2020.06.24.txt

==========New Papers==========
1, TITLE:       Scale-Space Autoencoders for Unsupervised Anomaly Segmentation in Brain MRI
http://arxiv.org/abs/2006.12852
AUTHORS:        Christoph Baur ; Benedikt Wiestler ; Shadi Albarqouni ; Nassir Navab
HIGHLIGHT:      The modeling of the laplacian pyramid further enables the delineation and aggregation of lesions at multiple scales, which allows to effectively cope with different pathologies and lesion sizes using a single model.

2, TITLE:       Can you tell? SSNet -- a Sagittal Stratum-inspired Neural Network Framework for Sentiment Analysis
http://arxiv.org/abs/2006.12958
AUTHORS:        Apostol Vassilev ; Munawar Hasan
COMMENTS:       11 pages, 6 figures, 2 tables, 21 references
HIGHLIGHT:      We use this biological formation as the inspiration for designing a neural network architecture that combines predictions of different models on the same text to construct a robust, accurate and computationally efficient classifier for sentiment analysis.

3, TITLE:       Unsupervised Evaluation of Interactive Dialog with DialoGPT
http://arxiv.org/abs/2006.12719
AUTHORS:        Shikib Mehri ; Maxine Eskenazi
COMMENTS:       Published at to SIGdial 2020
HIGHLIGHT:      This paper introduces the FED metric (fine-grained evaluation of dialog), an automatic evaluation metric which uses DialoGPT, without any fine-tuning or supervision.

4, TITLE:       Distilling Object Detectors with Task Adaptive Regularization
http://arxiv.org/abs/2006.13108
AUTHORS:        Ruoyu Sun ; Fuhui Tang ; Xiaopeng Zhang ; Hongkai Xiong ; Qi Tian
HIGHLIGHT:      In this paper, we investigate each module of a typical detector in depth, and propose a general distillation framework that adaptively transfers knowledge from teacher to student according to the task specific priors.

5, TITLE:       On the Relationship Between Active Inference and Control as Inference
http://arxiv.org/abs/2006.12964
AUTHORS:        Beren Millidge ; Alexander Tschantz ; Anil K Seth ; Christopher L Buckley
COMMENTS:       initial upload
HIGHLIGHT:      In the context of this comparison, we highlight several ways in which these frameworks can inform one another.

6, TITLE:       PFGDF: Pruning Filter via Gaussian Distribution Feature for Deep Neural Networks Acceleration
http://arxiv.org/abs/2006.12963
AUTHORS:        Jianrong Xu ; Chao Li ; Bifeng Cui ; Kang Yang ; Yongjun Xu
HIGHLIGHT:      To solve this issue, we proposed a novel deep learning model compression acceleration method based on data distribution characteristics, namely Pruning Filter via Gaussian Distribution Feature(PFGDF) which was to found the smaller interval of the convolution layer of a certain layer to describe the original on the grounds of distribution characteristics .

7, TITLE:       Long-Term Prediction of Lane Change Maneuver Through a Multilayer Perceptron
http://arxiv.org/abs/2006.12769
AUTHORS:        Zhenyu Shou ; Ziran Wang ; Kyungtae Han ; Yongkang Liu ; Prashant Tiwari ; Xuan Di
COMMENTS:       Accepted by 31st IEEE Intelligent Vehicles Symposium
HIGHLIGHT:      In this study, we propose a longer-term (5~10 seconds) prediction model without any lateral or angle information.

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

9, TITLE:       Logical Neural Networks
http://arxiv.org/abs/2006.13155
AUTHORS:        Ryan Riegel ; Alexander Gray ; Francois Luus ; Naweed Khan ; Ndivhuwo Makondo ; Ismail Yunus Akhalwaya ; Haifeng Qian ; Ronald Fagin ; Francisco Barahona ; Udit Sharma ; Shajith Ikbal ; Hima Karanam ; Sumit Neelam ; Ankita Likhyani ; Santosh Srivastava
COMMENTS:       10 pages (incl. references), 38 pages supplementary, 7 figures, 9 tables, 6 algorithms. In submission to NeurIPS 2020
HIGHLIGHT:      We propose a novel framework seamlessly providing key properties of both neural nets (learning) and symbolic logic (knowledge and reasoning).

10, TITLE:      Surpassing Real-World Source Training Data: Random 3D Characters for Generalizable Person Re-Identification
http://arxiv.org/abs/2006.12774
AUTHORS:        Yanan Wang ; Shengcai Liao ; Ling Shao
COMMENTS:       https://github.com/VideoObjectSearch/RandPerson
HIGHLIGHT:      To address this, we propose to automatically synthesize a large-scale person re-identification dataset following a set-up similar to real surveillance but with virtual environments, and then use the synthesized person images to train a generalizable person re-identification model. As a result, we obtain a virtual dataset, called RandPerson, with 1,756,759 person images of 8,000 identities.

11, TITLE:      Maximizing Submodular or Monotone Functions under Partition Matroid Constraints by Multi-objective Evolutionary Algorithms
http://arxiv.org/abs/2006.12773
AUTHORS:        Anh Viet Do ; Frank Neumann
COMMENTS:       Paper accepted for publication in the proceedings of PPSN 2020
HIGHLIGHT:      In this work, we extend the theoretical results to partition matroid constraints which generalize cardinality constraints, and show that GSEMO can generally guarantee good approximation performance within polynomial expected run time.

12, TITLE:      hxtorch: PyTorch for ANNs on BrainScaleS-2
http://arxiv.org/abs/2006.13138
AUTHORS:        Philipp Spilger ; Eric Müller ; Arne Emmel ; Aron Leibfried ; Christian Mauch ; Christian Pehle ; Johannes Weis ; Oliver Breitwieser ; Sebastian Billaudelle ; Sebastian Schmitt ; Timo C. Wunderlich ; Yannik Stradmann ; Johannes Schemmel
HIGHLIGHT:      As an application of the introduced framework, we present a model that classifies activities of daily living with smartphone sensor data.

13, TITLE:      MANTRA: A Machine Learning reference lightcurve dataset for astronomical transient event recognition
http://arxiv.org/abs/2006.13163
AUTHORS:        Mauricio Neira ; Catalina Gómez ; John F. Suárez-Pérez ; Diego A. Gómez ; Juan Pablo Reyes ; Marcela Hernández Hoyos ; Pablo Arbeláez ; Jaime E. Forero-Romero
HIGHLIGHT:      We introduce MANTRA, an annotated dataset of 4869 transient and 71207 non-transient object lightcurves built from the Catalina Real Time Transient Survey.

14, TITLE:      Calibrated Adversarial Refinement for Multimodal Semantic Segmentation
http://arxiv.org/abs/2006.13144
AUTHORS:        Elias Kassapis ; Georgi Dikov ; Deepak K. Gupta ; Cedric Nugteren
HIGHLIGHT:      In this work, we aim to learn a calibrated multimodal predictive distribution, where the empirical frequency of the sampled predictions closely reflects that of the corresponding labels in the training set.

15, TITLE:      Provably Efficient Reinforcement Learning for Discounted MDPs with Feature Mapping
http://arxiv.org/abs/2006.13165
AUTHORS:        Dongruo Zhou ; Jiafan He ; Quanquan Gu
COMMENTS:       28 pages, 1 figure
HIGHLIGHT:      In this paper, we study reinforcement learning with feature mapping for discounted Markov Decision Processes (MDPs).

16, TITLE:      Joint Detection and Multi-Object Tracking with Graph Neural Networks
http://arxiv.org/abs/2006.13164
AUTHORS:        Yongxin Wang ; Xinshuo Weng ; Kris Kitani
HIGHLIGHT:      In this work, we propose a new approach for joint MOT based on Graph Neural Networks (GNNs).

17, TITLE:      Experience Replay with Likelihood-free Importance Weights
http://arxiv.org/abs/2006.13169
AUTHORS:        Samarth Sinha ; Jiaming Song ; Animesh Garg ; Stefano Ermon
HIGHLIGHT:      Prioritization or reweighting of important experiences has shown to improve performance of TD learning algorithms.In this work, we propose to reweight experiences based on their likelihood under the stationary distribution of the current policy.

18, TITLE:      LAMP: Large Deep Nets with Automated Model Parallelism for Image Segmentation
http://arxiv.org/abs/2006.12575
AUTHORS:        Wentao Zhu ; Can Zhao ; Wenqi Li ; Holger Roth ; Ziyue Xu ; Daguang Xu
COMMENTS:       MICCAI 2020 Early Accepted paper. Code is available\footnote{https://github.com/wentaozhu/lamp-automated-model-parallelism}
HIGHLIGHT:      In this work, we introduce Large deep 3D ConvNets with Automated Model Parallelism (LAMP) and investigate the impact of both input's and deep 3D ConvNets' size on segmentation accuracy.

19, TITLE:      ObjectNav Revisited: On Evaluation of Embodied Agents Navigating to Objects
http://arxiv.org/abs/2006.13171
AUTHORS:        Dhruv Batra ; Aaron Gokaslan ; Aniruddha Kembhavi ; Oleksandr Maksymets ; Roozbeh Mottaghi ; Manolis Savva ; Alexander Toshev ; Erik Wijmans
HIGHLIGHT:      We revisit the problem of Object-Goal Navigation (ObjectNav).

20, TITLE:      Encoding Legal Balancing: Automating an Abstract Ethico-Legal Value Ontology in Preference Logic
http://arxiv.org/abs/2006.12789
AUTHORS:        Christoph Benzmüller ; David Fuenmayor ; Bertram Lomfeld
COMMENTS:       13 pages, 14 figures
HIGHLIGHT:      We propose a holistic approach to formal modeling at different abstraction layers supported by a pluralistic framework in which the encoding of an ethico-legal value and upper ontology is developed in combination with the exploration of a formalization logic, with legal domain knowledge and with exemplary use cases until a reflective equilibrium is reached.

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

22, TITLE:      Boundary Regularized Building Footprint Extraction From Satellite Images Using Deep Neural Network
http://arxiv.org/abs/2006.13176
AUTHORS:        Kang Zhao ; Muhammad Kamran ; Gunho Sohn
HIGHLIGHT:      In this paper, we propose a novel deep neural network, which enables to jointly detect building instance and regularize noisy building boundary shapes from a single satellite imagery.

23, TITLE:      Laplacian Mixture Model Point Based Registration
http://arxiv.org/abs/2006.12582
AUTHORS:        Mohammad Sadegh Majdi ; Emad Fatemizadeh
HIGHLIGHT:      Here we introduce a novel method for matching of different data sets based on Laplacian distribution.

24, TITLE:      PipeSim: Trace-driven Simulation of Large-Scale AI Operations Platforms
http://arxiv.org/abs/2006.12587
AUTHORS:        Thomas Rausch ; Waldemar Hummer ; Vinod Muthusamy
COMMENTS:       11 pages, 13 figures, extended version of OpML'20 paper
HIGHLIGHT:      Our simulation model describes the interaction between pipelines and system infrastructure, and how pipeline tasks affect different ML model metrics.

25, TITLE:      Semantic Features Aided Multi-Scale Reconstruction of Inter-Modality Magnetic Resonance Images
http://arxiv.org/abs/2006.12585
AUTHORS:        Preethi Srinivasan ; Prabhjot Kaur ; Aditya Nigam ; Arnav Bhavsar
COMMENTS:       Accepted in IEEE CMBS 2020
HIGHLIGHT:      We propose a novel deep network based solution to reconstruct T2W images from T1W images (T1WI) using an encoder-decoder architecture.

26, TITLE:      Lumos: A Library for Diagnosing Metric Regressions in Web-Scale Applications
http://arxiv.org/abs/2006.12793
AUTHORS:        Jamie Pool ; Ebrahim Beyrami ; Vishak Gopal ; Ashkan Aazami ; Jayant Gupchup ; Jeff Rowland ; Binlong Li ; Pritesh Kanani ; Ross Cutler ; Johannes Gehrke
HIGHLIGHT:      In this work, we open source Lumos and present our results from applying it to two different components within the RTC group over millions of sessions.

27, TITLE:      Drive-Net: Convolutional Network for Driver Distraction Detection
http://arxiv.org/abs/2006.12586
AUTHORS:        Mohammed S. Majdi ; Sundaresh Ram ; Jonathan T. Gill ; Jeffery J. Rodriguez
HIGHLIGHT:      In this paper, we present an automated supervised learning method called Drive-Net for driver distraction detection.

28, TITLE:      PICO: Primitive Imitation for COntrol
http://arxiv.org/abs/2006.12551
AUTHORS:        Corban G. Rivera ; Katie M. Popek ; Chace Ashcraft ; Edward W. Staley ; Kapil D. Katyal ; Bart L. Paulhamus
HIGHLIGHT:      In this work, we explore a novel framework for control of complex systems called Primitive Imitation for Control PICO.

29, TITLE:      RayS: A Ray Searching Method for Hard-label Adversarial Attack
http://arxiv.org/abs/2006.12792
AUTHORS:        Jinghui Chen ; Quanquan Gu
COMMENTS:       9 pages, 4 figures, 9 tables. In KDD 2020
HIGHLIGHT:      In this paper, we present the Ray Searching attack (RayS), which greatly improves the hard-label attack effectiveness as well as efficiency.

30, TITLE:      Increased-Range Unsupervised Monocular Depth Estimation
http://arxiv.org/abs/2006.12791
AUTHORS:        Saad Imran ; Muhammad Umar Karim Khan ; Sikander Bin Mukarram ; Chong-Min Kyung
HIGHLIGHT:      In this work, we propose to integrate the advantages of the small and wide baselines.

31, TITLE:      MSMD-Net: Deep Stereo Matching with Multi-scale and Multi-dimension Cost Volume
http://arxiv.org/abs/2006.12797
AUTHORS:        Zhelun Shen ; Yuchao Dai ; Zhibo Rao
HIGHLIGHT:      In this paper, we propose MSMD-Net (Multi-Scale and Multi-Dimension) to construct multi-scale and multi-dimension cost volume.

32, TITLE:      Combining Neural Language Models for WordSense Induction
http://arxiv.org/abs/2006.13200
AUTHORS:        Nikolay Arefyev ; Boris Sheludko ; Tatiana Aleksashina
COMMENTS:       International Conference on Analysis of Images, Social Networks and Texts AIST 2019: Analysis of Images, Social Networks and Texts, pp 105-121
HIGHLIGHT:      In this work, we apply this approach to the Russian language and improve it in two ways.

33, TITLE:      Keyframe Segmentation and Positional Encoding for Video-guided Machine Translation Challenge 2020
http://arxiv.org/abs/2006.12799
AUTHORS:        Tosho Hirasawa ; Zhishen Yang ; Mamoru Komachi ; Naoaki Okazaki
COMMENTS:       4 pages; First Workshop on Advances in Language and Vision Research (ALVR 2020)
HIGHLIGHT:      In this work, we presented our video-guided machine translation system in approaching the Video-guided Machine Translation Challenge 2020.

34, TITLE:      Just How Toxic is Data Poisoning? A Unified Benchmark for Backdoor and Data Poisoning Attacks
http://arxiv.org/abs/2006.12557
AUTHORS:        Avi Schwarzschild ; Micah Goldblum ; Arjun Gupta ; John P Dickerson ; Tom Goldstein
COMMENTS:       19 pages, 4 figures
HIGHLIGHT:      In order to promote fair comparison in future work, we develop unified benchmarks for data poisoning and backdoor attacks.

35, TITLE:      Simple and Effective VAE Training with Calibrated Decoders
http://arxiv.org/abs/2006.13202
AUTHORS:        Oleh Rybkin ; Kostas Daniilidis ; Sergey Levine
COMMENTS:       Project website: \url{https://orybkin.github.io/sigma-vae/}
HIGHLIGHT:      In this work, we study how the performance of VAEs can be improved while not requiring the use of this heuristic hyperparameter, by learning calibrated decoders that accurately model the decoding distribution.

36, TITLE:      A Survey on Deep Learning for Localization and Mapping: Towards the Age of Spatial Machine Intelligence
http://arxiv.org/abs/2006.12567
AUTHORS:        Changhao Chen ; Bing Wang ; Chris Xiaoxuan Lu ; Niki Trigoni ; Andrew Markham
COMMENTS:       26 pages, 10 figures. Project website: https://github.com/changhao-chen/deep-learning-localization-mapping
HIGHLIGHT:      In this work, we provide a comprehensive survey, and propose a new taxonomy on the existing approaches on localization and mapping using deep learning.

37, TITLE:      Feature Expansive Reward Learning: Rethinking Human Input
http://arxiv.org/abs/2006.13208
AUTHORS:        Andreea Bobu ; Marius Wiggert ; Claire Tomlin ; Anca D. Dragan
COMMENTS:       16 pages, 15 figures
HIGHLIGHT:      We propose an algorithm for learning the feature from the raw state space and integrating it into the reward function.

38, TITLE:      Long-Horizon Visual Planning with Goal-Conditioned Hierarchical Predictors
http://arxiv.org/abs/2006.13205
AUTHORS:        Karl Pertsch ; Oleh Rybkin ; Frederik Ebert ; Chelsea Finn ; Dinesh Jayaraman ; Sergey Levine
COMMENTS:       Project page: orybkin.github.io/video-gcp
HIGHLIGHT:      In this work we propose a framework for visual prediction and planning that is able to overcome both of these limitations.

39, TITLE:      Calibration of Neural Networks using Splines
http://arxiv.org/abs/2006.12800
AUTHORS:        Kartik Gupta ; Amir Rahimi ; Thalaiyasingam Ajanthan ; Thomas Mensink ; Cristian Sminchisescu ; Richard Hartley
HIGHLIGHT:      In this work, we introduce a binning-free calibration measure inspired by the classical Kolmogorov-Smirnov (KS) statistical test in which the main idea is to compare the respective cumulative probability distributions.

40, TITLE:      3D Probabilistic Segmentation and Volumetry from 2D projection images
http://arxiv.org/abs/2006.12809
AUTHORS:        Athanasios Vlontzos ; Samuel Budd ; Benjamin Hou ; Daniel Rueckert ; Bernhard Kainz
HIGHLIGHT:      In this paper we explore probabilistic methods to reconstruct 3D volumetric images from 2D imaging modalities and measure the models' performance and confidence.

41, TITLE:      Post-hoc Calibration of Neural Networks
http://arxiv.org/abs/2006.12807
AUTHORS:        Amir Rahimi ; Kartik Gupta ; Thalaiyasingam Ajanthan ; Thomas Mensink ; Cristian Sminchisescu ; Richard Hartley
HIGHLIGHT:      In this work, we intend to understand the post-hoc calibration methods from a theoretical point of view.

42, TITLE:      Inductive Unsupervised Domain Adaptation for Few-Shot Classification via Clustering
http://arxiv.org/abs/2006.12816
AUTHORS:        Xin Cong ; Bowen Yu ; Tingwen Liu ; Shiyao Cui ; Hengzhu Tang ; Bin Wang
COMMENTS:       Accepted by ECML-PKDD 2020
HIGHLIGHT:      In this paper, we set out to tackle this issue by introducing a inductive framework, DaFeC, to improve Domain adaptation performance for Few-shot classification via Clustering.

43, TITLE:      NeuralScale: Efficient Scaling of Neurons for Resource-Constrained Deep Neural Networks
http://arxiv.org/abs/2006.12813
AUTHORS:        Eugene Lee ; Chen-Yi Lee
COMMENTS:       17 pages, 11 figures, accepted by CVPR as oral paper
HIGHLIGHT:      In this work, we attempt to search for the neuron (filter) configuration of a fixed network architecture that maximizes accuracy.

44, TITLE:      Joint Left Atrial Segmentation and Scar Quantification Based on a DNN with Spatial Encoding and Shape Attention
http://arxiv.org/abs/2006.13011
AUTHORS:        Lei Li ; Xin Weng ; Julia A. Schnabel ; Xiahai Zhuang
COMMENTS:       10 pages
HIGHLIGHT:      We propose an end-to-end deep neural network (DNN) which can simultaneously segment the left atrial (LA) cavity and quantify LA scars.

45, TITLE:      Extension of Direct Feedback Alignment to Convolutional and Recurrent Neural Network for Bio-plausible Deep Learning
http://arxiv.org/abs/2006.12830
AUTHORS:        Donghyeon Han ; Gwangtae Park ; Junha Ryu ; Hoi-jun Yoo
COMMENTS:       Submitted to WACV2021
HIGHLIGHT:      In this work, we propose a new DFA algorithm for BP-level accurate CNN and RNN training.

46, TITLE:      Sparse-RS: a versatile framework for query-efficient sparse black-box adversarial attacks
http://arxiv.org/abs/2006.12834
AUTHORS:        Francesco Croce ; Maksym Andriushchenko ; Naman D. Singh ; Nicolas Flammarion ; Matthias Hein
HIGHLIGHT:      In this paper we instead focus on query-efficient sparse attacks in the black-box setting.

47, TITLE:      Expert-Supervised Reinforcement Learning for Offline Policy Learning and Evaluation
http://arxiv.org/abs/2006.13189
AUTHORS:        Aaron Sonabend W ; Junwei Lu ; Leo A. Celi ; Tianxi Cai ; Peter Szolovits
HIGHLIGHT:      To overcome these issues, we propose an Expert-Supervised RL (ESRL) framework which uses uncertainty quantification for offline policy learning.

48, TITLE:      Efficient Spatially Adaptive Convolution and Correlation
http://arxiv.org/abs/2006.13188
AUTHORS:        Thomas W. Mitchel ; Benedict Brown ; David Koller ; Tim Weyrich ; Szymon Rusinkiewicz ; Michael Kazhdan
HIGHLIGHT:      In this work, we provide a general, representation-theoretic, framework that allows for spatially varying linear transformations to be applied to the filter.

49, TITLE:      Towards Robust Sensor Fusion in Visual Perception
http://arxiv.org/abs/2006.13192
AUTHORS:        Shaojie Wang ; Tong Wu ; Yevgeniy Vorobeychik
HIGHLIGHT:      In this work, we are interested in the behavior of different fusion methods under adversarial attacks on different sensors.

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

51, TITLE:      Instant 3D Object Tracking with Applications in Augmented Reality
http://arxiv.org/abs/2006.13194
AUTHORS:        Adel Ahmadyan ; Tingbo Hou ; Jianing Wei ; Liangkai Zhang ; Artsiom Ablavatski ; Matthias Grundmann
COMMENTS:       4 pages, five figures, CVPR Fourth Workshop on Computer Vision for AR/VR
HIGHLIGHT:      We propose an instant motion tracking system that tracks an object's pose in space (represented by its 3D bounding box) in real-time on mobile devices.

52, TITLE:      ELSIM: End-to-end learning of reusable skills through intrinsic motivation
http://arxiv.org/abs/2006.12903
AUTHORS:        Arthur Aubret ; Laetitia Matignon ; Salima Hassas
COMMENTS:       Accepted at ECML 2020
HIGHLIGHT:      Taking inspiration from developmental learning, we present a novel reinforcement learning architecture which hierarchically learns and represents self-generated skills in an end-to-end way.

53, TITLE:      Probabilistic Crowd GAN: Multimodal Pedestrian Trajectory Prediction using a Graph Vehicle-Pedestrian Attention Network
http://arxiv.org/abs/2006.12906
AUTHORS:        Stuart Eiffert ; Kunming Li ; Mao Shan ; Stewart Worrall ; Salah Sukkarieh ; Eduardo Nebot
COMMENTS:       Accepted for publication in IEEE Robotics and Automation Letters (RA-L) Copyright may be transferred without notice, after which this version may no longer be accessible
HIGHLIGHT:      This problem becomes increasingly complex when we consider the uncertainty and multimodality of pedestrian motion, as well as the implicit interactions between members of a crowd, including any response to a vehicle.

54, TITLE:      Security and Privacy Preserving Deep Learning
http://arxiv.org/abs/2006.12698
AUTHORS:        Saichethan Miriyala Reddy ; Saisree Miriyala
HIGHLIGHT:      In this chapter, we introduce differential privacy, which ensures that different kinds of statistical analyses dont compromise privacy and federated learning, training a machine learning model on a data to which we do not have access to.

55, TITLE:      Deep Attentive Wasserstein Generative Adversarial Networks for MRI Reconstruction with Recurrent Context-Awareness
http://arxiv.org/abs/2006.12915
AUTHORS:        Yifeng Guo ; Chengjia Wang ; Heye Zhang ; Guang Yang
HIGHLIGHT:      In this study, we propose a new deep learning-based CS-MRI reconstruction method to fully utilise the relationship among sequential MRI slices by coupling Wasserstein Generative Adversarial Networks (WGAN) with Recurrent Neural Networks.

56, TITLE:      Bridging the Theoretical Bound and Deep Algorithms for Open Set Domain Adaptation
http://arxiv.org/abs/2006.13022
AUTHORS:        Li Zhong ; Zhen Fang ; Feng Liu ; Bo Yuan ; Guangquan Zhang ; Jie Lu
HIGHLIGHT:      To address this issue, we propose a new upper bound of target-domain risk for UOSDA, which includes four terms: source-domain risk, $\epsilon$-open set difference ($\Delta_\epsilon$), a distributional discrepancy between domains, and a constant.

57, TITLE:      Particle Swarm Optimization for Energy Disaggregation in Industrial and Commercial Buildings
http://arxiv.org/abs/2006.12940
AUTHORS:        Karoline Brucke ; Stefan Arens ; Jan-Simon Telle ;  Sunke~Schlüters ; Benedikt Hanke ; Karsten von Maydell ; Carsten Agert
COMMENTS:       10 pages, 13 figures, 3 tables
HIGHLIGHT:      In this work we use two unlabelled power datasets with a granularity of 1 s. Therefore, the results are validated in the power domain in which good results regarding multiple error measures like root mean squared error or the percentage energy error can be shown.

58, TITLE:      Deep Learning of Unified Region, Edge, and Contour Models for Automated Image Segmentation
http://arxiv.org/abs/2006.12706
AUTHORS:        Ali Hatamizadeh
COMMENTS:       PhD dissertation, UCLA, 2020
HIGHLIGHT:      In this thesis, we devise novel methodologies that address these issues and establish robust representation learning frameworks for fully-automatic semantic segmentation in medical imaging and mainstream computer vision.

59, TITLE:      Learning Physical Constraints with Neural Projections
http://arxiv.org/abs/2006.12745
AUTHORS:        Shuqi Yang ; Xingzhe He ; Bo Zhu
HIGHLIGHT:      We propose a new family of neural networks to predict the behaviors of physical systems by learning their underpinning constraints.

60, TITLE:      Semi-Supervised Learning for Fetal Brain MRI Quality Assessment with ROI consistency
http://arxiv.org/abs/2006.12704
AUTHORS:        Junshen Xu ; Sayeri Lala ; Borjan Gagoski ; Esra Abaci Turk ; P. Ellen Grant ; Polina Golland ; Elfar Adalsteinsson
HIGHLIGHT:      Therefore, in this work, a semi-supervised deep learning method that detects slices with artifacts during the brain volume scan is proposed.

61, TITLE:      Efficient Hyperparameter Optimization in Deep Learning Using a Variable Length Genetic Algorithm
http://arxiv.org/abs/2006.12703
AUTHORS:        Xueli Xiao ; Ming Yan ; Sunitha Basodi ; Chunyan Ji ; Yi Pan
HIGHLIGHT:      In this article, we propose to use a variable length genetic algorithm (GA) to systematically and automatically tune the hyperparameters of a CNN to improve its performance.

62, TITLE:      CIE XYZ Net: Unprocessing Images for Low-Level Computer Vision Tasks
http://arxiv.org/abs/2006.12709
AUTHORS:        Mahmoud Afifi ; Abdelrahman Abdelhamed ; Abdullah Abuolaim ; Abhijith Punnappurath ; Michael S. Brown
HIGHLIGHT:      Leveraging this canonical image state, we propose a deep learning framework, CIE XYZ Net, that can unprocess a nonlinear image back to the canonical CIE XYZ image.

63, TITLE:      iffDetector: Inference-aware Feature Filtering for Object Detection
http://arxiv.org/abs/2006.12708
AUTHORS:        Mingyuan Mao ; Yuxin Tian ; Baochang Zhang ; Qixiang Ye ; Wanquan Liu ; Guodong Guo ; David Doermann
COMMENTS:       14 pages, 6 figures
HIGHLIGHT:      In this paper, we propose a new feature optimization approach to enhance features and suppress background noise in both the training and inference stages.

64, TITLE:      PoseGAN: A Pose-to-Image Translation Framework for Camera Localization
http://arxiv.org/abs/2006.12712
AUTHORS:        Kanglin Liu ; Qing Li ; Guoping Qiu
HIGHLIGHT:      This paper introduces a pose-to-image translation framework to tackle the camera localization problem.

65, TITLE:      A Framework for Fairness in Two-Sided Marketplaces
http://arxiv.org/abs/2006.12756
AUTHORS:        Kinjal Basu ; Cyrus DiCiccio ; Heloise Logan ; Noureddine El Karoui
COMMENTS:       15 pages, 7 Tables
HIGHLIGHT:      In this paper, we propose a definition and develop an end-to-end framework for achieving fairness while building such machine learning systems at scale.

66, TITLE:      Symmetries, graph properties, and quantum speedups
http://arxiv.org/abs/2006.12760
AUTHORS:        Shalev Ben-David ; Andrew M. Childs ; András Gilyén ; William Kretschmer ; Supartha Podder ; Daochen Wang
COMMENTS:       46 pages. Subsumes arXiv:2001.09642 and arXiv:2001.10520; adds a characterization of permutation groups with speedup and an exponential speedup for adjacency-list graph property testing
HIGHLIGHT:      In this work, we prove that hypergraph symmetries in the adjacency matrix model allow at most a polynomial separation between randomized and quantum query complexities.

67, TITLE:      DCNNs: A Transfer Learning comparison of Full Weapon Family threat detection forDual-Energy X-Ray Baggage Imagery
http://arxiv.org/abs/2006.13065
AUTHORS:        A. Williamson ; P. Dickinson ; T. Lambrou ; J. C. Murray
COMMENTS:       Submitted to BMVC 2019 Workshop on "Object Detection and Recognition for Security Screening"
HIGHLIGHT:      In this work we propose the first pipeline to effectively process Dual-Energy X-Ray scanner output, and perform classification capable of distinguishing between firearm families (Assault Rifle, Revolver, Self-Loading Pistol,Shotgun, and Sub-Machine Gun) from this output.

68, TITLE:      Benchmarking features from different radiomics toolkits / toolboxes using Image Biomarkers Standardization Initiative
http://arxiv.org/abs/2006.12761
AUTHORS:        Mingxi Lei ; Bino Varghese ; Darryl Hwang ; Steven Cen ; Xiaomeng Lei ; Afshin Azadikhah ; Bhushan Desai ; Assad Oberai ; Vinay Duddalwar
COMMENTS:       21 pages, 8 figures
HIGHLIGHT:      In this study, the image biomarker standardization initiative (IBSI) established phantom and benchmark values were used to compare the variation of the radiomic features while using 6 publicly available software programs and 1 in-house radiomics pipeline.

69, TITLE:      Perceptual Adversarial Robustness: Defense Against Unseen Threat Models
http://arxiv.org/abs/2006.12655
AUTHORS:        Cassidy Laidlaw ; Sahil Singla ; Soheil Feizi
HIGHLIGHT:      We present adversarial attacks and defenses for the perceptual adversarial threat model: the set of all perturbations to natural images which can mislead a classifier but are imperceptible to human eyes.

70, TITLE:      Domain Adaptation for Semantic Parsing
http://arxiv.org/abs/2006.13071
AUTHORS:        Zechang Li ; Yuxuan Lai ; Yansong Feng ; Dongyan Zhao
COMMENTS:       Accepted by IJCAI2020
HIGHLIGHT:      In this paper, we propose a novel semantic parser for domain adaptation, where we have much fewer annotated data in the target domain compared to the source domain.

71, TITLE:      Reduction From Non-Unique Games To Boolean Unique Games
http://arxiv.org/abs/2006.13073
AUTHORS:        Ronen Eldan ; Dana Moshkovitz
HIGHLIGHT:      We reduce the problem of proving a "Boolean Unique Games Conjecture" (with gap 1-delta vs. 1-C*delta, for any C> 1, and sufficiently small delta>0) to the problem of proving a PCP Theorem for a certain non-unique game.

72, TITLE:      Automatic Data Augmentation for Generalization in Deep Reinforcement Learning
http://arxiv.org/abs/2006.12862
AUTHORS:        Roberta Raileanu ; Max Goldstein ; Denis Yarats ; Ilya Kostrikov ; Rob Fergus
HIGHLIGHT:      In this paper, we compare three approaches for automatically finding an appropriate augmentation.

73, TITLE:      Object recognition through pose and shape estimation
http://arxiv.org/abs/2006.12864
AUTHORS:        Anitta D ; Annis Fathima A
HIGHLIGHT:      The purpose of this paper is to review many state of art which is already available for finding the pose of object based on shape, based on appearance, based on feature and comparison for its accuracy, complexity and performance

74, TITLE:      Single-Shot 3D Detection of Vehicles from Monocular RGB Images via Geometry Constrained Keypoints in Real-Time
http://arxiv.org/abs/2006.13084
AUTHORS:        Nils Gählert ; Jun-Jun Wan ; Nicolas Jourdan ; Jan Finkbeiner ; Uwe Franke ; Joachim Denzler
COMMENTS:       2020 IEEE IV Symposium
HIGHLIGHT:      In this paper we propose a novel 3D single-shot object detection method for detecting vehicles in monocular RGB images.

75, TITLE:      AFDet: Anchor Free One Stage 3D Object Detection
http://arxiv.org/abs/2006.12671
AUTHORS:        Runzhou Ge ; Zhuangzhuang Ding ; Yihan Hu ; Yu Wang ; Sijia Chen ; Li Huang ; Yuan Li
COMMENTS:       Accepted on May 6th, 2020 by CVPRW 2020, published on June 7th, 2020; Baseline detector for the 1st place solutions of Waymo Open Dataset Challenges 2020
HIGHLIGHT:      Most previous works try to solve it using anchor-based detection methods which come with two drawbacks: post-processing is relatively complex and computationally expensive; tuning anchor parameters is tricky.

76, TITLE:      Generalized Grasping for Mechanical Grippers for Unknown Objects with Partial Point Cloud Representations
http://arxiv.org/abs/2006.12676
AUTHORS:        Michael Hegedus ; Kamal Gupta ; Mehran Mehrandezh
COMMENTS:       8 pages, 12 figures, submitted to 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2018) on 2/24/2020
HIGHLIGHT:      We present a generalized grasping algorithm that uses point clouds (i.e. a group of points and their respective surface normals) to discover grasp pose solutions for multiple grasp types, executed by a mechanical gripper, in near real-time.

77, TITLE:      Inexact Derivative-Free Optimization for Bilevel Learning
http://arxiv.org/abs/2006.12674
AUTHORS:        Matthias J. Ehrhardt ; Lindon Roberts
HIGHLIGHT:      In this work we propose to solve these problems using inexact derivative-free optimization algorithms which never require to solve the lower-level problem exactly.

78, TITLE:      NLPContributions: An Annotation Scheme for Machine Reading of Scholarly Contributions in Natural Language Processing Literature
http://arxiv.org/abs/2006.12870
AUTHORS:        Jennifer D'Souza ; Sören Auer
COMMENTS:       Submitted for review at 1st Workshop on Extraction and Evaluation of Knowledge Entities from Scientific Documents (EEKE2020) at the ACM/IEEE Joint Conference on Digital Libraries 2020 (JCDL2020), Wuhan, China
HIGHLIGHT:      In this article, we describe the outcomes of this pilot annotation phase.

79, TITLE:      Non-parametric spatially constrained local prior for scene parsing on real-world data
http://arxiv.org/abs/2006.12874
AUTHORS:        Ligang Zhang
COMMENTS:       10 pages, journal
HIGHLIGHT:      In this paper, we present the non-parametric Spatially Constrained Local Prior (SCLP) for scene parsing on realistic data.

80, TITLE:      Towards Contrastive Explanations for Comparing the Ethics of Plans
http://arxiv.org/abs/2006.12632
AUTHORS:        Benjamin Krarup ; Senka Krivic ; Felix Lindner ; Derek Long
COMMENTS:       Accepted to the ICRA-AGAINST-20 workshop
HIGHLIGHT:      In this paper, we present how contrastive explanations can be used for comparing the ethics of plans.

81, TITLE:      Direct Feedback Alignment Scales to Modern Deep Learning Tasks and Architectures
http://arxiv.org/abs/2006.12878
AUTHORS:        Julien Launay ; Iacopo Poli ; François Boniface ; Florent Krzakala
COMMENTS:       22 pages, 5 figures, 10 tables. For associated code, see https://github.com/lightonai/dfa-scales-to-modern-deep-learning
HIGHLIGHT:      Here, we challenge this perspective, and study the applicability of Direct Feedback Alignment to neural view synthesis, recommender systems, geometric learning, and natural language processing.

82, TITLE:      RP2K: A Large-Scale Retail Product Dataset forFine-Grained Image Classification
http://arxiv.org/abs/2006.12634
AUTHORS:        Jingtian Peng ; Chang Xiao ; Xun Wei ; Yifan Li
HIGHLIGHT:      We introduce RP2K, a new large-scale retail product dataset for fine-grained image classification.

83, TITLE:      Information-theoretic User Interaction: Significant Inputs for Program Synthesis
http://arxiv.org/abs/2006.12638
AUTHORS:        Ashish Tiwari ; Arjun Radhakrishna ; Sumit Gulwani ; Daniel Perelman
HIGHLIGHT:      Motivated by the need to find the most pertinent question to ask the user, in this paper, we introduce the {\em significant questions problem}, and show that it is hard in general.

84, TITLE:      Contrastive Generative Adversarial Networks
http://arxiv.org/abs/2006.12681
AUTHORS:        Minguk Kang ; Jaesik Park
COMMENTS:       23 pages, 13 figures
HIGHLIGHT:      In this paper, we propose a novel conditional contrastive loss to maximize a lower bound on mutual information between samples from the same class.

85, TITLE:      SLV: Spatial Likelihood Voting for Weakly Supervised Object Detection
http://arxiv.org/abs/2006.12884
AUTHORS:        Ze Chen ; Zhihang Fu ; Rongxin Jiang ; Yaowu Chen ; Xian-sheng Hua
COMMENTS:       Accepted by CVPR2020
HIGHLIGHT:      In this paper, we propose a spatial likelihood voting (SLV) module to converge the proposal localizing process without any bounding box annotations.

86, TITLE:      Limits of Transfer Learning
http://arxiv.org/abs/2006.12694
AUTHORS:        Jake Williams ; Abel Tadesse ; Tyler Sam ; Huey Sun ; George D. Montanez
COMMENTS:       Accepted for presentation at the Sixth International Conference on Machine Learning, Optimization, and Data Science (LOD 2020), July 19-23, 2020
HIGHLIGHT:      To address this, we prove several novel results related to transfer learning, showing the need to carefully select which sets of information to transfer and the need for dependence between transferred information and target problems.

87, TITLE:      Exploring Software Naturalness throughNeural Language Models
http://arxiv.org/abs/2006.12641
AUTHORS:        Luca Buratti ; Saurabh Pujar ; Mihaela Bornea ; Scott McCarley ; Yunhui Zheng ; Gaetano Rossiello ; Alessandro Morari ; Jim Laredo ; Veronika Thost ; Yufan Zhuang ; Giacomo Domeniconi
HIGHLIGHT:      We explore this hypothesis through the use of a pre-trained transformer-based language model to perform code analysis tasks.

88, TITLE:      The Effect of Multi-step Methods on Overestimation in Deep Reinforcement Learning
http://arxiv.org/abs/2006.12692
AUTHORS:        Lingheng Meng ; Rob Gorbet ; Dana Kulić
COMMENTS:       7 pages, 4 figures, the 25th International Conference on Pattern Recognition (ICPR)
HIGHLIGHT:      In this work, we analyze the effect of multi-step methods on alleviating the overestimation problem in DRL, where multi-step experiences are sampled from a replay buffer.

89, TITLE:      Automatic Kernel Generation for Volta Tensor Cores
http://arxiv.org/abs/2006.12645
AUTHORS:        Somashekaracharya G. Bhaskaracharya ; Julien Demouth ; Vinod Grover
HIGHLIGHT:      In this paper, we describe a polyhedral approach to generate efficient CUDA kernels for matrix multiplication using inline assembly instructions for programming tensor cores on NVIDIA Volta GPUs.

90, TITLE:      Scribble2Label: Scribble-Supervised Cell Segmentation via Self-Generating Pseudo-Labels with Consistency
http://arxiv.org/abs/2006.12890
AUTHORS:        Hyeonsoo Lee ; Won-Ki Jeong
COMMENTS:       MICCAI 2020 accepted
HIGHLIGHT:      In this paper, we introduce Scribble2Label, a novel weakly-supervised cell segmentation framework that exploits only a handful of scribble annotations without full segmentation labels.

91, TITLE:      Coverage Path Planning with Track Spacing Adaptation for Autonomous Underwater Vehicles
http://arxiv.org/abs/2006.12896
AUTHORS:        Veronika Yordanova ; Bart Gips
COMMENTS:       Accepted submission at IEEE Robotics and Automation Letters (RA-L); 8 pages, 6 figures
HIGHLIGHT:      In this paper we address the mine countermeasures (MCM) search problem for an autonomous underwater vehicle (AUV) surveying the seabed using a side-looking sonar. To assess the algorithm, we collected data from three at-sea experiments.



==========Updates to Previous Papers==========
1, TITLE:       A Fast Stochastic Plug-and-Play ADMM for Imaging Inverse Problems
http://arxiv.org/abs/2006.11630
AUTHORS:        Junqi Tang ; Mike Davies
HIGHLIGHT:      In this work we propose an efficient stochastic plug-and-play (PnP) algorithm for imaging inverse problems.

2, TITLE:       Deep Gradient Compression: Reducing the Communication Bandwidth for Distributed Training
http://arxiv.org/abs/1712.01887
AUTHORS:        Yujun Lin ; Song Han ; Huizi Mao ; Yu Wang ; William J. Dally
COMMENTS:       we find 99.9% of the gradient exchange in distributed SGD is redundant; we reduce the communication bandwidth by two orders of magnitude without losing accuracy. Code is available at: https://github.com/synxlin/deep-gradient-compression
HIGHLIGHT:      In this paper, we find 99.9% of the gradient exchange in distributed SGD is redundant, and propose Deep Gradient Compression (DGC) to greatly reduce the communication bandwidth.

3, TITLE:       Towards Mesh Saliency Detection in 6 Degrees of Freedom
http://arxiv.org/abs/2005.13127
AUTHORS:        Xiaoying Ding ; Zhenzhong Chen
HIGHLIGHT:      In this work, a novel 6DoF mesh saliency database is developed which provides both the subject's 6DoF data and eye-movement data.

4, TITLE:       Learn to cycle: Time-consistent feature discovery for action recognition
http://arxiv.org/abs/2006.08247
AUTHORS:        Alexandros Stergiou ; Ronald Poppe
HIGHLIGHT:      We introduce Squeeze and Recursion Temporal Gates (SRTG), an approach that favors inputs with similar activations with potential temporal variations.

5, TITLE:       Deep Iterative Surface Normal Estimation
http://arxiv.org/abs/1904.07172
AUTHORS:        Jan Eric Lenssen ; Christian Osendorfer ; Jonathan Masci
COMMENTS:       Presented at CVPR 2020
HIGHLIGHT:      This paper presents an end-to-end differentiable algorithm for robust and detail-preserving surface normal estimation on unstructured point-clouds.

6, TITLE:       Image Classification in the Dark using Quanta Image Sensors
http://arxiv.org/abs/2006.02026
AUTHORS:        Abhiram Gnanasambandam ; Stanley H. Chan
HIGHLIGHT:      In this paper, we present a new low-light image classification solution using Quanta Image Sensors (QIS).

7, TITLE:       Learning Efficient Multi-agent Communication: An Information Bottleneck Approach
http://arxiv.org/abs/1911.06992
AUTHORS:        Rundong Wang ; Xu He ; Runsheng Yu ; Wei Qiu ; Bo An ; Zinovi Rabinovich
COMMENTS:       ICML 2020
HIGHLIGHT:      In this paper, we develop an Informative Multi-Agent Communication (IMAC) method to learn efficient communication protocols as well as scheduling.

8, TITLE:       Distill, Adapt, Distill: Training Small, In-Domain Models for Neural Machine Translation
http://arxiv.org/abs/2003.02877
AUTHORS:        Mitchell A. Gordon ; Kevin Duh
COMMENTS:       Accepted to WNGT 2020 Workshop at ACL 2020 Conference. Code is at http://github.com/mitchellgordon95/kd-aug
HIGHLIGHT:      We explore best practices for training small, memory efficient machine translation models with sequence-level knowledge distillation in the domain adaptation setting.

9, TITLE:       Multi-band Weighted $l_p$ Norm Minimization for Image Denoising
http://arxiv.org/abs/1901.04206
AUTHORS:        Yanchi Su ; Zhanshan Li ; Haihong Yu ; Zeyu Wang
COMMENTS:       accepted by Information Sciences
HIGHLIGHT:      To address this problem, we propose a flexible and precise model named multi-band weighted $l_p$ norm minimization (MBWPNM).

10, TITLE:      Universal Lower-Bounds on Classification Error under Adversarial Attacks and Random Corruption
http://arxiv.org/abs/2006.09989
AUTHORS:        Elvis Dohmatob
HIGHLIGHT:      Our contributions are three-fold.

11, TITLE:      Quantum Logspace Algorithm for Powering Matrices with Bounded Norm
http://arxiv.org/abs/2006.04880
AUTHORS:        Uma Girish ; Ran Raz ; Wei Zhan
HIGHLIGHT:      We give a quantum logspace algorithm for powering contraction matrices, that is, matrices with spectral norm at most~1.

12, TITLE:      Ranked List Loss for Deep Metric Learning
http://arxiv.org/abs/1903.03238
AUTHORS:        Xinshao Wang ; Yang Hua ; Elyor Kodirov ; Neil M. Robertson
COMMENTS:       TPAMI Extension of CVPR-19. Concretely, the new improvements are summarised at the end/appendix. Our source code is available online: https://github.com/XinshaoAmosWang/Ranked-List-Loss-for-DML
HIGHLIGHT:      In this work, we unveil two limitations of existing ranking-motivated structured losses and propose a novel ranked list loss to solve both of them.

13, TITLE:      Always-On, Sub-300-nW, Event-Driven Spiking Neural Network based on Spike-Driven Clock-Generation and Clock- and Power-Gating for an Ultra-Low-Power Intelligent Device
http://arxiv.org/abs/2006.12314
AUTHORS:        Dewei Wang ; Pavan Kumar Chundi ; Sung Justin Kim ; Minhao Yang ; Joao Pedro Cerqueira ; Joonsung Kang ; Seungchul Jung ; Sangjoon Kim ; Mingoo Seok
HIGHLIGHT:      Toward this goal, we present a novel SNN classifier architecture for always-on functions, demonstrating sub-300nW power consumption at the competitive inference accuracy for a KWS and other always-on classification workloads.

14, TITLE:      AdvJND: Generating Adversarial Examples with Just Noticeable Difference
http://arxiv.org/abs/2002.00179
AUTHORS:        Zifei Zhang ; Kai Qiao ; Lingyun Jiang ; Linyuan Wang ; Bin Yan
HIGHLIGHT:      To alleviate the tradeoff between the attack success rate and image fidelity, we propose a method named AdvJND, adding visual model coefficients, just noticeable difference coefficients, in the constraint of a distortion function when generating adversarial examples.

15, TITLE:      Siamese Meta-Learning and Algorithm Selection with 'Algorithm-Performance Personas' [Proposal]
http://arxiv.org/abs/2006.12328
AUTHORS:        Joeran Beel ; Bryan Tyrell ; Edward Bergman ; Andrew Collins ; Shahad Nagoor
COMMENTS:       7th Workshop on Automated Machine Learning (AutoML 2020)
HIGHLIGHT:      We propose a Siamese Neural Network architecture for automated algorithm selection that focuses more on 'alike performing' instances than meta-features.

16, TITLE:      Real-time Universal Style Transfer on High-resolution Images via Zero-channel Pruning
http://arxiv.org/abs/2006.09029
AUTHORS:        Jie An ; Tao Li ; Haozhi Huang ; Li Shen ; Xuan Wang ; Yongyi Tang ; Jinwen Ma ; Wei Liu ; Jiebo Luo
HIGHLIGHT:      In this work, we propose a lightweight alternative architecture - ArtNet, which is based on GoogLeNet, and later pruned by a novel channel pruning method named Zero-channel Pruning specially designed for style transfer approaches.

17, TITLE:      Soft Threshold Weight Reparameterization for Learnable Sparsity
http://arxiv.org/abs/2002.03231
AUTHORS:        Aditya Kusupati ; Vivek Ramanujan ; Raghav Somani ; Mitchell Wortsman ; Prateek Jain ; Sham Kakade ; Ali Farhadi
COMMENTS:       19 pages, 10 figures, Published at International Conference on Machine Learning (ICML) 2020
HIGHLIGHT:      This work proposes Soft Threshold Reparameterization (STR), a novel use of the soft-threshold operator on DNN weights.

18, TITLE:      Detecting Emergent Intersectional Biases: Contextualized Word Embeddings Contain a Distribution of Human-like Biases
http://arxiv.org/abs/2006.03955
AUTHORS:        Wei Guo ; Aylin Caliskan
COMMENTS:       20 pages, 2 figures, 3 tables
HIGHLIGHT:      We propose a new comprehensive method, Contextualized Embedding Association Test (CEAT), based on the distribution of 10,000 pooled effect magnitudes of bias in embedding variations and a random-effects model, dispensing with templates.

19, TITLE:      Multilogue-Net: A Context Aware RNN for Multi-modal Emotion Detection and Sentiment Analysis in Conversation
http://arxiv.org/abs/2002.08267
AUTHORS:        Aman Shenoy ; Ashish Sardana
COMMENTS:       10 pages, 3 figures, 5 tables; Published in Proceedings of the Second Grand Challenge and Workshop on Multimodal Language (Challenge-HML) in the 58th Annual Meeting of the Association for Computational Linguistics (ACL 2020)
HIGHLIGHT:      In this paper, we propose an end to end RNN architecture that attempts to take into account all the mentioned drawbacks.

20, TITLE:      Towards Understanding Linear Value Decomposition in Cooperative Multi-Agent Q-Learning
http://arxiv.org/abs/2006.00587
AUTHORS:        Jianhao Wang ; Zhizhou Ren ; Beining Han ; Chongjie Zhang
HIGHLIGHT:      In this paper, we introduce a variant of the fitted Q-iteration framework for analyzing multi-agent Q-learning with value decomposition.

21, TITLE:      Adaptive Reward-Poisoning Attacks against Reinforcement Learning
http://arxiv.org/abs/2003.12613
AUTHORS:        Xuezhou Zhang ; Yuzhe Ma ; Adish Singla ; Xiaojin Zhu
HIGHLIGHT:      We categorize such attacks by the infinity-norm constraint on $\delta_t$: We provide a lower threshold below which reward-poisoning attack is infeasible and RL is certified to be safe; we provide a corresponding upper threshold above which the attack is feasible.

22, TITLE:      Phase Portraits as Movement Primitives for Fast Humanoid Robot Control
http://arxiv.org/abs/1912.03535
AUTHORS:        Guilherme Maeda ; Okan Koc ; Jun Morimoto
HIGHLIGHT:      This article introduces Phase Portrait Movement Primitives (PPMP), a primitive that predicts dynamics on a low dimensional phase space which in turn is used to govern the high dimensional kinematics of the task.

23, TITLE:      Iterative Effect-Size Bias in Ridehailing: Measuring Social Bias in Dynamic Pricing of 100 Million Rides
http://arxiv.org/abs/2006.04599
AUTHORS:        Akshat Pandey ; Aylin Caliskan
COMMENTS:       16 pages, 6 tables, 6 figures
HIGHLIGHT:      In this work we develop a random-effects based metric for the analysis of social bias in supervised machine learning prediction models where model outputs depend on U.S. locations.

24, TITLE:      BlockGAN: Learning 3D Object-aware Scene Representations from Unlabelled Images
http://arxiv.org/abs/2002.08988
AUTHORS:        Thu Nguyen-Phuoc ; Christian Richardt ; Long Mai ; Yong-Liang Yang ; Niloy Mitra
COMMENTS:       For project page, see https://www.monkeyoverflow.com/#/blockgan/
HIGHLIGHT:      We present BlockGAN, an image generative model that learns object-aware 3D scene representations directly from unlabelled 2D images.

25, TITLE:      SEEK: Segmented Embedding of Knowledge Graphs
http://arxiv.org/abs/2005.00856
AUTHORS:        Wentao Xu ; Shun Zheng ; Liang He ; Bin Shao ; Jian Yin ; Tie-Yan Liu
HIGHLIGHT:      To mitigate this problem, we propose a lightweight modeling framework that can achieve highly competitive relational expressiveness without increasing the model complexity.

26, TITLE:      Category-wise Attack: Transferable Adversarial Examples for Anchor Free Object Detection
http://arxiv.org/abs/2003.04367
AUTHORS:        Quanyu Liao ; Xin Wang ; Bin Kong ; Siwei Lyu ; Youbing Yin ; Qi Song ; Xi Wu
HIGHLIGHT:      In this work, we aim to present an effective and efficient algorithm to generate adversarial examples to attack anchor-free object models based on two approaches.

27, TITLE:      EndoL2H: Deep Super-Resolution for Capsule Endoscopy
http://arxiv.org/abs/2002.05459
AUTHORS:        Yasin Almalioglu ; Kutsev Bengisu Ozyoruk ; Abdulkadir Gokce ; Kagan Incetan ; Guliz Irem Gokceler ; Muhammed Ali Simsek ; Kivanc Ararat ; Richard J. Chen ; Nicholas J. Durr ; Faisal Mahmood ; Mehmet Turan
COMMENTS:       23 pages, submitted to IEEE Transactions on Medical Imaging, corresponding Author: Mehmet Turan
HIGHLIGHT:      In this work, we propose and quantitatively validate a novel framework to learn a mapping from low-to-high resolution endoscopic images.

28, TITLE:      On the interaction between supervision and self-play in emergent communication
http://arxiv.org/abs/2002.01093
AUTHORS:        Ryan Lowe ; Abhinav Gupta ; Jakob Foerster ; Douwe Kiela ; Joelle Pineau
COMMENTS:       The first two authors contributed equally. Accepted at ICLR 2020
HIGHLIGHT:      In this paper, we investigate the relationship between two categories of learning signals with the ultimate goal of improving sample efficiency: imitating human language data via supervised learning, and maximizing reward in a simulated multi-agent environment via self-play (as done in emergent communication), and introduce the term supervised self-play (S2P) for algorithms using both of these signals.

29, TITLE:      First Steps Towards a Runtime Analysis When Starting With a Good Solution
http://arxiv.org/abs/2006.12161
AUTHORS:        Denis Antipov ; Maxim Buzdalov ; Benjamin Doerr
COMMENTS:       The extended version of the PPSN 2020 conference paper
HIGHLIGHT:      We start a mathematical runtime analysis for such situations.

30, TITLE:      Can we trust deep learning models diagnosis? The impact of domain shift in chest radiograph classification
http://arxiv.org/abs/1909.01940
AUTHORS:        Eduardo H. P. Pooch ; Pedro L. Ballester ; Rodrigo C. Barros
COMMENTS:       10 pages, 3 figures
HIGHLIGHT:      In this work, we evaluate the extent of domain shift on four of the largest datasets of chest radiographs.

31, TITLE:      Deep Learning for 3D Point Clouds: A Survey
http://arxiv.org/abs/1912.12033
AUTHORS:        Yulan Guo ; Hanyun Wang ; Qingyong Hu ; Hao Liu ; Li Liu ; Mohammed Bennamoun
COMMENTS:       Accepted by IEEE TPAMI. Project page: https://github.com/QingyongHu/SoTA-Point-Cloud
HIGHLIGHT:      To stimulate future research, this paper presents a comprehensive review of recent progress in deep learning methods for point clouds.

32, TITLE:      On best uniform affine approximants of convex or concave real valued functions from $\mathbb R^k$, Chebyshev equioscillation and graphics
http://arxiv.org/abs/1812.02302
AUTHORS:        Steven B. Damelin ; David L. Ragozin ; Michael Werman
COMMENTS:       In this final version, version 6.24.20 (the version to be published) we have included some new material and clarified that this paper is one on approximation theory and vision. We replace "min-max" by "best uniform" throughout. The former term was used in a previous version
HIGHLIGHT:      We study best uniform affine approximants of a continuous convex or concave function $f:\Delta\subset \mathbb R^k\xrightarrow{} \mathbb R$ where $\Delta$ is a convex compact subset of $\mathbb R^k$.

33, TITLE:      Deep Coordination Graphs
http://arxiv.org/abs/1910.00091
AUTHORS:        Wendelin Böhmer ; Vitaly Kurin ; Shimon Whiteson
COMMENTS:       Accepted at ICML 2020
HIGHLIGHT:      This paper introduces the deep coordination graph (DCG) for collaborative multi-agent reinforcement learning.

34, TITLE:      Complex Deep Learning Models for Denoising of Human Heart ECG signals
http://arxiv.org/abs/1908.10417
AUTHORS:        Corneliu Arsene
COMMENTS:       51 pages, 23 figures
HIGHLIGHT:      This paper presents several DL models namely Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM), Restricted Boltzmann Machine (RBM) together with the more conventional filtering methods (low pass filtering, high pass filtering, Notch filtering) and the standard wavelet-based technique for denoising EEG signals.

35, TITLE:      Sample Factory: Egocentric 3D Control from Pixels at 100000 FPS with Asynchronous Reinforcement Learning
http://arxiv.org/abs/2006.11751
AUTHORS:        Aleksei Petrenko ; Zhehui Huang ; Tushar Kumar ; Gaurav Sukhatme ; Vladlen Koltun
COMMENTS:       Paper published in ICML2020. Visualizations of trained policies can be found at https://sites.google.com/view/sample-factory
HIGHLIGHT:      In this work we aim to solve this problem by optimizing the efficiency and resource utilization of reinforcement learning algorithms instead of relying on distributed computation.

36, TITLE:      Improving Query Safety at Pinterest
http://arxiv.org/abs/2006.11511
AUTHORS:        Abhijit Mahabal ; Yinrui Li ; Rajat Raina ; Daniel Sun ; Revati Mahajan ; Jure Leskovec
HIGHLIGHT:      We present PinSets, a system for query-set expansion, which applies a simple yet powerful mechanism to search user sessions, expanding a tiny seed set into thousands of related queries at nearly perfect precision, deep into the tail, along with explanations that are easy to interpret.

37, TITLE:      Modeling Lost Information in Lossy Image Compression
http://arxiv.org/abs/2006.11999
AUTHORS:        Yaolong Wang ; Mingqing Xiao ; Chang Liu ; Shuxin Zheng ; Tie-Yan Liu
HIGHLIGHT:      In this work, we propose a novel invertible framework called Invertible Lossy Compression (ILC) to largely mitigate the information loss problem.

38, TITLE:      Deep Low-rank Prior in Dynamic MR Imaging
http://arxiv.org/abs/2006.12090
AUTHORS:        Ziwen Ke ; Wenqi Huang ; Jing Cheng ; Sen Jia ; Haifeng Wang ; Xin Liu ; Hairong Zheng ; Leslie Ying ; Yanjie Zhu ; Dong Liang
COMMENTS:       12 pages, 7 figures
HIGHLIGHT:      In this paper, we explore deep low-rank prior in dynamic MR imaging to obtain improved reconstruction results.

39, TITLE:      Progressive Skeletonization: Trimming more fat from a network at initialization
http://arxiv.org/abs/2006.09081
AUTHORS:        Pau de Jorge ; Amartya Sanyal ; Harkirat S. Behl ; Philip H. S. Torr ; Gregory Rogez ; Puneet K. Dokania
HIGHLIGHT:      To this end, we propose to find a skeletonized network with maximum foresight connection sensitivity (FORCE).

40, TITLE:      Pruned Neural Networks are Surprisingly Modular
http://arxiv.org/abs/2003.04881
AUTHORS:        Daniel Filan ; Shlomi Hod ; Cody Wild ; Andrew Critch ; Stuart Russell
COMMENTS:       25 pages, 12 figures
HIGHLIGHT:      To discern structure in these weights, we introduce a measurable notion of modularity for multi-layer perceptrons (MLPs), and investigate the modular structure of MLPs trained on datasets of small images.

41, TITLE:      FairRec: Two-Sided Fairness for Personalized Recommendations in Two-Sided Platforms
http://arxiv.org/abs/2002.10764
AUTHORS:        Gourab K Patro ; Arpita Biswas ; Niloy Ganguly ; Krishna P. Gummadi ; Abhijnan Chakraborty
COMMENTS:       In Proceedings of The Web Conference (WWW) 2020
HIGHLIGHT:      We investigate the problem of fair recommendation in the context of two-sided online platforms, comprising customers on one side and producers on the other.

42, TITLE:      Why Having 10,000 Parameters in Your Camera Model is Better Than Twelve
http://arxiv.org/abs/1912.02908
AUTHORS:        Thomas Schöps ; Viktor Larsson ; Marc Pollefeys ; Torsten Sattler
COMMENTS:       15 pages, 12 figures, accepted to CVPR 2020 as an oral
HIGHLIGHT:      In this paper, we argue that this should change.

43, TITLE:      Convolutional neural net face recognition works in non-human-like ways
http://arxiv.org/abs/2004.04069
AUTHORS:        P. J. B. Hancock ; R. S. Somai ; V. R. Mileva
COMMENTS:       8 pages, 2 figures. Submitted to Royal Society Open Science
HIGHLIGHT:      We tested six commercial face recognition CNNs and found that they outperform typical human participants on standard face matching tasks.

44, TITLE:      Why should we add early exits to neural networks?
http://arxiv.org/abs/2004.12814
AUTHORS:        Simone Scardapane ; Michele Scarpiniti ; Enzo Baccarelli ; Aurelio Uncini
COMMENTS:       Published in Cognitive Computation
HIGHLIGHT:      In this paper, we provide a comprehensive introduction to this family of neural networks, by describing in a unified fashion the way these architectures can be designed, trained, and actually deployed in time-constrained scenarios.

45, TITLE:      Learning Global Transparent Models from Local Contrastive Explanations
http://arxiv.org/abs/2002.08247
AUTHORS:        Tejaswini Pedapati ; Avinash Balakrishnan ; Karthikeyan Shanmugam ; Amit Dhurandhar
HIGHLIGHT:      In this work, we explore the question: Can we produce a transparent global model that is simultaneously accurate and consistent with the local (contrastive) explanations of the black-box model?

46, TITLE:      Opportunities and Challenges in Explainable Artificial Intelligence (XAI): A Survey
http://arxiv.org/abs/2006.11371
AUTHORS:        Arun Das ; Paul Rad
COMMENTS:       24 pages, 20 figures, survey paper, submitting to IEEE
HIGHLIGHT:      We start by proposing a taxonomy and categorizing the XAI techniques based on their scope of explanations, methodology behind the algorithms, and explanation level or usage which helps build trustworthy, interpretable, and self-explanatory deep learning models.

47, TITLE:      Hierarchical Patch VAE-GAN: Generating Diverse Videos from a Single Sample
http://arxiv.org/abs/2006.12226
AUTHORS:        Shir Gur ; Sagie Benaim ; Lior Wolf
HIGHLIGHT:      We introduce a novel patch-based variational autoencoder (VAE) which allows for a much greater diversity in generation.

48, TITLE:      Condensing Two-stage Detection with Automatic Object Key Part Discovery
http://arxiv.org/abs/2006.05597
AUTHORS:        Zhe Chen ; Jing Zhang ; Dacheng Tao
HIGHLIGHT:      To address this problem, we propose that the model parameters of two-stage detection heads can be condensed and reduced by concentrating on object key parts.

49, TITLE:      Recent Advances and Challenges in Task-oriented Dialog System
http://arxiv.org/abs/2003.07490
AUTHORS:        Zheng Zhang ; Ryuichi Takanobu ; Qi Zhu ; Minlie Huang ; Xiaoyan Zhu
COMMENTS:       Under review of SCIENCE CHINA Technological Science (SCTS)
HIGHLIGHT:      In this paper, we survey recent advances and challenges in task-oriented dialog systems.

50, TITLE:      Exponential Upper Bounds for the Runtime of Randomized Search Heuristics
http://arxiv.org/abs/2004.05733
AUTHORS:        Benjamin Doerr
COMMENTS:       Extended version of a paper appearing at PPSN2020
HIGHLIGHT:      We argue that proven exponential upper bounds on runtimes, an established area in classic algorithms, are interesting also in heuristic search and we prove several such results.

51, TITLE:      Robust Generative Restricted Kernel Machines using Weighted Conjugate Feature Duality
http://arxiv.org/abs/2002.01180
AUTHORS:        Arun Pandey ; Joachim Schreurs ; Johan A. K. Suykens
HIGHLIGHT:      In this paper, we introduce weighted conjugate feature duality in the framework of Restricted Kernel Machines (RKMs).

52, TITLE:      iSeeBetter: Spatio-temporal video super-resolution using recurrent generative back-projection networks
http://arxiv.org/abs/2006.11161
AUTHORS:        Aman Chadha ; John Britto ; M. Mani Roja
COMMENTS:       11 pages, 6 figures, 4 tables
HIGHLIGHT:      We present iSeeBetter, a novel GAN-based spatio-temporal approach to video super-resolution (VSR) that renders temporally consistent super-resolution videos.

53, TITLE:      The PSPACE-hardness of understanding neural circuits
http://arxiv.org/abs/2006.08266
AUTHORS:        Vidya Sagar Sharma ; Piyush Srivastava
COMMENTS:       2 figures
HIGHLIGHT:      In this paper, we prove that the problems of finding minimal or minimum-size degenerate sets, and of finding the set of vital neurons, of a neural circuit given as input, are in fact PSPACE-hard.

54, TITLE:      TOMA: Topological Map Abstraction for Reinforcement Learning
http://arxiv.org/abs/2005.06061
AUTHORS:        Zhao-Heng Yin ; Wu-Jun Li
HIGHLIGHT:      In this paper, we propose a new method, called topological map abstraction (TOMA), for graph generation.

55, TITLE:      Pruning untrained neural networks: Principles and Analysis
http://arxiv.org/abs/2002.08797
AUTHORS:        Soufiane Hayou ; Jean-Francois Ton ; Arnaud Doucet ; Yee Whye Teh
COMMENTS:       31 pages, 12 figures
HIGHLIGHT:      In this paper we provide a comprehensive theoretical analysis of pruning at initialization and training of sparse architectures.

56, TITLE:      Simplification of Polyline Bundles
http://arxiv.org/abs/1907.05296
AUTHORS:        Joachim Spoerhase ; Sabine Storandt ; Johannes Zink
HIGHLIGHT:      We propose and study a generalization to the well-known problem of polyline simplification.

57, TITLE:      SoccerDB: A Large-Scale Database for Comprehensive Video Understanding
http://arxiv.org/abs/1912.04465
AUTHORS:        Yudong Jiang ; Kaixu Cui ; Leilei Chen ; Canjin Wang ; Changliang Xu
COMMENTS:       pre-print draft version
HIGHLIGHT:      In this paper, we propose a new soccer video database named SoccerDB, comprising 171,192 video segments from 346 high-quality soccer games.

58, TITLE:      BEV-Seg: Bird's Eye View Semantic Segmentation Using Geometry and Semantic Point Cloud
http://arxiv.org/abs/2006.11436
AUTHORS:        Mong H. Ng ; Kaahan Radia ; Jianfei Chen ; Dequan Wang ; Ionel Gog ; Joseph E. Gonzalez
COMMENTS:       Accepted into CVPR 2020 Workshop Scalability in Autonomous Driving by Waymo
HIGHLIGHT:      In this work, we focus on bird's eye semantic segmentation, a task that predicts pixel-wise semantic segmentation in BEV from side RGB images.

59, TITLE:      Deployment-Efficient Reinforcement Learning via Model-Based Offline Optimization
http://arxiv.org/abs/2006.03647
AUTHORS:        Tatsuya Matsushima ; Hiroki Furuta ; Yutaka Matsuo ; Ofir Nachum ; Shixiang Gu
HIGHLIGHT:      We propose a novel model-based algorithm, Behavior-Regularized Model-ENsemble (BREMEN) that can effectively optimize a policy offline using 10-20 times fewer data than prior works.

60, TITLE:      Wasserstein-2 Generative Networks
http://arxiv.org/abs/1909.13082
AUTHORS:        Alexander Korotin ; Vage Egiazarian ; Arip Asadulaev ; Alexander Safin ; Evgeny Burnaev
COMMENTS:       29 pages, 21 figures, 3 tables
HIGHLIGHT:      In this paper, we propose a novel end-to-end algorithm for training generative models which uses a non-minimax objective simplifying model training.

61, TITLE:      MCENET: Multi-Context Encoder Network for Homogeneous Agent Trajectory Prediction in Mixed Traffic
http://arxiv.org/abs/2002.05966
AUTHORS:        Hao Cheng ; Wentong Liao ; Michael Ying Yang ; Monika Sester ; Bodo Rosenhahn
COMMENTS:       8 pages, 5 figures, code is available on https://github.com/haohao11/MCENET
HIGHLIGHT:      To this end, we propose an approach named Multi-Context Encoder Network (MCENET) that is trained by encoding both past and future scene context, interaction context and motion information to capture the patterns and variations of the future trajectories using a set of stochastic latent variables.

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

63, TITLE:      Automatic Validation of Textual Attribute Values in E-commerce Catalog by Learning with Limited Labeled Data
http://arxiv.org/abs/2006.08779
AUTHORS:        Yaqing Wang ; Yifan Ethan Xu ; Xian Li ; Xin Luna Dong ; Jing Gao
COMMENTS:       KDD 2020
HIGHLIGHT:      To address the aforementioned challenges, we propose a novel meta-learning latent variable approach, called MetaBridge, which can learn transferable knowledge from a subset of categories with limited labeled data and capture the uncertainty of never-seen categories with unlabeled data.

64, TITLE:      Simulating Anisoplanatic Turbulence by Sampling Inter-modal and Spatially Correlated Zernike Coefficients
http://arxiv.org/abs/2004.11210
AUTHORS:        Nicholas Chimitt ; Stanley H. Chan
HIGHLIGHT:      In this paper, we present a propagation-free method for simulating imaging through turbulence.

65, TITLE:      Visor: Privacy-Preserving Video Analytics as a Cloud Service
http://arxiv.org/abs/2006.09628
AUTHORS:        Rishabh Poddar ; Ganesh Ananthanarayanan ; Srinath Setty ; Stavros Volos ; Raluca Ada Popa
COMMENTS:       USENIX Security 2020
HIGHLIGHT:      We present Visor, a system that provides confidentiality for the user's video stream as well as the ML models in the presence of a compromised cloud platform and untrusted co-tenants.

66, TITLE:      Robust Face Verification via Disentangled Representations
http://arxiv.org/abs/2006.03638
AUTHORS:        Marius Arvinte ; Ahmed H. Tewfik ; Sriram Vishwanath
COMMENTS:       Preprint
HIGHLIGHT:      We introduce a robust algorithm for face verification, i.e., deciding whether twoimages are of the same person or not.

67, TITLE:      PICK: Processing Key Information Extraction from Documents using Improved Graph Learning-Convolutional Networks
http://arxiv.org/abs/2004.07464
AUTHORS:        Wenwen Yu ; Ning Lu ; Xianbiao Qi ; Ping Gong ; Rong Xiao
COMMENTS:       Accepted in the 25th International Conference on Pattern Recognition (ICPR 2020), Milan, Italy
HIGHLIGHT:      In this paper, we introduce PICK, a framework that is effective and robust in handling complex documents layout for KIE by combining graph learning with graph convolution operation, yielding a richer semantic representation containing the textual and visual features and global layout without ambiguity.

68, TITLE:      Consolidating Commonsense Knowledge
http://arxiv.org/abs/2006.06114
AUTHORS:        Filip Ilievski ; Pedro Szekely ; Jingwei Cheng ; Fu Zhang ; Ehsan Qasemi
COMMENTS:       14 pages
HIGHLIGHT:      In this paper, we list representative sources and their properties.

69, TITLE:      Reinforcement Learning Upside Down: Don't Predict Rewards -- Just Map Them to Actions
http://arxiv.org/abs/1912.02875
AUTHORS:        Juergen Schmidhuber
COMMENTS:       22 pages, 81 references
HIGHLIGHT:      A separate paper [63] on first experiments with UDRL shows that even a pilot version of UDRL can outperform traditional baseline algorithms on certain challenging RL problems.

70, TITLE:      Neural Execution Engines: Learning to Execute Subroutines
http://arxiv.org/abs/2006.08084
AUTHORS:        Yujun Yan ; Kevin Swersky ; Danai Koutra ; Parthasarathy Ranganathan ; Milad Hashemi
HIGHLIGHT:      To address the issue, we propose a learned conditional masking mechanism, which enables the model to strongly generalize far outside of its training range with near-perfect accuracy on a variety of algorithms.

71, TITLE:      A Further Study of Unsupervised Pre-training for Transformer Based Speech Recognition
http://arxiv.org/abs/2005.09862
AUTHORS:        Dongwei Jiang ; Wubo Li ; Ruixiong Zhang ; Miao Cao ; Ne Luo ; Yang Han ; Wei Zou ; Xiangang Li
HIGHLIGHT:      In this paper, we conduct a further study on MPC and focus on three important aspects: the effect of pre-training data speaking style, its extension on streaming model, and how to better transfer learned knowledge from pre-training stage to downstream tasks.

72, TITLE:      Data Efficient Stagewise Knowledge Distillation
http://arxiv.org/abs/1911.06786
AUTHORS:        Akshay Kulkarni ; Navid Panchi ; Sharath Chandra Raparthy ; Shital Chiddarwar
COMMENTS:       15 pages, 1 figure, 6 tables and 1 algorithm
HIGHLIGHT:      In this work, we propose a new method called Stagewise Knowledge Distillation (SKD) which builds on traditional KD methods by progressive stagewise training to leverage the knowledge gained from the teacher, resulting in data-efficient distillation process.

73, TITLE:      Envy-freeness up to one item: Shall we add or remove resources?
http://arxiv.org/abs/2006.11312
AUTHORS:        Martin Aleksandrov
COMMENTS:       10 pages, 1 table, 2 figures, v1 is a working version, v2 is the polished version
HIGHLIGHT:      We propose two new axiomatic properties for allocations in this model: EF1+- and EFX+-.

74, TITLE:      Towards vision-based robotic skins: a data-driven, multi-camera tactile sensor
http://arxiv.org/abs/1910.14526
AUTHORS:        Camill Trueeb ; Carmelo Sferrazza ; Raffaello D'Andrea
COMMENTS:       Accompanying video: https://youtu.be/lbavqAlKl98
HIGHLIGHT:      The design proposed in this paper exhibits a larger contact surface and a thinner structure than most of the existing camera-based tactile sensors, without the use of additional reflecting components such as mirrors.

75, TITLE:      Non-linear aggregation of filters to improve image denoising
http://arxiv.org/abs/1904.00865
AUTHORS:        Benjamin Guedj ; Juliette Rengot
COMMENTS:       To appear at Computing Conference 2020
HIGHLIGHT:      We introduce a novel aggregation method to efficiently perform image denoising.

76, TITLE:      Representation Learning with Statistical Independence to Mitigate Bias
http://arxiv.org/abs/1910.03676
AUTHORS:        Ehsan Adeli ; Qingyu Zhao ; Adolf Pfefferbaum ; Edith V. Sullivan ; Li Fei-Fei ; Juan Carlos Niebles ; Kilian M. Pohl
HIGHLIGHT:      In this paper, we propose such a model based on adversarial training with two competing objectives to learn features that have (1) maximum discriminative power with respect to the task and (2) minimal statistical mean dependence with the protected (bias) variable(s).

77, TITLE:      Train Large, Then Compress: Rethinking Model Size for Efficient Training and Inference of Transformers
http://arxiv.org/abs/2002.11794
AUTHORS:        Zhuohan Li ; Eric Wallace ; Sheng Shen ; Kevin Lin ; Kurt Keutzer ; Dan Klein ; Joseph E. Gonzalez
COMMENTS:       ICML 2020
HIGHLIGHT:      We study the impact of model size in this setting, focusing on Transformer models for NLP tasks that are limited by compute: self-supervised pretraining and high-resource machine translation.

78, TITLE:      Learning to Fly via Deep Model-Based Reinforcement Learning
http://arxiv.org/abs/2003.08876
AUTHORS:        Philip Becker-Ehmck ; Maximilian Karl ; Jan Peters ; Patrick van der Smagt
HIGHLIGHT:      In this work, by leveraging a learnt probabilistic model of drone dynamics, we learn a thrust-attitude controller for a quadrotor through model-based reinforcement learning.