2020.03.23.txt

==========New Papers==========
1, TITLE:       TNT-KID: Transformer-based Neural Tagger for Keyword Identification
http://arxiv.org/abs/2003.09166
AUTHORS:        Matej Martinc ; Blaž Škrlj ; Senja Pollak
COMMENTS:       Submitted to Natural Language Engineering journal
HIGHLIGHT:      In this research we present a novel algorithm for keyword identification, i.e., an extraction of one or multi-word phrases representing key aspects of a given document, called Transformer-based Neural Tagger for Keyword IDentification (TNT-KID).

2, TITLE:       Techniques for Vocabulary Expansion in Hybrid Speech Recognition Systems
http://arxiv.org/abs/2003.09024
AUTHORS:        Nikolay Malkovsky ; Vladimir Bataev ; Dmitrii Sviridkin ; Natalia Kizhaeva ; Aleksandr Laptev ; Ildar Valiev ; Oleg Petrov
COMMENTS:       Submitted to Interspeech 2020
HIGHLIGHT:      In this paper we explore different existing methods of this solution on both graph construction and search method levels.

3, TITLE:       Parallel Intent and Slot Prediction using MLB Fusion
http://arxiv.org/abs/2003.09211
AUTHORS:        Anmol Bhasin ; Bharatram Natarajan ; Gaurav Mathur ; Himanshu Mangla
HIGHLIGHT:      In this work, we propose a parallel Intent and Slot Prediction technique where separate Bidirectional Gated Recurrent Units (GRU) are used for each task.

4, TITLE:       Language Technology Programme for Icelandic 2019-2023
http://arxiv.org/abs/2003.09244
AUTHORS:        Anna Björk Nikulásdóttir ; Jón Guðnason ; Anton Karl Ingason ; Hrafn Loftsson ; Eiríkur Rögnvaldsson ; Einar Freyr Sigurðsson ; Steinþór Steingrímsson
COMMENTS:       Accepted at LREC 2020
HIGHLIGHT:      In this paper, we describe a new national language technology programme for Icelandic.

5, TITLE:       NSURL-2019 Task 7: Named Entity Recognition (NER) in Farsi
http://arxiv.org/abs/2003.09029
AUTHORS:        Nasrin Taghizadeh ; Zeinab Borhanifard ; Melika GolestaniPour ; Heshaam Faili
HIGHLIGHT:      This paper describes the process of making training and test data, a list of participating teams (6 teams), and evaluation results of their systems.

6, TITLE:       Ordered Functional Decision Diagrams
http://arxiv.org/abs/2003.09340
AUTHORS:        Joan Thibault ; Khalil Ghorbal
HIGHLIGHT:      Several BDD variants were designed to exploit special features of Boolean functions to achieve better compression rates.Deciding a priori which variant to use is as hard as constructing the diagrams themselves and the conversion between variants comes in general with a prohibitive cost.This observation leads naturally to a growing interest into when and how one can combine existing variants to benefit from their respective sweet spots.In this paper, we introduce a novel framework, termed \lambdaDD (LDD), that revisits BDD from a purely functional point of view.The framework allows to classify the already existing variants, including the most recent ones like ChainDD and ESRBDD, as implementations of a special class of ordered models.We enumerate, in a principled way, all the models of this class and isolate its most expressive model.This new model, termed \lambdaDD-O-NUCX, is suitable for both dense and sparse Boolean functions, and, unlike ChainDD and ESRBDD, is invariant by negation.The canonicity of \lambdaDD-O-NUCX is formally verified using the Coq proof assistant.We furthermore provide experimental evidence corroborating our theoretical findings: more expressive \lambdaDD models achieve, indeed, better memory compression rates.

7, TITLE:       Comments on Sejnowski's "The unreasonable effectiveness of deep learning in artificial intelligence" [arXiv:2002.04806]
http://arxiv.org/abs/2003.09415
AUTHORS:        Leslie S. Smith
COMMENTS:       5 pages, 2 figures
HIGHLIGHT:      There are detailed mathematical analyses, but this short paper attempts to look at the issue differently, considering the way that these networks are used, the subset of these functions that can be achieved by training (starting from some location in the original function space), as well as the functions that in reality will be modelled.

8, TITLE:       RGB-Topography and X-rays Image Registration for Idiopathic Scoliosis Children Patient Follow-up
http://arxiv.org/abs/2003.09404
AUTHORS:        Insaf Setitra ; Noureddine Aouaa ; Abdelkrim Meziane ; Afef Benrabia ; Houria Kaced ; Hanene Belabassi ; Sara Ait Ziane ; Nadia Henda Zenati ; Oualid Djekkoune
HIGHLIGHT:      In this work, we exploit both RGB images of scoliosis captured during clinical diagnosis, and X-rays hard copies provided by patients in order to register both images and give a rich comparison of diagnoses.

9, TITLE:       Across-scale Process Similarity based Interpolation for Image Super-Resolution
http://arxiv.org/abs/2003.09182
AUTHORS:        Sobhan Kanti Dhara ; Debashis Sen
HIGHLIGHT:      In this paper, we propose a technique that performs interpolation through an infusion of high frequency signal components computed by exploiting `process similarity'.

10, TITLE:      U-Det: A Modified U-Net architecture with bidirectional feature network for lung nodule segmentation
http://arxiv.org/abs/2003.09293
AUTHORS:        Nikhil Varma Keetha ; Samson Anosh Babu P ; Chandra Sekhara Rao Annavarapu
COMMENTS:       14 pages, 7 figures, 5 tables
HIGHLIGHT:      This article proposes U-Det, a resource-efficient model architecture, which is an end to end deep learning approach to solve the task at hand.

11, TITLE:      DIDFuse: Deep Image Decomposition for Infrared and Visible Image Fusion
http://arxiv.org/abs/2003.09210
AUTHORS:        Zixiang Zhao ; Shuang Xu ; Chunxia Zhang ; Junmin Liu ; Pengfei Li ; Jiangshe Zhang
HIGHLIGHT:      This paper proposes a novel auto-encoder (AE) based fusion network.

12, TITLE:      Diagnosis of Diabetic Retinopathy in Ethiopia: Before the Deep Learning based Automation
http://arxiv.org/abs/2003.09208
AUTHORS:        Misgina Tsighe Hagos
COMMENTS:       Accepted for poster presentation at the Practical Machine Learning for Developing Countries (PML4DC) workshop, ICLR 2020
HIGHLIGHT:      Diagnosis of Diabetic Retinopathy in Ethiopia: Before the Deep Learning based Automation

13, TITLE:      Masked Face Recognition Dataset and Application
http://arxiv.org/abs/2003.09093
AUTHORS:        Zhongyuan Wang ; Guangcheng Wang ; Baojin Huang ; Zhangyang Xiong ; Qi Hong ; Hao Wu ; Peng Yi ; Kui Jiang ; Nanxi Wang ; Yingjiao Pei ; Heling Chen ; Yu Miao ; Zhibing Huang ; Jinbi Liang
HIGHLIGHT:      To this end, this work proposes three types of masked face datasets, including Masked Face Detection Dataset (MFDD), Real-world Masked Face Recognition Dataset (RMFRD) and Simulated Masked Face Recognition Dataset (SMFRD).

14, TITLE:      CentripetalNet: Pursuing High-quality Keypoint Pairs for Object Detection
http://arxiv.org/abs/2003.09119
AUTHORS:        Zhiwei Dong ; Guoxuan Li ; Yue Liao ; Fei Wang ; Pengju Ren ; Chen Qian
COMMENTS:       Accepted by CVPR2020
HIGHLIGHT:      In this paper, we propose CentripetalNet which uses centripetal shift to pair corner keypoints from the same instance.

15, TITLE:      Data-Free Knowledge Amalgamation via Group-Stack Dual-GAN
http://arxiv.org/abs/2003.09088
AUTHORS:        Jingwen Ye ; Yixin Ji ; Xinchao Wang ; Xin Gao ; Mingli Song
COMMENTS:       Accepted by CVPR 2020
HIGHLIGHT:      In this paper, we propose a data-free knowledge amalgamate strategy to craft a well-behaved multi-task student network from multiple single/multi-task teachers.

16, TITLE:      FocalMix: Semi-Supervised Learning for 3D Medical Image Detection
http://arxiv.org/abs/2003.09108
AUTHORS:        Dong Wang ; Yuan Zhang ; Kexin Zhang ; Liwei Wang
COMMENTS:       Accepted by CVPR 2020
HIGHLIGHT:      In this paper, we propose a novel method, called FocalMix, which, to the best of our knowledge, is the first to leverage recent advances in semi-supervised learning (SSL) for 3D medical image detection.

17, TITLE:      Dual-discriminator GAN: A GAN way of profile face recognition
http://arxiv.org/abs/2003.09116
AUTHORS:        Xinyu Zhang ; Yang Zhao ; Hao Zhang
HIGHLIGHT:      In this paper, we proposed a method of generating frontal faces with image-to-image profile faces based on Generative Adversarial Network (GAN).

18, TITLE:      Automatic Identification of Types of Alterations in Historical Manuscripts
http://arxiv.org/abs/2003.09136
AUTHORS:        David Lassner ; Anne Baillot ; Sergej Dogadov ; Klaus-Robert Müller ; Shinichi Nakajima
HIGHLIGHT:      In this paper, we present a machine learning-based approach to help categorize alterations in documents.

19, TITLE:      Online Continual Learning on Sequences
http://arxiv.org/abs/2003.09114
AUTHORS:        German I. Parisi ; Vincenzo Lomonaco
COMMENTS:       L. Oneto et al. (eds.), Recent Trends in Learning From Data, Studies in Computational Intelligence 896
HIGHLIGHT:      In this chapter, we summarize and discuss recent deep learning models that address OCL on sequential input through the use (and combination) of synaptic regularization, structural plasticity, and experience replay.

20, TITLE:      Few-Shot Learning with Geometric Constraints
http://arxiv.org/abs/2003.09151
AUTHORS:        Hong-Gyu Jung ; Seong-Whan Lee
COMMENTS:       Accepted for publication in IEEE Transactions on Neural Networks and Learning Systems (T-NNLS)
HIGHLIGHT:      In this article, we consider the problem of few-shot learning for classification.

21, TITLE:      Unsupervised Latent Space Translation Network
http://arxiv.org/abs/2003.09149
AUTHORS:        Magda Friedjungová ; Daniel Vašata ; Tomáš Chobola ; Marcel Jiřina
COMMENTS:       To be published in conference proceedings of ESANN 2020
HIGHLIGHT:      In this paper, we present an enhancement of the UNIT framework that aids in removing its main drawbacks.

22, TITLE:      Exploring Categorical Regularization for Domain Adaptive Object Detection
http://arxiv.org/abs/2003.09152
AUTHORS:        Chang-Dong Xu ; Xing-Ran Zhao ; Xin Jin ; Xiu-Shen Wei
COMMENTS:       To appear in CVPR 2020. X.-S. Wei is the corresponding author
HIGHLIGHT:      In this paper, we tackle the domain adaptive object detection problem, where the main challenge lies in significant domain gaps between source and target domains.

23, TITLE:      Event-based Asynchronous Sparse Convolutional Networks
http://arxiv.org/abs/2003.09148
AUTHORS:        Nico Messikommer ; Daniel Gehrig ; Antonio Loquercio ; Davide Scaramuzza
HIGHLIGHT:      In this work, we present a general framework for converting models trained on synchronous image-like event representations into asynchronous models with identical output, thus directly leveraging the intrinsic asynchronous and sparse nature of the event data.

24, TITLE:      Three-branch and Mutil-scale learning for Fine-grained Image Recognition (TBMSL-Net)
http://arxiv.org/abs/2003.09150
AUTHORS:        Fan Zhang ; Guisheng Zhai ; Meng Li ; Yizhao Liu
HIGHLIGHT:      Our approach is end-to-end training, through the comprehensive experiments demonstrate that our approach achieves state-of-the-art results on CUB-200-2011, Stanford Cars and FGVC-Aircraft datasets.

25, TITLE:      Detection in Crowded Scenes: One Proposal, Multiple Predictions
http://arxiv.org/abs/2003.09163
AUTHORS:        Xuangeng Chu ; Anlin Zheng ; Xiangyu Zhang ; Jian Sun
COMMENTS:       12 pages; 5 figures; 10 tables
HIGHLIGHT:      We propose a simple yet effective proposal-based object detector, aiming at detecting highly-overlapped instances in crowded scenes.

26, TITLE:      Deep Reinforcement Learning with Weighted Q-Learning
http://arxiv.org/abs/2003.09280
AUTHORS:        Andrea Cini ; Carlo D'Eramo ; Jan Peters ; Cesare Alippi
HIGHLIGHT:      In this work, we provide the methodological advances to benefit from the WQL properties in Deep Reinforcement Learning (DRL), by using neural networks with Dropout Variational Inference as an effective approximation of deep Gaussian processes.

27, TITLE:      Computational Complexity of the $α$-Ham-Sandwich Problem
http://arxiv.org/abs/2003.09266
AUTHORS:        Man-Kwun Chiu ; Aruni Choudhary ; Wolfgang Mulzer
HIGHLIGHT:      They gave an algorithm to find the hyperplane in time $O(n (\log n)^{d-3})$, where $n$ is the total number of input points.

28, TITLE:      Superaccurate Camera Calibration via Inverse Rendering
http://arxiv.org/abs/2003.09177
AUTHORS:        Morten Hannemose ; Jakob Wilm ; Jeppe Revall Frisvad
COMMENTS:       10 pages, 6 figures
HIGHLIGHT:      Instead of relying solely on detected feature points, we use an estimate of the internal parameters and the pose of the calibration object to implicitly render a non-photorealistic equivalent of the optical features.

29, TITLE:      DMV: Visual Object Tracking via Part-level Dense Memory and Voting-based Retrieval
http://arxiv.org/abs/2003.09171
AUTHORS:        Gunhee Nam ; Seoung Wug Oh ; Joon-Young Lee ; Seon Joo Kim
COMMENTS:       19 pages, 9 figures
HIGHLIGHT:      In this paper, we relieve this limitation by maintaining an external memory that saves the tracking record.

30, TITLE:      Privileged Pooling: Supervised attention-based pooling for compensating dataset bias
http://arxiv.org/abs/2003.09168
AUTHORS:        Andres C. Rodriguez ; Stefano D'Aronco ; Jan Dirk Wegner ; Konrad Schindler
COMMENTS:       privileged pooling, weakly-supervised attention, training set bias, fine-grained species recognition, camera trap imagery
HIGHLIGHT:      In this paper we propose a novel supervised image classification method that overcomes dataset bias and scarcity of training data using privileged information in the form of keypoints annotations.

31, TITLE:      3dDepthNet: Point Cloud Guided Depth Completion Network for Sparse Depth and Single Color Image
http://arxiv.org/abs/2003.09175
AUTHORS:        Rui Xiang ; Feng Zheng ; Huapeng Su ; Zhe Zhang
COMMENTS:       8 pages, 10 figures and 4 tables
HIGHLIGHT:      In this paper, we propose an end-to-end deep learning network named 3dDepthNet, which produces an accurate dense depth image from a single pair of sparse LiDAR depth and color image for robotics and autonomous driving tasks.

32, TITLE:      VisuoSpatial Foresight for Multi-Step, Multi-Task Fabric Manipulation
http://arxiv.org/abs/2003.09044
AUTHORS:        Ryan Hoque ; Daniel Seita ; Ashwin Balakrishna ; Aditya Ganapathi ; Ajay Kumar Tanwani ; Nawid Jamali ; Katsu Yamane ; Soshi Iba ; Ken Goldberg
HIGHLIGHT:      We introduce VisuoSpatial Foresight (VSF), which extends prior work by learning visual dynamics on domain randomized RGB images and depth maps simultaneously and completely in simulation.

33, TITLE:      Fully Automated Hand Hygiene Monitoring\\in Operating Room using 3D Convolutional Neural Network
http://arxiv.org/abs/2003.09087
AUTHORS:        Minjee Kim ; Joonmyeong Choi ; Namkug Kim
HIGHLIGHT:      Leveraging this progress, we proposed a fully automated hand hygiene monitoring tool of the alcohol-based hand rubbing action of anesthesiologists on OR video using spatio-temporal features with 3D CNN.

34, 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 25 pages and submitted to MDPI remote sensing journal (under review)
HIGHLIGHT:      We propose an architecture with three components: ESRGAN, Edge Enhancement Network (EEN), and Detection network.

35, TITLE:      A Graduated Filter Method for Large Scale Robust Estimation
http://arxiv.org/abs/2003.09080
AUTHORS:        Huu Le ; Christopher Zach
HIGHLIGHT:      In this paper, we introduce a novel solver for robust estimation that possesses a strong ability to escape poor local minima.

36, TITLE:      How to Train Your Event Camera Neural Network
http://arxiv.org/abs/2003.09078
AUTHORS:        Timo Stoffregen ; Cedric Scheerlinck ; Davide Scaramuzza ; Tom Drummond ; Nick Barnes ; Lindsay Kleeman ; Robert Mahony
HIGHLIGHT:      To address this, we present a new High Quality Frames (HQF) dataset, containing events and groundtruth frames from a DAVIS240C that are well-exposed and minimally motion-blurred.

37, TITLE:      Exchangeable Input Representations for Reinforcement Learning
http://arxiv.org/abs/2003.09022
AUTHORS:        John Mern ; Dorsa Sadigh ; Mykel J. Kochenderfer
COMMENTS:       6 pages, 7 figures
HIGHLIGHT:      This work presents an attention-based method to project neural network inputs into an efficient representation space that is invariant under changes to input ordering.

38, TITLE:      MOT20: A benchmark for multi object tracking in crowded scenes
http://arxiv.org/abs/2003.09003
AUTHORS:        Patrick Dendorfer ; Hamid Rezatofighi ; Anton Milan ; Javen Shi ; Daniel Cremers ; Ian Reid ; Stefan Roth ; Konrad Schindler ; Laura Leal-Taixé
COMMENTS:       The sequences of the new MOT20 benchmark were previously presented in the CVPR 2019 tracking challenge ( arXiv:1906.04567 ). The differences between the two challenges are: - New and corrected annotations - New sequences, as we had to crop and transform some old sequences to achieve higher quality in the annotations. - New baselines evaluations and different sets of public detections
HIGHLIGHT:      In this paper, we present our MOT20benchmark, consisting of 8 new sequences depicting very crowded challenging scenes.

39, TITLE:      Semi-Supervised Semantic Segmentation with Cross-Consistency Training
http://arxiv.org/abs/2003.09005
AUTHORS:        Yassine Ouali ; Céline Hudelot ; Myriam Tami
COMMENTS:       18 Pages, 10 Figures. To appear in the Proceedings of 2020 Conference on Computer Vision and Pattern Recognition, 13-19 June 2020, Seattle, Washington
HIGHLIGHT:      In this paper, we present a novel cross-consistency based semi-supervised approach for semantic segmentation.

40, TITLE:      Local Implicit Grid Representations for 3D Scenes
http://arxiv.org/abs/2003.08981
AUTHORS:        Chiyu Max Jiang ; Avneesh Sud ; Ameesh Makadia ; Jingwei Huang ; Matthias Nießner ; Thomas Funkhouser
COMMENTS:       CVPR 2020. Supplementary Video: https://youtu.be/XCyl1-vxfII
HIGHLIGHT:      In this paper, we introduce Local Implicit Grid Representations, a new 3D shape representation designed for scalability and generality.

41, TITLE:      Multilayer Dense Connections for Hierarchical Concept Classification
http://arxiv.org/abs/2003.09015
AUTHORS:        Toufiq Parag ; Hongcheng Wang
HIGHLIGHT:      We propose a multilayer dense connectivity for a CNN to simultaneously predict the category and its conceptual superclasses in hierarchical order.

42, TITLE:      Temporal Extension Module for Skeleton-Based Action Recognition
http://arxiv.org/abs/2003.08951
AUTHORS:        Yuya Obinata ; Takuma Yamamoto
COMMENTS:       7 pages, 4 figures, pre-print
HIGHLIGHT:      In this work, we focus on adding extra edges to neighboring multiple vertices on the inter-frame and extracting additional features based on the extended temporal graph.

43, TITLE:      Affinity Graph Supervision for Visual Recognition
http://arxiv.org/abs/2003.09049
AUTHORS:        Chu Wang ; Babak Samari ; Vladimir G. Kim ; Siddhartha Chaudhuri ; Kaleem Siddiqi
HIGHLIGHT:      Here we propose a principled method to directly supervise the learning of weights in affinity graphs, to exploit meaningful connections between entities in the data source.

44, TITLE:      Cross-Shape Graph Convolutional Networks
http://arxiv.org/abs/2003.09053
AUTHORS:        Dmitry Petrov ; Evangelos Kalogerakis
HIGHLIGHT:      We present a method that processes 3D point clouds by performing graph convolution operations across shapes.

45, TITLE:      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:       25 pages
HIGHLIGHT:      Recently, substantial research efforts in Deep Metric Learning (DML) focused on designing complex pairwise-distance losses and convoluted sample-mining and implementation strategies to ease optimization.

46, TITLE:      The value of text for small business default prediction: A deep learning approach
http://arxiv.org/abs/2003.08964
AUTHORS:        Matthew Stevenson ; Christophe Mues ; Cristián Bravo
COMMENTS:       Submitted - v1
HIGHLIGHT:      We consider the performance in terms of AUC (Area Under the Curve) and Balanced Accuracy and find that the text alone is surprisingly effective for predicting default.

47, TITLE:      Generating new concepts with hybrid neuro-symbolic models
http://arxiv.org/abs/2003.08978
AUTHORS:        Reuben Feinman ; Brenden M. Lake
HIGHLIGHT:      In this paper, we explore a synthesis of these two traditions through a novel neuro-symbolic model for generating new concepts.

48, TITLE:      Detecting Mismatch between Text Script and Voice-over Using Utterance Verification Based on Phoneme Recognition Ranking
http://arxiv.org/abs/2003.09180
AUTHORS:        Yoonjae Jeong ; Hoon-Young Cho
COMMENTS:       Accepted by ICASSP 2020
HIGHLIGHT:      The purpose of this study is to detect the mismatch between text script and voice-over.

49, TITLE:      Layerwise Knowledge Extraction from Deep Convolutional Networks
http://arxiv.org/abs/2003.09000
AUTHORS:        Simon Odense ; Artur d'Avila Garcez
HIGHLIGHT:      In this paper, we propose a novel layerwise knowledge extraction method using M-of-N rules which seeks to obtain the best trade-off between the complexity and accuracy of rules describing the hidden features of a deep network.

50, TITLE:      FedNER: Medical Named Entity Recognition with Federated Learning
http://arxiv.org/abs/2003.09288
AUTHORS:        Suyu Ge ; Fangzhao Wu ; Chuhan Wu ; Tao Qi ; Yongfeng Huang ; Xing Xie
HIGHLIGHT:      In this paper, we propose a privacy-preserving medical NER method based on federated learning, which can leverage the labeled data in different platforms to boost the training of medical NER model and remove the need of exchanging raw data among different platforms.

51, TITLE:      Tactic Learning and Proving for the Coq Proof Assistant
http://arxiv.org/abs/2003.09140
AUTHORS:        Lasse Blaauwbroek ; Josef Urban ; Herman Geuvers
COMMENTS:       12 pages, 2 figures, 1 table. For the associated artefacts, see https://doi.org/10.5281/zenodo.3693760
HIGHLIGHT:      We present a system that utilizes machine learning for tactic proof search in the Coq Proof Assistant.

52, TITLE:      Detection of Information Hiding at Anti-Copying 2D Barcodes
http://arxiv.org/abs/2003.09316
AUTHORS:        Ning Xie ; Ji Hu ; Junjie Chen ; Qiqi Zhang ; Changsheng Chen
HIGHLIGHT:      In this paper, we propose two hidden information detection schemes at the existing anti-copying 2D barcodes.

53, TITLE:      Visual Navigation Among Humans with Optimal Control as a Supervisor
http://arxiv.org/abs/2003.09354
AUTHORS:        Varun Tolani ; Somil Bansal ; Aleksandra Faust ; Claire Tomlin
COMMENTS:       Project Website: https://smlbansal.github.io/LB-WayPtNav-DH/
HIGHLIGHT:      We propose a novel framework for navigation around humans which combines learning-based perception with model-based optimal control.

54, TITLE:      TF-Coder: Program Synthesis for Tensor Manipulations
http://arxiv.org/abs/2003.09040
AUTHORS:        Kensen Shi ; David Bieber ; Rishabh Singh
HIGHLIGHT:      In this work, we present a tool called TF-Coder for programming by example in TensorFlow.

55, TITLE:      Evolutionary Multi-Objective Optimization Framework for Mining Association Rules
http://arxiv.org/abs/2003.09158
AUTHORS:        Shaik Tanveer Ul Huq ; Vadlamani Ravi
COMMENTS:       37 pages
HIGHLIGHT:      In this paper, two multi-objective optimization frameworks in two variants (i.e., NSGA-III-ARM-V1, NSGA-III-ARM-V2; and MOEAD-ARM-V1, MOEAD-ARM-V2) are proposed to find association rules from transactional datasets.

56, TITLE:      Microvasculature Segmentation and Inter-capillary Area Quantification of the Deep Vascular Complex using Transfer Learning
http://arxiv.org/abs/2003.09033
AUTHORS:        Julian Lo ; Morgan Heisler ; Vinicius Vanzan ; Sonja Karst ; Ivana Zadro Matovinovic ; Sven Loncaric ; Eduardo V. Navajas ; Mirza Faisal Beg ; Marinko V. Sarunic
COMMENTS:       27 pages, 8 figures
HIGHLIGHT:      We demonstrate accurate segmentation of the vascular morphology for the superficial capillary plexus and deep vascular complex (SCP and DVC) using a convolutional neural network (CNN) for quantitative analysis.

57, TITLE:      Hierarchical Severity Staging of Anterior Cruciate Ligament Injuries using Deep Learning
http://arxiv.org/abs/2003.09089
AUTHORS:        Nikan K. Namiri ; Io Flament ; Bruno Astuto ; Rutwik Shah ; Radhika Tibrewala ; Francesco Caliva ; Thomas M. Link ; Valentina Pedoia ; Sharmila Majumdar
HIGHLIGHT:      Hierarchical Severity Staging of Anterior Cruciate Ligament Injuries using Deep Learning

58, TITLE:      Kidney segmentation using 3D U-Net localized with Expectation Maximization
http://arxiv.org/abs/2003.09075
AUTHORS:        Omid Bazgir ; Kai Barck ; Richard A. D. Carano ; Robby M. Weimer ; Luke Xie
HIGHLIGHT:      In this paper we propose a new framework to address some of the challenges for segmenting 3D MRI.

59, TITLE:      Learning the Loss Functions in a Discriminative Space for Video Restoration
http://arxiv.org/abs/2003.09124
AUTHORS:        Younghyun Jo ; Jaeyeon Kang ; Seoung Wug Oh ; Seonghyeon Nam ; Peter Vajda ; Seon Joo Kim
COMMENTS:       24 pages
HIGHLIGHT:      To this end, we propose a new framework for building effective loss functions by learning a discriminative space specific to a video restoration task.

60, TITLE:      Weakly Supervised Context Encoder using DICOM metadata in Ultrasound Imaging
http://arxiv.org/abs/2003.09070
AUTHORS:        Szu-Yeu Hu ; Shuhang Wang ; Wei-Hung Weng ; JingChao Wang ; XiaoHong Wang ; Arinc Ozturk ; Qian Li ; Viksit Kumar ; Anthony E. Samir
COMMENTS:       Accept as a workshop paper at AI4AH, ICLR 2020
HIGHLIGHT:      In this work, we leverage DICOM metadata from ultrasound images to help learn representations of the ultrasound image.

61, TITLE:      Efficient algorithm for calculating transposed PSF matrices for 3D light field deconvolution
http://arxiv.org/abs/2003.09133
AUTHORS:        Martin Eberhart
COMMENTS:       7 pages, 11 figures, 1 table
HIGHLIGHT:      This paper illustrates the significance and the construction of this special matrix and presents an efficient algorithm for its computation, which is based on the distinct relation of the corresponding indices within the original and the transposed matrix.

62, TITLE:      Comprehensive Instructional Video Analysis: The COIN Dataset and Performance Evaluation
http://arxiv.org/abs/2003.09392
AUTHORS:        Yansong Tang ; Jiwen Lu ; Jie Zhou
COMMENTS:       Accepted by T-PAMI, journal version of COIN dataset arXiv:1903.02874
HIGHLIGHT:      Accordingly, we propose two simple yet effective methods, which can be easily plugged into conventional proposal-based action detection models. To address this, we present a large-scale dataset named as "COIN" for COmprehensive INstructional video analysis.

63, TITLE:      SER-FIQ: Unsupervised Estimation of Face Image Quality Based on Stochastic Embedding Robustness
http://arxiv.org/abs/2003.09373
AUTHORS:        Philipp Terhörst ; Jan Niklas Kolf ; Naser Damer ; Florian Kirchbuchner ; Arjan Kuijper
COMMENTS:       Accepted at CVPR2020
HIGHLIGHT:      Avoiding the use of inaccurate quality labels, we proposed a novel concept to measure face quality based on an arbitrary face recognition model.

64, TITLE:      Domain Adaptation by Class Centroid Matching and Local Manifold Self-Learning
http://arxiv.org/abs/2003.09391
AUTHORS:        Lei Tian ; Yongqiang Tang ; Liangchen Hu ; Zhida Ren ; Wensheng Zhang
HIGHLIGHT:      Different from existing methods that make label prediction for target samples independently, in this paper, we propose a novel domain adaptation approach that assigns pseudo-labels to target data with the guidance of class centroids in two domains, so that the data distribution structure of both source and target domains can be emphasized.

65, TITLE:      Weakly Supervised 3D Hand Pose Estimation via Biomechanical Constraints
http://arxiv.org/abs/2003.09282
AUTHORS:        Adrian Spurr ; Umar Iqbal ; Pavlo Molchanov ; Otmar Hilliges ; Jan Kautz
HIGHLIGHT:      Embracing this challenge we propose a set of novel losses.

66, TITLE:      Explainable Object-induced Action Decision for Autonomous Vehicles
http://arxiv.org/abs/2003.09405
AUTHORS:        Yiran Xu ; Xiaoyin Yang ; Lihang Gong ; Hsuan-Chu Lin ; Tz-Ying Wu ; Yunsheng Li ; Nuno Vasconcelos
HIGHLIGHT:      An extension of the BDD100K dataset, annotated for a set of 4 actions and 21 explanations, is proposed.

67, TITLE:      Selecting Relevant Features from a Universal Representation for Few-shot Classification
http://arxiv.org/abs/2003.09338
AUTHORS:        Nikita Dvornik ; Cordelia Schmid ; Julien Mairal
HIGHLIGHT:      In this work, we propose a new strategy based on feature selection, which is both simpler and more effective than previous feature adaptation approaches.

68, TITLE:      Out-of-Distribution Detection for Skin Lesion Images with Deep Isolation Forest
http://arxiv.org/abs/2003.09365
AUTHORS:        Xuan Li ; Yuchen Lu ; Christian Desrosiers ; Xue Liu
HIGHLIGHT:      In this paper, we study the problem of out-of-distribution detection in skin disease images.

69, TITLE:      Exploring Inherent Properties of the Monophonic Melody of Songs
http://arxiv.org/abs/2003.09287
AUTHORS:        Zehao Wang ; Shicheng Zhang ; Xiaoou Chen
HIGHLIGHT:      To boost the performance of deep-learning-related musical tasks, we propose a set of interpretable features on monophonic melody for computational purposes.



==========Updates to Previous Papers==========
1, TITLE:       AI2D-RST: A multimodal corpus of 1000 primary school science diagrams
http://arxiv.org/abs/1912.03879
AUTHORS:        Tuomo Hiippala ; Malihe Alikhani ; Jonas Haverinen ; Timo Kalliokoski ; Evanfiya Logacheva ; Serafina Orekhova ; Aino Tuomainen ; Matthew Stone ; John A. Bateman
COMMENTS:       24 pages; revised version submitted to Language Resources & Evaluation
HIGHLIGHT:      This article introduces AI2D-RST, a multimodal corpus of 1000 English-language diagrams that represent topics in primary school natural sciences, such as food webs, life cycles, moon phases and human physiology.

2, TITLE:       An Approach for Process Model Extraction By Multi-Grained Text Classification
http://arxiv.org/abs/1906.02127
AUTHORS:        Chen Qian ; Lijie Wen ; Akhil Kumar ; Leilei Lin ; Li Lin ; Zan Zong ; Shuang Li ; Jianmin Wang
COMMENTS:       Accepted to CAiSE-2020
HIGHLIGHT:      In this paper, we formalize the PME task into the multi-grained text classification problem, and propose a hierarchical neural network to effectively model and extract multi-grained information without manually-defined procedural features. To evaluate our approach, we construct two multi-grained datasets from two different domains and conduct extensive experiments from different dimensions.

3, TITLE:       End-to-end ASR: from Supervised to Semi-Supervised Learning with Modern Architectures
http://arxiv.org/abs/1911.08460
AUTHORS:        Gabriel Synnaeve ; Qiantong Xu ; Jacob Kahn ; Tatiana Likhomanenko ; Edouard Grave ; Vineel Pratap ; Anuroop Sriram ; Vitaliy Liptchinsky ; Ronan Collobert
HIGHLIGHT:      We study pseudo-labeling for the semi-supervised training of ResNet, Time-Depth Separable ConvNets, and Transformers for speech recognition, with either CTC or Seq2Seq loss functions.

4, TITLE:       Neutron: An Implementation of the Transformer Translation Model and its Variants
http://arxiv.org/abs/1903.07402
AUTHORS:        Hongfei Xu ; Qiuhui Liu
HIGHLIGHT:      We implement the Neutron in this work, including the Transformer model and its several variants from most recent researches.

5, TITLE:       Exploiting Oxide Based Resistive RAM Variability for Bayesian Neural Network Hardware Design
http://arxiv.org/abs/1911.08555
AUTHORS:        Akul Malhotra ; Sen Lu ; Kezhou Yang ; Abhronil Sengupta
HIGHLIGHT:      This paper elaborates on a hardware design that exploits cycle-to-cycle variability of oxide based Resistive Random Access Memories (RRAMs) as a means to realize such a probabilistic sampling function, instead of viewing it as a disadvantage.

6, TITLE:       Redistribution Systems and PRAM
http://arxiv.org/abs/2003.08783
AUTHORS:        Paul Cohen ; Tomasz Loboda
COMMENTS:       arXiv admin note: substantial text overlap with arXiv:1902.05677
HIGHLIGHT:      We discuss the relationships between redistribution systems, agent-based systems, compartmental models and Bayesian models.

7, TITLE:       MUTATT: Visual-Textual Mutual Guidance for Referring Expression Comprehension
http://arxiv.org/abs/2003.08027
AUTHORS:        Shuai Wang ; Fan Lyu ; Wei Feng ; Song Wang
COMMENTS:       6 pages, Accepted by ICME-2020
HIGHLIGHT:      In this paper, we argue that for REC the referring expression and the target region are semantically correlated and subject, location and relationship consistency exist between vision and language.On top of this, we propose a novel approach called MutAtt to construct mutual guidance between vision and language, which treat vision and language equally thus yield compact information matching.

8, TITLE:       LT-Net: Label Transfer by Learning Reversible Voxel-wise Correspondence for One-shot Medical Image Segmentation
http://arxiv.org/abs/2003.07072
AUTHORS:        Shuxin Wang ; Shilei Cao ; Dong Wei ; Renzhen Wang ; Kai Ma ; Liansheng Wang ; Deyu Meng ; Yefeng Zheng
COMMENTS:       Accepted to Proc. IEEE Conf. Computer Vision and Pattern Recognition 2020
HIGHLIGHT:      We introduce a one-shot segmentation method to alleviate the burden of manual annotation for medical images.

9, TITLE:       Learning to Deblur and Generate High Frame Rate Video with an Event Camera
http://arxiv.org/abs/2003.00847
AUTHORS:        Chen Haoyu ; Teng Minggui ; Shi Boxin ; Wang YIzhou ; Huang Tiejun
HIGHLIGHT:      In this paper, we formulate the deblurring task on traditional cameras directed by events to be a residual learning one, and we propose corresponding network architectures for effective learning of deblurring and high frame rate video generation tasks.

10, TITLE:      Fixing the train-test resolution discrepancy: FixEfficientNet
http://arxiv.org/abs/2003.08237
AUTHORS:        Hugo Touvron ; Andrea Vedaldi ; Matthijs Douze ; Hervé Jégou
HIGHLIGHT:      This note complements the paper "Fixing the train-test resolution discrepancy" that introduced the FixRes method.

11, TITLE:      Provable Robust Learning Based on Transformation-Specific Smoothing
http://arxiv.org/abs/2002.12398
AUTHORS:        Linyi Li ; Maurice Weber ; Xiaojun Xu ; Luka Rimanic ; Tao Xie ; Ce Zhang ; Bo Li
COMMENTS:       Corrected typos in Appendix
HIGHLIGHT:      In this paper we aim to provide a unified framework for certifying ML model robustness against general adversarial transformations.

12, TITLE:      Aggregating explanation methods for stable and robust explainability
http://arxiv.org/abs/1903.00519
AUTHORS:        Laura Rieger ; Lars Kai Hansen
HIGHLIGHT:      Our contributions in this paper are twofold.

13, TITLE:      Reinforced Feature Points: Optimizing Feature Detection and Description for a High-Level Task
http://arxiv.org/abs/1912.00623
AUTHORS:        Aritra Bhowmik ; Stefan Gumhold ; Carsten Rother ; Eric Brachmann
COMMENTS:       CVPR 2020 (oral)
HIGHLIGHT:      We propose a new training methodology which embeds the feature detector in a complete vision pipeline, and where the learnable parameters are trained in an end-to-end fashion.

14, TITLE:      Teacher-Student chain for efficient semi-supervised histology image classification
http://arxiv.org/abs/2003.08797
AUTHORS:        Shayne Shaw ; Maciej Pajak ; Aneta Lisowska ; Sotirios A Tsaftaris ; Alison Q O'Neil
COMMENTS:       AI for Affordable Healthcare (AI4AH) workshop at ICLR 2020
HIGHLIGHT:      In this paper, we apply the semi-supervised teacher-student knowledge distillation technique proposed by Yalniz et al. (2019) to the task of quantifying prognostic features in colorectal cancer.

15, TITLE:      Scene Text Recognition via Transformer
http://arxiv.org/abs/2003.08077
AUTHORS:        Xinjie Feng ; Hongxun Yao ; Yuankai Qi ; Jun Zhang ; Shengping Zhang
HIGHLIGHT:      In this paper, we find that the rectification is completely unnecessary.

16, TITLE:      RatLesNetv2: A Fully Convolutional Network for Rodent Brain Lesion Segmentation
http://arxiv.org/abs/2001.09138
AUTHORS:        Juan Miguel Valverde ; Artem Shatillo ; Riccardo de Feo ; Olli Gröhn ; Alejandra Sierra ; Jussi Tohka
COMMENTS:       Fixed a data-related issue
HIGHLIGHT:      We present a three-dimensional fully convolutional neural network (ConvNet) named RatLesNetv2 for segmenting rodent brain lesions.

17, TITLE:      Spotting Macro- and Micro-expression Intervals in Long Video Sequences
http://arxiv.org/abs/1912.11985
AUTHORS:        Ying He ; Su-Jing Wang ; Jingting Li ; Moi Hoon Yap
COMMENTS:       7 pages, 4 figures and 3 tables
HIGHLIGHT:      This paper presents baseline results for the Third Facial Micro-Expression Grand Challenge (MEGC 2020).

18, TITLE:      Bias-Reduced Hindsight Experience Replay with Virtual Goal Prioritization
http://arxiv.org/abs/1905.05498
AUTHORS:        Binyamin Manela ; Armin Biess
HIGHLIGHT:      In this paper, we present two improvements over the existing HER algorithm.

19, TITLE:      Colored Noise Injection for Training Adversarially Robust Neural Networks
http://arxiv.org/abs/2003.02188
AUTHORS:        Evgenii Zheltonozhskii ; Chaim Baskin ; Yaniv Nemcovsky ; Brian Chmiel ; Avi Mendelson ; Alex M. Bronstein
HIGHLIGHT:      In this work we extend the idea of adding white Gaussian noise to the network weights and activations during adversarial training (PNI) to the injection of colored noise for defense against common white-box and black-box attacks.

20, TITLE:      Unrestricted Adversarial Examples via Semantic Manipulation
http://arxiv.org/abs/1904.06347
AUTHORS:        Anand Bhattad ; Min Jin Chong ; Kaizhao Liang ; Bo Li ; D. A. Forsyth
COMMENTS:       Accepted to ICLR 2020. First two authors contributed equally. Code: https://github.com/aisecure/Big-but-Invisible-Adversarial-Attack and Openreview: https://openreview.net/forum?id=Sye_OgHFwH
HIGHLIGHT:      In this paper, we instead introduce "unrestricted" perturbations that manipulate semantically meaningful image-based visual descriptors - color and texture - in order to generate effective and photorealistic adversarial examples.

21, TITLE:      Approximating the Existential Theory of the Reals
http://arxiv.org/abs/1810.01393
AUTHORS:        Argyrios Deligkas ; John Fearnley ; Themistoklis Melissourgos ; Paul G. Spirakis
COMMENTS:       In the proceedings of the 14th Conference on Web and Internet Economics (WINE 2018)
HIGHLIGHT:      In this paper we propose and study the approximate existential theory of the reals ($\epsilon$-ETR), in which the constraints only need to be satisfied approximately.

22, TITLE:      Adversarial Learning with Contextual Embeddings for Zero-resource Cross-lingual Classification and NER
http://arxiv.org/abs/1909.00153
AUTHORS:        Phillip Keung ; Yichao Lu ; Vikas Bhardwaj
COMMENTS:       In EMNLP 2019
HIGHLIGHT:      We improve upon multilingual BERT's zero-resource cross-lingual performance via adversarial learning.

23, TITLE:      Counting Polygon Triangulations is Hard
http://arxiv.org/abs/1903.04737
AUTHORS:        David Eppstein
COMMENTS:       24 pages, 11 figures. Expanded version of a paper from Proc. 35th International Symposium on Computational Geometry
HIGHLIGHT:      We prove that it is $\#\mathsf{P}$-complete to count the triangulations of a (non-simple) polygon.

24, TITLE:      Determination of the Mitotically Most Active Region for Computer-Aided Mitotic Count
http://arxiv.org/abs/1902.05414
AUTHORS:        Marc Aubreville ; Christof A. Bertram ; Christian Marzahl ; Corinne Gurtner ; Martina Dettwiler ; Anja Schmidt ; Florian Bartenschlager ; Sophie Merz ; Marco Fragoso ; Olivia Kershaw ; Robert Klopfleisch ; Andreas Maier
COMMENTS:       11 pages, 6 figures
HIGHLIGHT:      We aimed to assess the question, how significantly the area selection impacts the mitotic count, which has a known high inter-rater disagreement.

25, TITLE:      TrajectoryNet: a new spatio-temporal feature learning network for human motion prediction
http://arxiv.org/abs/1910.06583
AUTHORS:        Xiaoli Liu ; Jianqin Yin ; Jin Liu ; Pengxiang Ding ; Jun Liu ; Huaping Liu
HIGHLIGHT:      In this paper, we propose a new 2D CNN based network, TrajectoryNet, to predict future poses in the trajectory space.

26, TITLE:      Diversified Arbitrary Style Transfer via Deep Feature Perturbation
http://arxiv.org/abs/1909.08223
AUTHORS:        Zhizhong Wang ; Lei Zhao ; Haibo Chen ; Lihong Qiu ; Qihang Mo ; Sihuan Lin ; Wei Xing ; Dongming Lu
COMMENTS:       Accepted by CVPR2020
HIGHLIGHT:      In this paper, we tackle these limitations and propose a simple yet effective method for diversified arbitrary style transfer.

27, TITLE:      Image Harmonization Dataset iHarmony4: HCOCO, HAdobe5k, HFlickr, and Hday2night
http://arxiv.org/abs/1908.10526
AUTHORS:        Wenyan Cong ; Jianfu Zhang ; Li Niu ; Liu Liu ; Zhixin Ling ; Weiyuan Li ; Liqing Zhang
COMMENTS:       Our full paper arXiv:1911.13239 "DoveNet: Deep Image Harmonization via Domain Verification" is accepted by CVPR2020
HIGHLIGHT:      Therefore, we contribute an image harmonization dataset iHarmony4 by generating synthesized composite images based on existing COCO (resp., Adobe5k, day2night) dataset, leading to our HCOCO (resp., HAdobe5k, Hday2night) sub-dataset.

28, TITLE:      Measuring and improving the quality of visual explanations
http://arxiv.org/abs/2003.08774
AUTHORS:        Agnieszka Grabska-Barwińska
HIGHLIGHT:      Several methods have been proposed to highlight features important for a given network decision.

29, 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
HIGHLIGHT:      In this paper, we study an intermediate form of supervision, i.e., single-frame supervision, for temporal action localization (TAL).

30, TITLE:      A Review of Computational Approaches for Evaluation of Rehabilitation Exercises
http://arxiv.org/abs/2003.08767
AUTHORS:        Yalin Liao ; Aleksandar Vakanski ; Min Xian ; David Paul ; Russell Baker
COMMENTS:       29 pages, 1 figure
HIGHLIGHT:      Related work in the literature on data representation, feature engineering, movement segmentation, and scoring functions is presented.

31, TITLE:      Prior-based Domain Adaptive Object Detection for Hazy and Rainy Conditions
http://arxiv.org/abs/1912.00070
AUTHORS:        Vishwanath A. Sindagi ; Poojan Oza ; Rajeev Yasarla ; Vishal M. Patel
HIGHLIGHT:      To address this issue, we propose an unsupervised prior-based domain adversarial object detection framework for adapting the detectors to hazy and rainy conditions.

32, TITLE:      PointAugment: an Auto-Augmentation Framework for Point Cloud Classification
http://arxiv.org/abs/2002.10876
AUTHORS:        Ruihui Li ; Xianzhi Li ; Pheng-Ann Heng ; Chi-Wing Fu
COMMENTS:       Camera-Ready Version for CVPR 2020 (Oral); code is https://github.com/liruihui/PointAugment/
HIGHLIGHT:      We present PointAugment, a new auto-augmentation framework that automatically optimizes and augments point cloud samples to enrich the data diversity when we train a classification network.

33, TITLE:      LANCE: Efficient Low-Precision Quantized Winograd Convolution for Neural Networks Based on Graphics Processing Units
http://arxiv.org/abs/2003.08646
AUTHORS:        Guangli Li ; Lei Liu ; Xueying Wang ; Xiu Ma ; Xiaobing Feng
COMMENTS:       Accepted by ICASSP 2020
HIGHLIGHT:      In this paper, we propose an efficient low-precision quantized Winograd convolution algorithm, called LANCE, which combines the advantages of fast convolution and quantization techniques.

34, TITLE:      UGRWO-Sampling: A modified random walk under-sampling approach based on graphs to imbalanced data classification
http://arxiv.org/abs/2002.03521
AUTHORS:        Saeideh Roshanfekr ; Shahriar Esmaeili ; Hassan Ataeian ; Ali Amiri
COMMENTS:       34 pages, 3 figures, 9 tables
HIGHLIGHT:      In this paper, we propose a new RWO-Sampling (Random Walk Over-Sampling) based on graphs for imbalanced datasets.

35, TITLE:      Improving Model Training by Periodic Sampling over Weight Distributions
http://arxiv.org/abs/1905.05774
AUTHORS:        Samarth Tripathi ; Jiayi Liu ; Unmesh Kurup ; Mohak Shah ; Sauptik Dhar
HIGHLIGHT:      In this paper, we explore techniques centered around periodic sampling of model weights that provide convergence improvements on gradient update methods (vanilla \acs{SGD}, Momentum, Adam) for a variety of vision problems (classification, detection, segmentation).

36, TITLE:      Discretizing Continuous Action Space for On-Policy Optimization
http://arxiv.org/abs/1901.10500
AUTHORS:        Yunhao Tang ; Shipra Agrawal
COMMENTS:       Accepted at AAAI Conference on Artificial Intelligence (2020) in New York, NY, USA. An open source implementation can be found at https://github.com/robintyh1/onpolicybaselines
HIGHLIGHT:      In this work, we show that discretizing action space for continuous control is a simple yet powerful technique for on-policy optimization.

37, TITLE:      Weakly Supervised Object Detection with 2D and 3D Regression Neural Networks
http://arxiv.org/abs/1906.01891
AUTHORS:        Florian Dubost ; Hieab Adams ; Pinar Yilmaz ; Gerda Bortsova ; Gijs van Tulder ; M. Arfan Ikram ; Wiro Niessen ; Meike Vernooij ; Marleen de Bruijne
COMMENTS:       New formatting. A few changes in introduction, discussion and conclusion
HIGHLIGHT:      We propose a new weakly supervised detection method using neural networks, that computes attention maps revealing the locations of brain lesions.

38, TITLE:      QLMC-HD: Quasi Large Margin Classifier based on Hyperdisk
http://arxiv.org/abs/1902.09692
AUTHORS:        Hassan Ataeian ; Shahriar Esmaeili ; Ali Amiri ; Hossein Safari
COMMENTS:       10 pages, 1 figures, 10 tables
HIGHLIGHT:      Among these classifiers, we propose a method that is based on the maximum margin between two classes.

39, TITLE:      Parameterized Indexed Value Function for Efficient Exploration in Reinforcement Learning
http://arxiv.org/abs/1912.10577
AUTHORS:        Tian Tan ; Zhihan Xiong ; Vikranth R. Dwaracherla
COMMENTS:       17 pages, 4 figures, Proceedings of the 34th AAAI Conference on Artificial Intelligence
HIGHLIGHT:      In this paper, we present an alternative, computationally efficient way to induce exploration using index sampling.

40, TITLE:      Health-Informed Policy Gradients for Multi-Agent Reinforcement Learning
http://arxiv.org/abs/1908.01022
AUTHORS:        Ross E. Allen ; Javona White Bear ; Jayesh K. Gupta ; Mykel J. Kochenderfer
HIGHLIGHT:      This paper proposes a definition of system health in the context of multiple agents optimizing a joint reward function.

41, TITLE:      Option-Critic in Cooperative Multi-agent Systems
http://arxiv.org/abs/1911.12825
AUTHORS:        Jhelum Chakravorty ; Nadeem Ward ; Julien Roy ; Maxime Chevalier-Boisvert ; Sumana Basu ; Andrei Lupu ; Doina Precup
HIGHLIGHT:      In this paper, we investigate learning temporal abstractions in cooperative multi-agent systems, using the options framework (Sutton et al, 1999).

42, TITLE:      Recurrent Neural Networks: An Embedded Computing Perspective
http://arxiv.org/abs/1908.07062
AUTHORS:        Nesma M. Rezk ; Madhura Purnaprajna ; Tomas Nordström ; Zain Ul-Abdin
COMMENTS:       Accepted for publication in IEEE Access
HIGHLIGHT:      In this paper, we review existing implementations of RNN models on embedded platforms and discuss the methods adopted to overcome the limitations of embedded systems.

43, TITLE:      A Quadruplet Loss for Enforcing Semantically Coherent Embeddings in Multi-output Classification Problems
http://arxiv.org/abs/2002.11644
AUTHORS:        Hugo Proença ; Ehsan Yaghoubi ; Pendar Alirezazadeh
COMMENTS:       10 pages, 10 figures, 2 tables
HIGHLIGHT:      This paper describes one objective function for learning semantically coherent feature embeddings in multi-output classification problems, i.e., when the response variables have dimension higher than one.

44, TITLE:      Exemplar Normalization for Learning Deep Representation
http://arxiv.org/abs/2003.08761
AUTHORS:        Ruimao Zhang ; Zhanglin Peng ; Lingyun Wu ; Zhen Li ; Ping Luo
COMMENTS:       Accepted by CVPR2020, normalization methods, image classification
HIGHLIGHT:      This work investigates a novel dynamic learning-to-normalize (L2N) problem by proposing Exemplar Normalization (EN), which is able to learn different normalization methods for different convolutional layers and image samples of a deep network.