENTS:       15 pages, 7 figures, 4 tables
HIGHLIGHT:      By contrast, in this work, we propose VD-BERT, a simple yet effective framework of unified vision-dialog Transformer that leverages the pretrained BERT language models for Visual Dialog tasks.

39, TITLE:      DualSMC: Tunneling Differentiable Filtering and Planning under Continuous POMDPs
http://arxiv.org/abs/1909.13003
AUTHORS:        Yunbo Wang ; Bo Liu ; Jiajun Wu ; Yuke Zhu ; Simon S. Du ; Li Fei-Fei ; Joshua B. Tenenbaum
COMMENTS:       IJCAI 2020
HIGHLIGHT:      We cast POMDP filtering and planning problems as two closely related Sequential Monte Carlo (SMC) processes, one over the real states and the other over the future optimal trajectories, and combine the merits of these two parts in a new model named the DualSMC network.

40, TITLE:      Robust Lane Detection from Continuous Driving Scenes Using Deep Neural Networks
http://arxiv.org/abs/1903.02193
AUTHORS:        Qin Zou ; Hanwen Jiang ; Qiyu Dai ; Yuanhao Yue ; Long Chen ; Qian Wang
COMMENTS:       IEEE Transactions on Vehicular Technology, 69(1), 2020
HIGHLIGHT:      In recent years, many sophisticated lane detection methods have been proposed.

41, TITLE:      Standardizing and Benchmarking Crisis-related Social Media Datasets for Humanitarian Information Processing
http://arxiv.org/abs/2004.06774
AUTHORS:        Firoj Alam ; Hassan Sajjad ; Muhammad Imran ; Ferda Ofli
HIGHLIGHT:      In this work, we attempt to bridge this gap by standardizing various existing crisis-related datasets.

42, TITLE:      FastBERT: a Self-distilling BERT with Adaptive Inference Time
http://arxiv.org/abs/2004.02178
AUTHORS:        Weijie Liu ; Peng Zhou ; Zhe Zhao ; Zhiruo Wang ; Haotang Deng ; Qi Ju
COMMENTS:       This manuscript has been accepted to appear at ACL 2020
HIGHLIGHT:      To improve their efficiency with an assured model performance, we propose a novel speed-tunable FastBERT with adaptive inference time.

43, TITLE:      Self-training with Noisy Student improves ImageNet classification
http://arxiv.org/abs/1911.04252
AUTHORS:        Qizhe Xie ; Minh-Thang Luong ; Eduard Hovy ; Quoc V. Le
COMMENTS:       CVPR 2020
HIGHLIGHT:      We present Noisy Student Training, a semi-supervised learning approach that works well even when labeled data is abundant.

44, TITLE:      Distance Guided Channel Weighting for Semantic Segmentation
http://arxiv.org/abs/2004.12679
AUTHORS:        Lanyun Zhu ; Shiping Zhu ; Xuanyi Liu ; Li Luo
HIGHLIGHT:      In this paper, we address above issue by introducing the Distance Guided Channel Weighting (DGCW) Module.

45, TITLE:      ConvLab-2: An Open-Source Toolkit for Building, Evaluating, and Diagnosing Dialogue Systems
http://arxiv.org/abs/2002.04793
AUTHORS:        Qi Zhu ; Zheng Zhang ; Yan Fang ; Xiang Li ; Ryuichi Takanobu ; Jinchao Li ; Baolin Peng ; Jianfeng Gao ; Xiaoyan Zhu ; Minlie Huang
COMMENTS:       Accepted by ACL 2020 demo track
HIGHLIGHT:      We present ConvLab-2, an open-source toolkit that enables researchers to build task-oriented dialogue systems with state-of-the-art models, perform an end-to-end evaluation, and diagnose the weakness of systems.

46, TITLE:      Do We Need Fully Connected Output Layers in Convolutional Networks?
http://arxiv.org/abs/2004.13587
AUTHORS:        Zhongchao Qian ; Tyler L. Hayes ; Kushal Kafle ; Christopher Kanan
HIGHLIGHT:      Do We Need Fully Connected Output Layers in Convolutional Networks?

47, TITLE:      Question Answering over Curated and Open Web Sources
http://arxiv.org/abs/2004.11980
AUTHORS:        Rishiraj Saha Roy ; Avishek Anand
COMMENTS:       SIGIR 2020 Tutorial
HIGHLIGHT:      This tutorial would cover the highlights of this really active period of growth for QA to give the audience a grasp over the families of algorithms that are currently being used.

48, TITLE:      Learning Bayesian networks from demographic and health survey data
http://arxiv.org/abs/1912.00715
AUTHORS:        Neville Kenneth Kitson ; Anthony C. Constantinou
HIGHLIGHT:      We combine knowledge with available Demographic and Health Survey (DHS) data from India, to construct Causal Bayesian Networks (CBNs) and investigate the factors associated with childhood diarrhoea.

49, TITLE:      Accelerating Generative Neural Networks on Unmodified Deep Learning Processors -- A Software Approach
http://arxiv.org/abs/1907.01773
AUTHORS:        Dawen Xu ; Ying Wang ; Kaijie Tu ; Cheng Liu ; Bingsheng He ; Lei Zhang
COMMENTS:       13 pages
HIGHLIGHT:      In contrast, this work proposes a novel deconvolution implementation with a software approach and enables fast and efficient deconvolution execution on the legacy deep learning processors.

50, TITLE:      Closed-loop Matters: Dual Regression Networks for Single Image Super-Resolution
http://arxiv.org/abs/2003.07018
AUTHORS:        Yong Guo ; Jian Chen ; Jingdong Wang ; Qi Chen ; Jiezhang Cao ; Zeshuai Deng ; Yanwu Xu ; Mingkui Tan
COMMENTS:       This paper is accepted by CVPR 2020
HIGHLIGHT:      To address the above issues, we propose a dual regression scheme by introducing an additional constraint on LR data to reduce the space of the possible functions.

51, TITLE:      Projected Neural Network for a Class of Sparse Regression with Cardinality Penalty
http://arxiv.org/abs/2004.00858
AUTHORS:        Wenjing Li ; Wei Bian
HIGHLIGHT:      In this paper, we consider a class of sparse regression problems, whose objective function is the summation of a convex loss function and a cardinality penalty.

52, TITLE:      Simplifying Casts and Coercions
http://arxiv.org/abs/2001.10594
AUTHORS:        Robert Y. Lewis ; Paul-Nicolas Madelaine
HIGHLIGHT:      This paper introduces norm_cast, a toolbox of tactics for the Lean proof assistant designed to manipulate expressions containing coercions and casts.

53, TITLE:      Optimal Streaming Approximations for all Boolean Max-2CSPs
http://arxiv.org/abs/2004.11796
AUTHORS:        Chi-Ning Chou ; Alexander Golovnev ; Santhoshini Velusamy
COMMENTS:       Fixing typos
HIGHLIGHT:      We prove tight upper and lower bounds on approximation ratios of all Boolean Max-2CSP problems in the streaming model.

54, TITLE:      AMR Parsing via Graph-Sequence Iterative Inference
http://arxiv.org/abs/2004.05572
AUTHORS:        Deng Cai ; Wai Lam
COMMENTS:       ACL2020
HIGHLIGHT:      We propose a new end-to-end model that treats AMR parsing as a series of dual decisions on the input sequence and the incrementally constructed graph.

55, TITLE:      Domain Adaptive Transfer Attack (DATA)-based Segmentation Networks for Building Extraction from Aerial Images
http://arxiv.org/abs/2004.11819
AUTHORS:        Younghwan Na ; Jun Hee Kim ; Kyungsu Lee ; Juhum Park ; Jae Youn Hwang ; Jihwan P. Choi
COMMENTS:       11pages, 12 figures
HIGHLIGHT:      In this paper, we propose segmentation networks based on a domain adaptive transfer attack (DATA) scheme for building extraction from aerial images.

56, TITLE:      GRASPEL: Graph Spectral Learning at Scale
http://arxiv.org/abs/1911.10373
AUTHORS:        Yongyu Wang ; Zhiqiang Zhao ; Zhuo Feng
HIGHLIGHT:      In this work, for the first time, we present a highly-scalable spectral approach (GRASPEL) for learning large graphs from data.

57, TITLE:      Fast(er) Reconstruction of Shredded Text Documents via Self-Supervised Deep Asymmetric Metric Learning
http://arxiv.org/abs/2003.10063
AUTHORS:        Thiago M. Paixão ; Rodrigo F. Berriel ; Maria C. S. Boeres ; Alessando L. Koerich ; Claudine Badue ; Alberto F. De Souza ; Thiago Oliveira-Santos
COMMENTS:       Accepted to CVPR 2020. Main Paper (9 pages, 10 figures) and Supplementary Material (5 pages, 9 figures)
HIGHLIGHT:      This work proposes a scalable deep learning approach for measuring pairwise compatibility in which the number of inferences scales linearly (rather than quadratically) with the number of shreds.

58, TITLE:      Small-Object Detection in Remote Sensing Images with End-to-End Edge-Enhanced GAN and Object Detector Network
http://arxiv.org/abs/2003.09085
AUTHORS:        Jakaria Rabbi ; Nilanjan Ray ; Matthias Schubert ; Subir Chowdhury ; Dennis Chao
COMMENTS:       This paper contains 27 pages and accepted for publication in MDPI remote sensing journal. GitHub Repository: https://github.com/Jakaria08/EESRGAN (Implementation)
HIGHLIGHT:      We propose an architecture with three components: ESRGAN, Edge Enhancement Network (EEN), and Detection network.

59, TITLE:      Making Smart Homes Smarter: Optimizing Energy Consumption with Human in the Loop
http://arxiv.org/abs/1912.03298
AUTHORS:        Mudit Verma ; Siddhant Bhambri ; Saurabh Gupta ; Arun Balaji Buduru
HIGHLIGHT:      This paper presents a novel approach to accurately configure IoT devices while achieving the twin objectives of energy optimization along with conforming to user preferences.

60, TITLE:      ConveRT: Efficient and Accurate Conversational Representations from Transformers
http://arxiv.org/abs/1911.03688
AUTHORS:        Matthew Henderson ; Iñigo Casanueva ; Nikola Mrkšić ; Pei-Hao Su ; Tsung-Hsien Wen ; Ivan Vulić
HIGHLIGHT:      We propose ConveRT (Conversational Representations from Transformers), a pretraining framework for conversational tasks satisfying all the following requirements: it is effective, affordable, and quick to train.

61, TITLE:      Subjective Annotation for a Frame Interpolation Benchmark using Artefact Amplification
http://arxiv.org/abs/2001.06409
AUTHORS:        Hui Men ; Vlad Hosu ; Hanhe Lin ; Andrés Bruhn ; Dietmar Saupe
COMMENTS:       arXiv admin note: text overlap with arXiv:1901.05362
HIGHLIGHT:      To increase the sensitivity of observers when judging minute difference in paired comparisons we introduced a new method to the field of full-reference quality assessment, called artefact amplification.

62, TITLE:      Gradual Channel Pruning while Training using Feature Relevance Scores for Convolutional Neural Networks
http://arxiv.org/abs/2002.09958
AUTHORS:        Sai Aparna Aketi ; Sourjya Roy ; Anand Raghunathan ; Kaushik Roy
COMMENTS:       15 pages, 2 figures, 4 tables
HIGHLIGHT:      To address all the above issues, we present a simple-yet-effective gradual channel pruning while training methodology using a novel data-driven metric referred to as feature relevance score.

63, TITLE:      From English To Foreign Languages: Transferring Pre-trained Language Models
http://arxiv.org/abs/2002.07306
AUTHORS:        Ke Tran
HIGHLIGHT:      In this work, we tackle the problem of transferring an existing pre-trained model from English to other languages under a limited computational budget.

64, TITLE:      Improving Disfluency Detection by Self-Training a Self-Attentive Model
http://arxiv.org/abs/2004.05323
AUTHORS:        Paria Jamshid Lou ; Mark Johnson
HIGHLIGHT:      Since the contextualized word embeddings are pre-trained on a large amount of unlabeled data, using additional unlabeled data to train a neural model might seem redundant.

65, TITLE:      POCOVID-Net: Automatic Detection of COVID-19 From a New Lung Ultrasound Imaging Dataset (POCUS)
http://arxiv.org/abs/2004.12084
AUTHORS:        Jannis Born ; Gabriel Brändle ; Manuel Cossio ; Marion Disdier ; Julie Goulet ; Jérémie Roulin ; Nina Wiedemann
COMMENTS:       7 pages, 4 figures
HIGHLIGHT:      Our contribution is threefold.