2020.07.27.txt

==========New Papers==========
1, TITLE:       Commonality-Parsing Network across Shape and Appearance for Partially Supervised Instance Segmentation
http://arxiv.org/abs/2007.12387
AUTHORS:        Qi Fan ; Lei Ke ; Wenjie Pei ; Chi-Keung Tang ; Yu-Wing Tai
COMMENTS:       Accepted by ECCV 2020
HIGHLIGHT:      We propose to learn the underlying class-agnostic commonalities that can be generalized from mask-annotated categories to novel categories.

2, TITLE:       Body2Hands: Learning to Infer 3D Hands from Conversational Gesture Body Dynamics
http://arxiv.org/abs/2007.12287
AUTHORS:        Evonne Ng ; Hanbyul Joo ; Shiry Ginosar ; Trevor Darrell
HIGHLIGHT:      We propose a novel learned deep prior of body motion for 3D hand shape synthesis and estimation in the domain of conversational gestures.

3, TITLE:       Frequency Domain-based Perceptual Loss for Super Resolution
http://arxiv.org/abs/2007.12296
AUTHORS:        Shane D. Sims
HIGHLIGHT:      We introduce Frequency Domain Perceptual Loss (FDPL), a loss function for single image super resolution (SR).

4, TITLE:       Learning the Solution Manifold in Optimization and Its Application in Motion Planning
http://arxiv.org/abs/2007.12397
AUTHORS:        Takayuki Osa
HIGHLIGHT:      To address this issue, we propose to learn the solution manifold in optimization.

5, TITLE:       MurTree: Optimal Classification Trees via Dynamic Programming and Search
http://arxiv.org/abs/2007.12652
AUTHORS:        Emir Demirović ; Anna Lukina ; Emmanuel Hebrard ; Jeffrey Chan ; James Bailey ; Christopher Leckie ; Kotagiri Ramamohanarao ; Peter J. Stuckey
HIGHLIGHT:      We follow this line of work and provide a novel algorithm for learning optimal classification trees based on dynamic programming and search.

6, TITLE:       IR-BERT: Leveraging BERT for Semantic Search in Background Linking for News Articles
http://arxiv.org/abs/2007.12603
AUTHORS:        Anup Anand Deshmukh ; Udhav Sethi
COMMENTS:       6 pages, 6 figures
HIGHLIGHT:      This work describes our two approaches for the background linking task of TREC 2020 News Track.

7, TITLE:       Model Checkers Are Cool: How to Model Check Voting Protocols in Uppaal
http://arxiv.org/abs/2007.12412
AUTHORS:        Wojciech Jamroga ; Yan Kim ; Damian Kurpiewski ; Peter Y. A. Ryan
HIGHLIGHT:      In this paper, we propose that the state-of-art model checker Uppaal provides a good environment for modelling and preliminary verification of voting protocols.

8, TITLE:       What and Where: Learn to Plug Adapters via NAS for Multi-Domain Learning
http://arxiv.org/abs/2007.12415
AUTHORS:        Hanbin Zhao ; Hao Zeng ; Xin Qin ; Yongjian Fu ; Hui Wang ; Bourahla Omar ; Xi Li
HIGHLIGHT:      With this motivation, we propose to learn a data-driven adapter plugging strategy with Neural Architecture Search (NAS), which automatically determines where to plug for those adapter modules.

9, TITLE:       Real-World Multi-Domain Data Applications for Generalizations to Clinical Settings
http://arxiv.org/abs/2007.12672
AUTHORS:        Nooshin Mojab ; Vahid Noroozi ; Darvin Yi ; Manoj Prabhakar Nallabothula ; Abdullah Aleem ; Phillip S. Yu ; Joelle A. Hallak
HIGHLIGHT:      In this paper, we utilize self-supervised representation learning methods, formulated effectively in transfer learning settings, to address limited data availability.

10, TITLE:      COVID-19 Remote Patient Monitoring: Social Impact of AI
http://arxiv.org/abs/2007.12312
AUTHORS:        Ashlesha Nesarikar ; Waqas Haque ; Suchith Vuppala ; Abhijit Nesarikar
COMMENTS:       21 pages, 4 figures
HIGHLIGHT:      In a multifaceted technology approach, we start with effective technology use for remote patient monitoring (RPM) of COVID-19 with the following objectives: 1.

11, TITLE:      Visual Compositional Learning for Human-Object Interaction Detection
http://arxiv.org/abs/2007.12407
AUTHORS:        Zhi Hou ; Xiaojiang Peng ; Yu Qiao ; Dacheng Tao
COMMENTS:       Accepted in ECCV2020
HIGHLIGHT:      We devise a deep Visual Compositional Learning (VCL) framework, which is a simple yet efficient framework to effectively address this problem.

12, TITLE:      MULTISEM at SemEval-2020 Task 3: Fine-tuning BERT for Lexical Meaning
http://arxiv.org/abs/2007.12432
AUTHORS:        Aina Garí Soler ; Marianna Apidianaki
COMMENTS:       8 pages, 2 tables. Accepted at the 14th International Workshop on Semantic Evaluation (SemEval-2020)
HIGHLIGHT:      We present the MULTISEM systems submitted to SemEval 2020 Task 3: Graded Word Similarity in Context (GWSC).

13, TITLE:      Style Transfer for Co-Speech Gesture Animation: A Multi-Speaker Conditional-Mixture Approach
http://arxiv.org/abs/2007.12553
AUTHORS:        Chaitanya Ahuja ; Dong Won Lee ; Yukiko I. Nakano ; Louis-Philippe Morency
COMMENTS:       24 pages, 12 figures
HIGHLIGHT:      In this paper, we propose a new model, named Mix-StAGE, which trains a single model for multiple speakers while learning unique style embeddings for each speaker's gestures in an end-to-end manner. We also introduce a new dataset, Pose-Audio-Transcript-Style (PATS), designed to study gesture generation and style transfer.

14, TITLE:      The Representation Theory of Neural Networks
http://arxiv.org/abs/2007.12213
AUTHORS:        Marco Antonio Armenta ; Pierre-Marc Jodoin
HIGHLIGHT:      In this work, we show that neural networks can be represented via the mathematical theory of quiver representations.

15, TITLE:      Fully Convolutional Networks for Continuous Sign Language Recognition
http://arxiv.org/abs/2007.12402
AUTHORS:        Ka Leong Cheng ; Zhaoyang Yang ; Qifeng Chen ; Yu-Wing Tai
COMMENTS:       Accepted to ECCV2020
HIGHLIGHT:      In this paper, we propose a fully convolutional network (FCN) for online SLR to concurrently learn spatial and temporal features from weakly annotated video sequences with only sentence-level annotations given.

16, TITLE:      Clinician-in-the-Loop Decision Making: Reinforcement Learning with Near-Optimal Set-Valued Policies
http://arxiv.org/abs/2007.12678
AUTHORS:        Shengpu Tang ; Aditya Modi ; Michael W. Sjoding ; Jenna Wiens
COMMENTS:       ICML 2020. Code available at https://github.com/shengpu1126/RL-Set-Valued-Policy
HIGHLIGHT:      We propose a model-free algorithm based on temporal difference learning and a near-greedy heuristic for action selection.

17, TITLE:      Predictive Information Accelerates Learning in RL
http://arxiv.org/abs/2007.12401
AUTHORS:        Kuang-Huei Lee ; Ian Fischer ; Anthony Liu ; Yijie Guo ; Honglak Lee ; John Canny ; Sergio Guadarrama
HIGHLIGHT:      We hypothesize that capturing the predictive information is useful in RL, since the ability to model what will happen next is necessary for success on many tasks.

18, TITLE:      ZSCRGAN: A GAN-based Expectation Maximization Model for Zero-Shot Retrieval of Images from Textual Descriptions
http://arxiv.org/abs/2007.12212
AUTHORS:        Anurag Roy ; Vinay Kumar Verma ; Kripabandhu Ghosh ; Saptarshi Ghosh
HIGHLIGHT:      In this paper, we propose a novel GAN-based model for zero-shot text to image retrieval.

19, TITLE:      Study of Different Deep Learning Approach with Explainable AI for Screening Patients with COVID-19 Symptoms: Using CT Scan and Chest X-ray Image Dataset
http://arxiv.org/abs/2007.12525
AUTHORS:        Md Manjurul Ahsan ; Kishor Datta Gupta ; Mohammad Maminur Islam ; Sajib Sen ; Md. Lutfar Rahman ; Mohammad Shakhawat Hossain
COMMENTS:       This is a work in progress, it should not be relied upon without context to guide clinical practice or health-related behavior and should not be reported in news media as established information without consulting multiple experts in the field
HIGHLIGHT:      Some studies proposed CT scan or chest X-ray images as an alternative solution.

20, TITLE:      JUNLP@SemEval-2020 Task 9:Sentiment Analysis of Hindi-English code mixed data
http://arxiv.org/abs/2007.12561
AUTHORS:        Avishek Garain ; Sainik Kumar Mahata ; Dipankar Das
HIGHLIGHT:      In this work, we focus on working out a plausible solution to the domain of Code-Mixed Sentiment Analysis.

21, TITLE:      Stain Style Transfer of Histopathology Images Via Structure-Preserved Generative Learning
http://arxiv.org/abs/2007.12578
AUTHORS:        Hanwen Liang ; Konstantinos N. Plataniotis ; Xingyu Li
HIGHLIGHT:      To address the issue of color variations in histopathology images, this study proposes two stain style transfer models, SSIM-GAN and DSCSI-GAN, based on the generative adversarial networks.

22, TITLE:      A Lightweight Neural Network for Monocular View Generation with Occlusion Handling
http://arxiv.org/abs/2007.12577
AUTHORS:        Simon Evain ; Christine Guillemot
COMMENTS:       Accepted at IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) in December 2019
HIGHLIGHT:      In this article, we present a very lightweight neural network architecture, trained on stereo data pairs, which performs view synthesis from one single image.

23, TITLE:      Computing nearest neighbour interchange distances between ranked phylogenetic trees
http://arxiv.org/abs/2007.12307
AUTHORS:        Lena Collienne ; Alex Gavryushkin
HIGHLIGHT:      In this paper, we settle this problem for the ranked nearest neighbour interchange operation by establishing that the complexity depends on the weight difference between the two types of tree rearrangements (rank moves and edge moves), and varies from quadratic, which is the lowest possible complexity for this problem, to NP-hard, which is the highest.

24, TITLE:      Deforming the Loss Surface
http://arxiv.org/abs/2007.12515
AUTHORS:        Liangming Chen ; Long Jin ; Xiujuan Du ; Shuai Li ; Mei Liu
COMMENTS:       2020NIPS
HIGHLIGHT:      Differently, a novel concept of deformation operator is first proposed in this paper to deform the loss surface, thereby improving the optimization.

25, TITLE:      Value-Decomposition Multi-Agent Actor-Critics
http://arxiv.org/abs/2007.12306
AUTHORS:        Jianyu Su ; Stephen Adams ; Peter A. Beling
COMMENTS:       Submitting to aaai2021
HIGHLIGHT:      To obtain a reasonable trade-off between training efficiency and algorithm performance, we extend value-decomposition to actor-critics that are compatible with A2C and propose a novel actor-critic framework, value-decomposition actor-critics (VDACs).

26, TITLE:      Distributional Reinforcement Learning with Maximum Mean Discrepancy
http://arxiv.org/abs/2007.12354
AUTHORS:        Thanh Tang Nguyen ; Sunil Gupta ; Svetha Venkatesh
COMMENTS:       21 pages, 6 figures, 5 tables
HIGHLIGHT:      Most existing methods focus on learning a set of predefined statistic functionals of the return distributions requiring involved projections to maintain the order statistics.

27, TITLE:      HEU Emotion: A Large-scale Database for Multi-modal Emotion Recognition in the Wild
http://arxiv.org/abs/2007.12519
AUTHORS:        Jing Chen ; Chenhui Wang ; Kejun Wang ; Chaoqun Yin ; Cong Zhao ; Tao Xu ; Xinyi Zhang ; Ziqiang Huang ; Meichen Liu ; Tao Yang
HIGHLIGHT:      We proposed a Multi-modal Attention module to fuse multi-modal features adaptively.

28, TITLE:      Corpse Reviver: Sound and Efficient Gradual Typing via Contract Verification
http://arxiv.org/abs/2007.12630
AUTHORS:        Cameron Moy ; Phúc C. Nguyen ; Sam Tobin-Hochstadt ; David Van Horn
HIGHLIGHT:      In this paper, we show that by building on existing work on soft contract verification, we can reduce or eliminate this overhead.

29, TITLE:      Reparameterizing Convolutions for Incremental Multi-Task Learning without Task Interference
http://arxiv.org/abs/2007.12540
AUTHORS:        Menelaos Kanakis ; David Bruggemann ; Suman Saha ; Stamatios Georgoulis ; Anton Obukhov ; Luc Van Gool
COMMENTS:       European Conference on Computer Vision (ECCV), 2020
HIGHLIGHT:      In this paper, we show that both can be achieved simply by reparameterizing the convolutions of standard neural network architectures into a non-trainable shared part (filter bank) and task-specific parts (modulators), where each modulator has a fraction of the filter bank parameters.

30, TITLE:      Machine Learning Explanations to Prevent Overtrust in Fake News Detection
http://arxiv.org/abs/2007.12358
AUTHORS:        Sina Mohseni ; Fan Yang ; Shiva Pentyala ; Mengnan Du ; Yi Liu ; Nic Lupfer ; Xia Hu ; Shuiwang Ji ; Eric Ragan
HIGHLIGHT:      We present evaluation results and analysis from multiple controlled crowdsourced studies. We design a news reviewing and sharing interface, create a dataset of news stories, and train four interpretable fake news detection algorithms to study the effects of algorithmic transparency on end-users.

31, TITLE:      The foundations of cost-sensitive causal classification
http://arxiv.org/abs/2007.12582
AUTHORS:        Wouter Verbeke ; Diego Olaya ; Jeroen Berrevoets ; Sebastián Maldonado
HIGHLIGHT:      Cost-sensitive and causal classification methods have independently been proposed to improve the performance of classification models.

32, TITLE:      IDS at SemEval-2020 Task 10: Does Pre-trained Language Model Know What to Emphasize?
http://arxiv.org/abs/2007.12390
AUTHORS:        Jaeyoul Shin ; Taeuk Kim ; Sang-goo Lee
HIGHLIGHT:      We propose a novel method that enables us to determine words that deserve to be emphasized from written text in visual media, relying only on the information from the self-attention distributions of pre-trained language models (PLMs).

33, TITLE:      An LSTM Approach to Temporal 3D Object Detection in LiDAR Point Clouds
http://arxiv.org/abs/2007.12392
AUTHORS:        Rui Huang ; Wanyue Zhang ; Abhijit Kundu ; Caroline Pantofaru ; David A Ross ; Thomas Funkhouser ; Alireza Fathi
COMMENTS:       To appear in ECCV 2020
HIGHLIGHT:      To address this problem, in this paper we propose a sparse LSTM-based multi-frame 3d object detection algorithm.

34, TITLE:      Artificial Intelligence in the Creative Industries: A Review
http://arxiv.org/abs/2007.12391
AUTHORS:        Nantheera Anantrasirichai ; David Bull
COMMENTS:       A white paper for the Creative Industries Clusters Programme
HIGHLIGHT:      This paper reviews the current state of the art in Artificial Intelligence (AI) technologies and applications in the context of the creative industries.

35, TITLE:      Micro-expression spotting: A new benchmark
http://arxiv.org/abs/2007.12421
AUTHORS:        Thuong-Khanh Tran ; Quang-Nhat Vo ; Xiaopeng Hong ; Xiaobai Li ; Guoying Zhao
HIGHLIGHT:      Our contributions in this paper are three folds: (1) We introduce an extension of the SMIC-E database, namely SMIC-E-Long database, which is a new challenging benchmark for ME spotting.

36, TITLE:      COVID TV-UNet: Segmenting COVID-19 Chest CT Images Using Connectivity Imposed U-Net
http://arxiv.org/abs/2007.12303
AUTHORS:        Narges Saeedizadeh ; Shervin Minaee ; Rahele Kafieh ; Shakib Yazdani ; Milan Sonka
HIGHLIGHT:      In this work we propose a segmentation framework to detect chest regions in CT images, which are infected by COVID-19.

37, TITLE:      On the Effectiveness of Image Rotation for Open Set Domain Adaptation
http://arxiv.org/abs/2007.12360
AUTHORS:        Silvia Bucci ; Mohammad Reza Loghmani ; Tatiana Tommasi
COMMENTS:       accepted at ECCV 2020
HIGHLIGHT:      We propose a novel method to addresses both these problems using the self-supervised task of rotation recognition.

38, TITLE:      FiSSA at SemEval-2020 Task 9: Fine-tuned For Feelings
http://arxiv.org/abs/2007.12544
AUTHORS:        Bertelt Braaksma ; Richard Scholtens ; Stan van Suijlekom ; Remy Wang ; Ahmet Üstün
COMMENTS:       In Proceedings of the 14th International Workshop on Semantic Evaluation (SemEval-2020), Barcelona, Spain, December. Association for Computational Linguistics
HIGHLIGHT:      In this paper, we present our approach for sentiment classification on Spanish-English code-mixed social media data in the SemEval-2020 Task 9.

39, TITLE:      Learning Crisp Edge Detector Using Logical Refinement Network
http://arxiv.org/abs/2007.12449
AUTHORS:        Luyan Liu ; Kai Ma ; Yefeng Zheng
COMMENTS:       Accepted by MICCAI2020
HIGHLIGHT:      In this work, we propose a novel logical refinement network for crisp edge detection, which is motivated by the logical relationship between segmentation and edge maps and can be applied to both 2D and 3D images.

40, TITLE:      SummEval: Re-evaluating Summarization Evaluation
http://arxiv.org/abs/2007.12626
AUTHORS:        Alexander R. Fabbri ; Wojciech Kryściński ; Bryan McCann ; Richard Socher ; Dragomir Radev
COMMENTS:       10 pages, 4 tables, 1 figure
HIGHLIGHT:      We address the existing shortcomings of summarization evaluation methods along five dimensions: 1) we re-evaluate 12 automatic evaluation metrics in a comprehensive and consistent fashion using neural summarization model outputs along with expert and crowd-sourced human annotations, 2) we consistently benchmark 23 recent summarization models using the aforementioned automatic evaluation metrics, 3) we assemble the largest collection of summaries generated by models trained on the CNN/DailyMail news dataset and share it in a unified format, 4) we implement and share a toolkit that provides an extensible and unified API for evaluating summarization models across a broad range of automatic metrics, 5) we assemble and share the largest and most diverse, in terms of model types, collection of human judgments of model-generated summaries on the CNN/Daily Mail dataset annotated by both expert judges and crowd source workers.

41, TITLE:      Performance analysis of weighted low rank model with sparse image histograms for face recognition under lowlevel illumination and occlusion
http://arxiv.org/abs/2007.12362
AUTHORS:        K. V. Sridhar ; Raghu vamshi Hemadri
COMMENTS:       12 pages, 8 figres, 4 Tables, International conferences
HIGHLIGHT:      In this paper, a comparison of the low-rank recovery performance of two LRMA algorithms- RPCA and WSNM is brought out on occluded human facial images.

42, TITLE:      KPRNet: Improving projection-based LiDAR semantic segmentation
http://arxiv.org/abs/2007.12668
AUTHORS:        Deyvid Kochanov ; Fatemeh Karimi Nejadasl ; Olaf Booij
HIGHLIGHT:      In this work, we adopt recent advances in both image and point cloud segmentation to achieve a better accuracy in the task of segmenting LiDAR scans.

43, TITLE:      MiCo: Mixup Co-Training for Semi-Supervised Domain Adaptation
http://arxiv.org/abs/2007.12684
AUTHORS:        Luyu Yang ; Yan Wang ; Mingfei Gao ; Abhinav Shrivastava ; Kilian Q. Weinberger ; Wei-Lun Chao ; Ser-Nam Lim
HIGHLIGHT:      In this paper we propose a new approach for SSDA, which is to explicitly decompose SSDA into two sub-problems: a semi-supervised learning (SSL) problem in the target domain and an unsupervised domain adaptation (UDA) problem across domains.

44, TITLE:      SeismoGlow -- Data augmentation for the class imbalance problem
http://arxiv.org/abs/2007.12229
AUTHORS:        Ruy Luiz Milidiú ; Luis Felipe Müller
COMMENTS:       10 pages
HIGHLIGHT:      In this work, we propose the SeismoGlow a flow-based generative model to create synthetic samples, aiming to address the class imbalance. We introduce a dataset composed of5.223seismograms that are distributed between the good, medium, and bad classes and with their respective frequencies of 66.68%,31.54%, and 1.76%.

45, TITLE:      Globally Optimal Solution to Inverse Kinematics of 7DOF Serial Manipulator
http://arxiv.org/abs/2007.12550
AUTHORS:        Pavel Trutman ; Safey El Din Mohab ; Didier Henrion ; Tomas Pajdla
HIGHLIGHT:      We present a global solution to the optimal IK problem for a general serial 7DOF manipulator with revolute joints and a quadratic polynomial objective function.

46, TITLE:      Hallucinating Saliency Maps for Fine-Grained Image Classification for Limited Data Domains
http://arxiv.org/abs/2007.12562
AUTHORS:        Carola Figueroa-Flores ; Bogdan Raducanu ; David Berga ; Joost van de Weijer
HIGHLIGHT:      In the current paper, we propose an approach which does not require explicit saliency maps to improve image classification, but they are learned implicitely, during the training of an end-to-end image classification task.

47, TITLE:      CelebA-Spoof: Large-Scale Face Anti-Spoofing Dataset with Rich Annotations
http://arxiv.org/abs/2007.12342
AUTHORS:        Yuanhan Zhang ; Zhenfei Yin ; Yidong Li ; Guojun Yin ; Junjie Yan ; Jing Shao ; Ziwei Liu
COMMENTS:       To appear in ECCV 2020. Dataset is available at: https://github.com/Davidzhangyuanhan/CelebA-Spoof
HIGHLIGHT:      Equipped with CelebA-Spoof, we carefully benchmark existing methods in a unified multi-task framework, Auxiliary Information Embedding Network (AENet), and reveal several valuable observations.

48, TITLE:      Approximately Optimal Binning for the Piecewise Constant Approximation of the Normalized Unexplained Variance (nUV) Dissimilarity Measure
http://arxiv.org/abs/2007.12463
AUTHORS:        Attila Fazekas ; György Kovács
HIGHLIGHT:      By pointing out an important analogy between the well known mutual information (MI) and MTM, we introduce the term "normalized unexplained variance" (nUV) for MTM to emphasize its relevance and applicability beyond image processing.

49, TITLE:      Impact of Medical Data Imprecision on Learning Results
http://arxiv.org/abs/2007.12375
AUTHORS:        Mei Wang ; Jianwen Su ; Haiqin Lu
COMMENTS:       2020 KDD Workshop on Applied Data Science for Healthcare
HIGHLIGHT:      In this paper, we initiate a study on the impact of imprecision on prediction results in a healthcare application where a pre-trained model is used to predict future state of hyperthyroidism for patients.

50, TITLE:      Towards Recognizing Unseen Categories in Unseen Domains
http://arxiv.org/abs/2007.12256
AUTHORS:        Massimiliano Mancini ; Zeynep Akata ; Elisa Ricci ; Barbara Caputo
COMMENTS:       Accepted to ECCV 2020
HIGHLIGHT:      Towards Recognizing Unseen Categories in Unseen Domains

51, TITLE:      Machine-learned Regularization and Polygonization of Building Segmentation Masks
http://arxiv.org/abs/2007.12587
AUTHORS:        Stefano Zorzi ; Ksenia Bittner ; Friedrich Fraundorfer
HIGHLIGHT:      We propose a machine learning based approach for automatic regularization and polygonization of building segmentation masks.

52, TITLE:      A Comprehensive Study on Sign Language Recognition Methods
http://arxiv.org/abs/2007.12530
AUTHORS:        Nikolas Adaloglou ; Theocharis Chatzis ; Ilias Papastratis ; Andreas Stergioulas ; Georgios Th. Papadopoulos ; Vassia Zacharopoulou ; George J. Xydopoulos ; Klimnis Atzakas ; Dimitris Papazachariou ; Petros Daras
HIGHLIGHT:      In this paper, a comparative experimental assessment of computer vision-based methods for sign language recognition is conducted.

53, TITLE:      Tromino Tilings with Pegs via Flow Networks
http://arxiv.org/abs/2007.12651
AUTHORS:        Javier T. Akagi ; Eduardo A. Canale ; Marcos Villagra
COMMENTS:       12 pages, 7 figures
HIGHLIGHT:      In this work we study a slight variation of the tromino tiling problem where some positions of the region have pegs and each tromino comes with a hole that can only be placed on top of the pegs.

54, TITLE:      The Lottery Ticket Hypothesis for Pre-trained BERT Networks
http://arxiv.org/abs/2007.12223
AUTHORS:        Tianlong Chen ; Jonathan Frankle ; Shiyu Chang ; Sijia Liu ; Yang Zhang ; Zhangyang Wang ; Michael Carbin
HIGHLIGHT:      In this work, we combine these observations to assess whether such trainable, transferrable subnetworks exist in pre-trained BERT models.

55, TITLE:      Context-Aware Attentive Knowledge Tracing
http://arxiv.org/abs/2007.12324
AUTHORS:        Aritra Ghosh ; Neil Heffernan ; Andrew S. Lan
COMMENTS:       Published in KDD 2020
HIGHLIGHT:      In this paper, we propose attentive knowledge tracing (AKT), which couples flexible attention-based neural network models with a series of novel, interpretable model components inspired by cognitive and psychometric models.

56, TITLE:      Named entity recognition in chemical patents using ensemble of contextual language models
http://arxiv.org/abs/2007.12569
AUTHORS:        Jenny Copara ; Nona Naderi ; Julien Knafou ; Patrick Ruch ; Douglas Teodoro
HIGHLIGHT:      We compare transformer architectures trained on a generic corpus with models specialised in chemistry patents, and propose a new model based on the combination of existing architectures.

57, TITLE:      Interpreting Spatially Infinite Generative Models
http://arxiv.org/abs/2007.12411
AUTHORS:        Chaochao Lu ; Richard E. Turner ; Yingzhen Li ; Nate Kushman
COMMENTS:       ICML 2020 workshop on Human Interpretability in Machine Learning (WHI 2020)
HIGHLIGHT:      In this paper we provide a firm theoretical interpretation for infinite spatial generation, by drawing connections to spatial stochastic processes.

58, TITLE:      Towards Automated Discovery of Geometrical Theorems in GeoGebra
http://arxiv.org/abs/2007.12447
AUTHORS:        Zoltán Kovács ; Jonathan H. Yu
COMMENTS:       21 pages, 19 figures
HIGHLIGHT:      We describe a prototype of a new experimental GeoGebra command and tool Discover that analyzes geometric figures for salient patterns, properties, and theorems.

59, TITLE:      The Surprising Effectiveness of Linear Unsupervised Image-to-Image Translation
http://arxiv.org/abs/2007.12568
AUTHORS:        Eitan Richardson ; Yair Weiss
COMMENTS:       Preprint - under review
HIGHLIGHT:      In this paper we introduce linear encoder-decoder architectures for unsupervised image to image translation.

60, TITLE:      Locality-Aware Rotated Ship Detection in High-Resolution Remote Sensing Imagery Based on Multi-Scale Convolutional Network
http://arxiv.org/abs/2007.12326
AUTHORS:        Lingyi Liu ; Yunpeng Bai ; Ying Li
COMMENTS:       5 pages, 8 figures
HIGHLIGHT:      In this letter, we propose a locality-aware rotated ship detection (LARSD) framework based on a multi-scale convolutional neural network (CNN) to tackle these issues.

61, TITLE:      Unsupervised Discovery of 3D Physical Objects from Video
http://arxiv.org/abs/2007.12348
AUTHORS:        Yilun Du ; Kevin Smith ; Tomer Ulman ; Joshua Tenenbaum ; Jiajun Wu
HIGHLIGHT:      We study the problem of unsupervised physical object discovery.

62, TITLE:      Multi-view adaptive graph convolutions for graph classification
http://arxiv.org/abs/2007.12450
AUTHORS:        Nikolas Adaloglou ; Nicholas Vretos ; Petros Daras
COMMENTS:       Accepted as a poster on ECCV 2020, camera ready version
HIGHLIGHT:      In this paper, a novel multi-view methodology for graph-based neural networks is proposed.

63, TITLE:      Map-Repair: Deep Cadastre Maps Alignment and Temporal Inconsistencies Fix in Satellite Images
http://arxiv.org/abs/2007.12470
AUTHORS:        Stefano Zorzi ; Ksenia Bittner ; Friedrich Fraundorfer
HIGHLIGHT:      In this work, we propose an end-to-end deep learning approach which is able to solve inconsistencies between the input intensity image and the available building footprints by correcting label noises and, at the same time, misalignments if needed.

64, TITLE:      Are Visual Explanations Useful? A Case Study in Model-in-the-Loop Prediction
http://arxiv.org/abs/2007.12248
AUTHORS:        Eric Chu ; Deb Roy ; Jacob Andreas
HIGHLIGHT:      We present a randomized controlled trial for a model-in-the-loop regression task, with the goal of measuring the extent to which (1) good explanations of model predictions increase human accuracy, and (2) faulty explanations decrease human trust in the model.

65, TITLE:      Self-Supervised Learning Across Domains
http://arxiv.org/abs/2007.12368
AUTHORS:        Silvia Bucci ; Antonio D'Innocente ; Yujun Liao ; Fabio Maria Carlucci ; Barbara Caputo ; Tatiana Tommasi
HIGHLIGHT:      In this paper we propose to apply a similar approach to the problem of object recognition across domains: our model learns the semantic labels in a supervised fashion, and broadens its understanding of the data by learning from self-supervised signals on the same images.

66, TITLE:      On Solving Word Equations via Program Transformation
http://arxiv.org/abs/2007.12247
AUTHORS:        Antonina Nepeivoda
HIGHLIGHT:      The paper presents an experiment of solving word equations via specialization of a configuration WE(R,E), where the program WE can be considered as an interpreter testing whether a composition of substitutions R produces a solution of a word equation E. Several variants of such interpreters, when specialized using a basic unfold/fold strategy, are able to decide solvability for a number of sets of the word equations with the overlapping variables.

67, TITLE:      On the Parameterized Complexity of Synthesizing Boolean Petri Nets With Restricted Dependency (Technical Report)
http://arxiv.org/abs/2007.12372
AUTHORS:        Ronny Tredup ; Evgeny Erofeev
HIGHLIGHT:      In this paper, we show that DR$\tau$S parameterized by $d$ is in XP.

68, TITLE:      Dopant Network Processing Units: Towards Efficient Neural-network Emulators with High-capacity Nanoelectronic Nodes
http://arxiv.org/abs/2007.12371
AUTHORS:        Hans-Christian Ruiz Euler ; Unai Alegre-Ibarra ; Bram van de Ven ; Hajo Broersma ; Peter A. Bobbert ; Wilfred G. van der Wiel
HIGHLIGHT:      This paper presents a promising novel approach to neural information processing by introducing DNPUs as high-capacity neurons and moving from a single to a multi-neuron framework.

69, TITLE:      Learning Noise-Aware Encoder-Decoder from Noisy Labels by Alternating Back-Propagation for Saliency Detection
http://arxiv.org/abs/2007.12211
AUTHORS:        Jing Zhang ; Jianwen Xie ; Nick Barnes
COMMENTS:       ECCV2020
HIGHLIGHT:      In this paper, we propose a noise-aware encoder-decoder framework to disentangle a clean saliency predictor from noisy training examples, where the noisy labels are generated by unsupervised handcrafted feature-based methods.

70, TITLE:      Image-Based Benchmarking and Visualization for Large-Scale Global Optimization
http://arxiv.org/abs/2007.12332
AUTHORS:        Kyle Robert Harrison ; Azam Asilian Bidgoli ; Shahryar Rahnamayan ; Kalyanmoy Deb
COMMENTS:       Preprint submitted to Applied Intelligence. 43 pages, 30 figures
HIGHLIGHT:      In the context of optimization, visualization techniques can be useful for understanding the behaviour of optimization algorithms and can even provide a means to facilitate human interaction with an optimizer.

71, TITLE:      A Survey on Graph Neural Networks for Knowledge Graph Completion
http://arxiv.org/abs/2007.12374
AUTHORS:        Siddhant Arora
COMMENTS:       10 pages, 2 figures
HIGHLIGHT:      In this survey, we understand the various strengths and weaknesses of the proposed methodology and try to find new exciting research problems in this area that require further investigation.

72, TITLE:      Self-Supervised Monocular 3D Face Reconstruction by Occlusion-Aware Multi-view Geometry Consistency
http://arxiv.org/abs/2007.12494
AUTHORS:        Jiaxiang Shang ; Tianwei Shen ; Shiwei Li ; Lei Zhou ; Mingmin Zhen ; Tian Fang ; Long Quan
COMMENTS:       Accepted to ECCV 2020, supplementary materials included
HIGHLIGHT:      Our method is accurate and robust, especially under large variations of expressions, poses, and illumination conditions.



==========Updates to Previous Papers==========
1, TITLE:       A Semantics-Assisted Video Captioning Model Trained with Scheduled Sampling
http://arxiv.org/abs/1909.00121
AUTHORS:        Haoran Chen ; Ke Lin ; Alexander Maye ; Jianming Li ; Xiaolin Hu
COMMENTS:       11 pages
HIGHLIGHT:      Toward resolving these three problems, we suggest three corresponding improvements.

2, TITLE:       A survey on Semi-, Self- and Unsupervised Learning for Image Classification
http://arxiv.org/abs/2002.08721
AUTHORS:        Lars Schmarje ; Monty Santarossa ; Simon-Martin Schröder ; Reinhard Koch
COMMENTS:       Under consideration at Computer Vision and Image Understanding
HIGHLIGHT:      In this survey, we provide an overview of often used ideas and methods in image classification with fewer labels.

3, TITLE:       Multi-Agent Informational Learning Processes
http://arxiv.org/abs/2006.06870
AUTHORS:        Justin K Terry ; Nathaniel Grammel
HIGHLIGHT:      We introduce a new mathematical model of multi-agent reinforcement learning, the Multi-Agent Informational Learning Processor "MAILP" model.

4, TITLE:       Encoding Knowledge Graph Entity Aliases in Attentive Neural Network for Wikidata Entity Linking
http://arxiv.org/abs/1912.06214
AUTHORS:        Isaiah Onando Mulang ; Kuldeep Singh ; Akhilesh Vyas ; Saeedeh Shekarpour ; Maria Esther Vidal ; Soren Auer ; Jens Lehmann
COMMENTS:       15 pages
HIGHLIGHT:      In this paper, we examine the role of knowledge graph context on an attentive neural network approach for entity linking on Wikidata.

5, TITLE:       Randomized Smoothing of All Shapes and Sizes
http://arxiv.org/abs/2002.08118
AUTHORS:        Greg Yang ; Tony Duan ; J. Edward Hu ; Hadi Salman ; Ilya Razenshteyn ; Jerry Li
COMMENTS:       9 pages main text, 49 pages total
HIGHLIGHT:      We propose a novel framework for devising and analyzing randomized smoothing schemes, and validate its effectiveness in practice.

6, TITLE:       HoughNet: Integrating near and long-range evidence for bottom-up object detection
http://arxiv.org/abs/2007.02355
AUTHORS:        Nermin Samet ; Samet Hicsonmez ; Emre Akbas
COMMENTS:       ECCV 2020 camera-ready version
HIGHLIGHT:      This paper presents HoughNet, a one-stage, anchor-free, voting-based, bottom-up object detection method.

7, TITLE:       Fictitious Play Outperforms Counterfactual Regret Minimization
http://arxiv.org/abs/2001.11165
AUTHORS:        Sam Ganzfried
COMMENTS:       Fixed a bug in the 5-player CFR implementation from prior version and reran the 5-player experiments
HIGHLIGHT:      We compare the performance of two popular algorithms, fictitious play and counterfactual regret minimization, in approximating Nash equilibrium in multiplayer games.

8, TITLE:       Declarative Mechanism Design
http://arxiv.org/abs/1912.13122
AUTHORS:        Andrés García-Camino
HIGHLIGHT:      The main purpose of this paper is to bring attention to Artificial Teaching (AT) and to give a tentative answer showing a proof-of-concept implementation of Regulated Deep Learning (RDL).

9, TITLE:       AP20-OLR Challenge: Three Tasks and Their Baselines
http://arxiv.org/abs/2006.03473
AUTHORS:        Zheng Li ; Miao Zhao ; Qingyang Hong ; Lin Li ; Zhiyuan Tang ; Dong Wang ; Liming Song ; Cheng Yang
COMMENTS:       arXiv admin note: substantial text overlap with arXiv:1907.07626, arXiv:1806.00616, arXiv:1706.09742
HIGHLIGHT:      This paper introduces the fifth oriental language recognition (OLR) challenge AP20-OLR, which intends to improve the performance of language recognition systems, along with APSIPA Annual Summit and Conference (APSIPA ASC).

10, TITLE:      URIE: Universal Image Enhancement for Visual Recognition in the Wild
http://arxiv.org/abs/2007.08979
AUTHORS:        Taeyoung Son ; Juwon Kang ; Namyup Kim ; Sunghyun Cho ; Suha Kwak
COMMENTS:       Accepted as a conference paper at ECCV 2020
HIGHLIGHT:      To tackle this issue, we present a Universal and Recognition-friendly Image Enhancement network, dubbed URIE, which is attached in front of existing recognition models and enhances distorted input to improve their performance without retraining them.

11, TITLE:      Physics-Incorporated Convolutional Recurrent Neural Networks for Source Identification and Forecasting of Dynamical Systems
http://arxiv.org/abs/2004.06243
AUTHORS:        Priyabrata Saha ; Saurabh Dash ; Saibal Mukhopadhyay
HIGHLIGHT:      In this paper, we present a hybrid framework combining physics-based numerical models with deep learning for source identification and forecasting of spatio-temporal dynamical systems with unobservable time-varying external sources.

12, TITLE:      The Automation of Acceleration: AI and the Future of Society
http://arxiv.org/abs/2007.04477
AUTHORS:        Nicholas Kluge Corrêa
HIGHLIGHT:      In this article, we present a review of several points, mainly the benefits and risks of social modernisation through AI, and how human society has been preparing to deal with such changes.

13, TITLE:      Continuous Adaptation for Interactive Object Segmentation by Learning from Corrections
http://arxiv.org/abs/1911.12709
AUTHORS:        Theodora Kontogianni ; Michael Gygli ; Jasper Uijlings ; Vittorio Ferrari
COMMENTS:       ECCV 2020 Camera Ready
HIGHLIGHT:      Instead, we recognize that user corrections can serve as sparse training examples and we propose a method that capitalizes on that idea to update the model parameters on-the-fly to the data at hand.

14, TITLE:      SF-Net: Single-Frame Supervision for Temporal Action Localization
http://arxiv.org/abs/2003.06845
AUTHORS:        Fan Ma ; Linchao Zhu ; Yi Yang ; Shengxin Zha ; Gourab Kundu ; Matt Feiszli ; Zheng Shou
COMMENTS:       ECCV 2020
HIGHLIGHT:      In this paper, we study an intermediate form of supervision, i.e., single-frame supervision, for temporal action localization (TAL).

15, TITLE:      Invariant Policy Optimization: Towards Stronger Generalization in Reinforcement Learning
http://arxiv.org/abs/2006.01096
AUTHORS:        Anoopkumar Sonar ; Vincent Pacelli ; Anirudha Majumdar
COMMENTS:       12 pages, 4 figures
HIGHLIGHT:      In this paper, we approach this challenge through the following invariance principle: an agent must find a representation such that there exists an action-predictor built on top of this representation that is simultaneously optimal across all training domains.

16, TITLE:      KaLM at SemEval-2020 Task 4: Knowledge-aware Language Models for Comprehension And Generation
http://arxiv.org/abs/2005.11768
AUTHORS:        Jiajing Wan ; Xinting Huang
COMMENTS:       6 pages, 1 figure
HIGHLIGHT:      This paper presents our strategies in SemEval 2020 Task 4: Commonsense Validation and Explanation.

17, TITLE:      Password-conditioned Anonymization and Deanonymization with Face Identity Transformers
http://arxiv.org/abs/1911.11759
AUTHORS:        Xiuye Gu ; Weixin Luo ; Michael S. Ryoo ; Yong Jae Lee
COMMENTS:       ECCV 2020
HIGHLIGHT:      We propose a novel face identity transformer which enables automated photo-realistic password-based anonymization as well as deanonymization of human faces appearing in visual data.

18, TITLE:      H3DNet: 3D Object Detection Using Hybrid Geometric Primitives
http://arxiv.org/abs/2006.05682
AUTHORS:        Zaiwei Zhang ; Bo Sun ; Haitao Yang ; Qixing Huang
HIGHLIGHT:      We introduce H3DNet, which takes a colorless 3D point cloud as input and outputs a collection of oriented object bounding boxes (or BB) and their semantic labels.

19, TITLE:      Neural Geometric Parser for Single Image Camera Calibration
http://arxiv.org/abs/2007.11855
AUTHORS:        Jinwoo Lee ; Minhyuk Sung ; Hyunjoon Lee ; Junho Kim
COMMENTS:       ECCV 2020
HIGHLIGHT:      We propose a neural geometric parser learning single image camera calibration for man-made scenes.

20, TITLE:      Inherent Adversarial Robustness of Deep Spiking Neural Networks: Effects of Discrete Input Encoding and Non-Linear Activations
http://arxiv.org/abs/2003.10399
AUTHORS:        Saima Sharmin ; Nitin Rathi ; Priyadarshini Panda ; Kaushik Roy
COMMENTS:       Accepted in 16th European Conference on Computer Vision (ECCV 2020)
HIGHLIGHT:      In this work, we demonstrate that adversarial accuracy of SNNs under gradient-based attacks is higher than their non-spiking counterparts for CIFAR datasets on deep VGG and ResNet architectures, particularly in blackbox attack scenario.

21, TITLE:      LEARN Codes: Inventing Low-latency Codes via Recurrent Neural Networks
http://arxiv.org/abs/1811.12707
AUTHORS:        Yihan Jiang ; Hyeji Kim ; Himanshu Asnani ; Sreeram Kannan ; Sewoong Oh ; Pramod Viswanath
HIGHLIGHT:      LEARN Codes: Inventing Low-latency Codes via Recurrent Neural Networks

22, TITLE:      Contact and Human Dynamics from Monocular Video
http://arxiv.org/abs/2007.11678
AUTHORS:        Davis Rempe ; Leonidas J. Guibas ; Aaron Hertzmann ; Bryan Russell ; Ruben Villegas ; Jimei Yang
COMMENTS:       ECCV 2020
HIGHLIGHT:      In this paper, we present a physics-based method for inferring 3D human motion from video sequences that takes initial 2D and 3D pose estimates as input.

23, TITLE:      Internal and external pressures on language emergence: least effort, object constancy and frequency
http://arxiv.org/abs/2004.03868
AUTHORS:        Diana Rodríguez Luna ; Edoardo Maria Ponti ; Dieuwke Hupkes ; Elia Bruni
HIGHLIGHT:      In this paper, we propose some realistic sources of pressure on communication that avert this outcome.

24, TITLE:      Investigating Object Compositionality in Generative Adversarial Networks
http://arxiv.org/abs/1810.10340
AUTHORS:        Sjoerd van Steenkiste ; Karol Kurach ; Jürgen Schmidhuber ; Sylvain Gelly
COMMENTS:       A preliminary version of this work (arXiv v1) appeared under the title "A Case for Object Compositionality in Deep Generative Models of Images" as a workshop paper at the NeurIPS2018 workshop on "Modeling the Physical World: Perception, Learning, and Control", and at the NeurIPS2018 workshop on "Relational Representation Learning"
HIGHLIGHT:      In this work, we investigate object compositionality as an inductive bias for Generative Adversarial Networks (GANs).

25, TITLE:      Language Models as Fact Checkers?
http://arxiv.org/abs/2006.04102
AUTHORS:        Nayeon Lee ; Belinda Z. Li ; Sinong Wang ; Wen-tau Yih ; Hao Ma ; Madian Khabsa
COMMENTS:       Accepted in FEVER Workshop (ACL2020)
HIGHLIGHT:      In this paper, we leverage this implicit knowledge to create an effective end-to-end fact checker using a solely a language model, without any external knowledge or explicit retrieval components.

26, TITLE:      Lifespan Age Transformation Synthesis
http://arxiv.org/abs/2003.09764
AUTHORS:        Roy Or-El ; Soumyadip Sengupta ; Ohad Fried ; Eli Shechtman ; Ira Kemelmacher-Shlizerman
COMMENTS:       ECCV 2020 Camera-Ready version. Main Changes: 1. Added Ethics & Bias statement in the supplementary material 2. Comparison figures to PyGAN [46] and S2GAN [13] were removed due to copyright issues. These figures can be found in the project's webpage (link is provided in the paper). 3. Added links to the code and dataset (Github)
HIGHLIGHT:      We propose a novel multi-domain image-to-image generative adversarial network architecture, whose learned latent space models a continuous bi-directional aging process.

27, TITLE:      Automatic Lyrics Transcription using Dilated Convolutional Neural Networks with Self-Attention
http://arxiv.org/abs/2007.06486
AUTHORS:        Emir Demirel ; Sven Ahlback ; Simon Dixon
HIGHLIGHT:      This paper proposes a complete pipeline for this task which may commonly be referred as automatic lyrics transcription (ALT).

28, TITLE:      Adversarially-learned Inference via an Ensemble of Discrete Undirected Graphical Models
http://arxiv.org/abs/2007.05033
AUTHORS:        Adarsh K. Jeewajee ; Leslie P. Kaelbling
COMMENTS:       11 pages, 4 figures, 2 tables. Submitted to NeurIPS 2020
HIGHLIGHT:      Instead, we propose an inference-agnostic adversarial training framework for producing an ensemble of graphical models (AGMs).

29, TITLE:      Intrinsically Motivated Goal Exploration Processes with Automatic Curriculum Learning
http://arxiv.org/abs/1708.02190
AUTHORS:        Sébastien Forestier ; Rémy Portelas ; Yoan Mollard ; Pierre-Yves Oudeyer
HIGHLIGHT:      We present an algorithmic approach called Intrinsically Motivated Goal Exploration Processes (IMGEP) to enable similar properties of autonomous or self-supervised learning in machines.

30, TITLE:      Occlusion-Aware Depth Estimation with Adaptive Normal Constraints
http://arxiv.org/abs/2004.00845
AUTHORS:        Xiaoxiao Long ; Lingjie Liu ; Christian Theobalt ; Wenping Wang
COMMENTS:       ECCV 2020
HIGHLIGHT:      We present a new learning-based method for multi-frame depth estimation from a color video, which is a fundamental problem in scene understanding, robot navigation or handheld 3D reconstruction.

31, TITLE:      Temporally Coherent Embeddings for Self-Supervised Video Representation Learning
http://arxiv.org/abs/2004.02753
AUTHORS:        Joshua Knights ; Ben Harwood ; Daniel Ward ; Anthony Vanderkop ; Olivia Mackenzie-Ross ; Peyman Moghadam
COMMENTS:       Under review! Project page: https://csiro-robotics.github.io/TCE_Webpage/
HIGHLIGHT:      This paper presents TCE: Temporally Coherent Embeddings for self-supervised video representation learning.

32, TITLE:      BATS: Binary ArchitecTure Search
http://arxiv.org/abs/2003.01711
AUTHORS:        Adrian Bulat ; Brais Martinez ; Georgios Tzimiropoulos
COMMENTS:       accepted to ECCV 2020
HIGHLIGHT:      This paper proposes Binary ArchitecTure Search (BATS), a framework that drastically reduces the accuracy gap between binary neural networks and their real-valued counterparts by means of Neural Architecture Search (NAS).

33, TITLE:      Spatio-Temporal Graph Transformer Networks for Pedestrian Trajectory Prediction
http://arxiv.org/abs/2005.08514
AUTHORS:        Cunjun Yu ; Xiao Ma ; Jiawei Ren ; Haiyu Zhao ; Shuai Yi
COMMENTS:       ECCV camera-ready
HIGHLIGHT:      In this paper, we present STAR, a Spatio-Temporal grAph tRansformer framework, which tackles trajectory prediction by only attention mechanisms.

34, TITLE:      Human Trajectory Forecasting in Crowds: A Deep Learning Perspective
http://arxiv.org/abs/2007.03639
AUTHORS:        Parth Kothari ; Sven Kreiss ; Alexandre Alahi
COMMENTS:       IEEE Notice added, Figures updated
HIGHLIGHT:      In this work, we cast the problem of human trajectory forecasting as learning a representation of human social interactions.

35, TITLE:      Parameter Sharing is Surprisingly Useful for Multi-Agent Deep Reinforcement Learning
http://arxiv.org/abs/2005.13625
AUTHORS:        Justin K Terry ; Nathaniel Grammel ; Ananth Hari ; Luis Santos
HIGHLIGHT:      We use the MAILP model to show that increasing training centralization arbitrarily mitigates the slowing of convergence due to nonstationarity.

36, TITLE:      WeightNet: Revisiting the Design Space of Weight Networks
http://arxiv.org/abs/2007.11823
AUTHORS:        Ningning Ma ; Xiangyu Zhang ; Jiawei Huang ; Jian Sun
COMMENTS:       ECCV 2020
HIGHLIGHT:      We present a conceptually simple, flexible and effective framework for weight generating networks.

37, TITLE:      Funnel Activation for Visual Recognition
http://arxiv.org/abs/2007.11824
AUTHORS:        Ningning Ma ; Xiangyu Zhang ; Jian Sun
COMMENTS:       ECCV 2020
HIGHLIGHT:      We present a conceptually simple but effective funnel activation for image recognition tasks, called Funnel activation (FReLU), that extends ReLU and PReLU to a 2D activation by adding a negligible overhead of spatial condition.

38, TITLE:      Connecting the Dots: Detecting Adversarial Perturbations Using Context Inconsistency
http://arxiv.org/abs/2007.09763
AUTHORS:        Shasha Li ; Shitong Zhu ; Sudipta Paul ; Amit Roy-Chowdhury ; Chengyu Song ; Srikanth Krishnamurthy ; Ananthram Swami ; Kevin S Chan
COMMENTS:       The paper is accepted by ECCV 2020
HIGHLIGHT:      Our approach builds a set of auto-encoders, one for each object class, appropriately trained so as to output a discrepancy between the input and output if an added adversarial perturbation violates context consistency rules.

39, TITLE:      Preparing Lessons: Improve Knowledge Distillation with Better Supervision
http://arxiv.org/abs/1911.07471
AUTHORS:        Tiancheng Wen ; Shenqi Lai ; Xueming Qian
HIGHLIGHT:      We introduce two novel approaches, Knowledge Adjustment (KA) and Dynamic Temperature Distillation (DTD), to penalize bad supervision and improve student model.

40, TITLE:      Efficient Image Gallery Representations at Scale Through Multi-Task Learning
http://arxiv.org/abs/2005.09027
AUTHORS:        Benjamin Gutelman ; Pavel Levin
COMMENTS:       Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval
HIGHLIGHT:      We study the problem of building a universal image gallery encoder through multi-task learning (MTL) approach and demonstrate that it is indeed a practical way to achieve generalizability of learned representations to new downstream tasks.

41, TITLE:      A global method to identify trees outside of closed-canopy forests with medium-resolution satellite imagery
http://arxiv.org/abs/2005.08702
AUTHORS:        John Brandt ; Fred Stolle
HIGHLIGHT:      Here we present a globally consistent method to identify trees with canopy diameters greater than three meters with medium-resolution optical and radar imagery.

42, TITLE:      Local Clustering with Mean Teacher for Semi-supervised Learning
http://arxiv.org/abs/2004.09665
AUTHORS:        Zexi Chen ; Benjamin Dutton ; Bharathkumar Ramachandra ; Tianfu Wu ; Ranga Raju Vatsavai
COMMENTS:       8 pages, 7 figures
HIGHLIGHT:      In this work, we propose a simple yet effective method called Local Clustering (LC) to mitigate the effect of confirmation bias.

43, TITLE:      Functional Asplund metrics for pattern matching, robust to variable lighting conditions
http://arxiv.org/abs/1909.01585
AUTHORS:        Guillaume Noyel ; Michel Jourlin
HIGHLIGHT:      In this paper, we propose a complete framework to process images captured under uncontrolled lighting and especially under low lighting.

44, TITLE:      Finding trainable sparse networks through Neural Tangent Transfer
http://arxiv.org/abs/2006.08228
AUTHORS:        Tianlin Liu ; Friedemann Zenke
COMMENTS:       Accepted by ICML 2020
HIGHLIGHT:      In this article, we introduce Neural Tangent Transfer, a method that instead finds trainable sparse networks in a label-free manner.

45, TITLE:      Exploiting No-Regret Algorithms in System Design
http://arxiv.org/abs/2007.11172
AUTHORS:        Le Cong Dinh ; Nick Bishop ; Long Tran-Thanh
HIGHLIGHT:      To design such a payoff matrix, we propose a novel solution that provably has a unique minimax solution with the desired behaviour.

46, TITLE:      Comprehensive SNN Compression Using ADMM Optimization and Activity Regularization
http://arxiv.org/abs/1911.00822
AUTHORS:        Lei Deng ; Yujie Wu ; Yifan Hu ; Ling Liang ; Guoqi Li ; Xing Hu ; Yufei Ding ; Peng Li ; Yuan Xie
COMMENTS:       Under review
HIGHLIGHT:      These methods can be applied in either a single way for moderate compression or a joint way for aggressive compression.

47, TITLE:      A Metric Learning Reality Check
http://arxiv.org/abs/2003.08505
AUTHORS:        Kevin Musgrave ; Serge Belongie ; Ser-Nam Lim
COMMENTS:       Visit https://www.github.com/KevinMusgrave/powerful-benchmarker for supplementary material, including the source code, configuration files, log files, and interactive bayesian optimization plots
HIGHLIGHT:      In this paper, we take a closer look at the field to see if this is actually true.

48, TITLE:      A unifying mutual information view of metric learning: cross-entropy vs. pairwise losses
http://arxiv.org/abs/2003.08983
AUTHORS:        Malik Boudiaf ; Jérôme Rony ; Imtiaz Masud Ziko ; Eric Granger ; Marco Pedersoli ; Pablo Piantanida ; Ismail Ben Ayed
COMMENTS:       ECCV 2020 (Spotlight) - Code available at: https://github.com/jeromerony/dml_cross_entropy
HIGHLIGHT:      Recently, substantial research efforts in Deep Metric Learning (DML) focused on designing complex pairwise-distance losses, which require convoluted schemes to ease optimization, such as sample mining or pair weighting.