Week Ending 08.18.19
RESEARCH WATCH: 08.18.19
Over the past week, 585 new papers were published in "Computer Science".
The paper discussed most in the news over the past week was by a team at UC Berkeley: "Natural Adversarial Examples" by Dan Hendrycks et al (Jul 2019), which was referenced 20 times, including in the article University Research Teams Open-Source Natural Adversarial Image DataSet for Computer-Vision AI in InfoQ. The paper author, Daniel Hendrycks, was quoted saying "inputs to machine learning models that an attacker has intentionally designed to cause the model to make a mistake". The paper got social media traction with 541 shares. A user, @DanHendrycks, tweeted "Natural Adversarial Examples are real-world and unmodified examples which cause classifiers to be consistently confused. The new dataset has 7,500 images, which we personally labeled over several months. Paper: Dataset and code".
Leading researcher Kyunghyun Cho (New York University) came out with "Neural Text Generation with Unlikelihood Training" @evolvingstuff tweeted "Neural Text Generation with Unlikelihood Training "We propose a new objective, unlikelihood training, which forces unlikely generations to be assigned lower probability by the model."".
The paper shared the most on social media this week is "Einconv: Exploring Unexplored Tensor Decompositions for Convolutional Neural Networks" by Kohei Hayashi et al (Aug 2019) with 190 shares.
Over the past week, 41 new papers were published in "Computer Science - Artificial Intelligence".
The paper discussed most in the news over the past week was by a team at DeepMind: "Behaviour Suite for Reinforcement Learning" by Ian Osband et al (Aug 2019), which was referenced 1 time, including in the article DeepMind ‘Bsuite’ Evaluates Reinforcement Learning Agents in SyncedReview.com. The paper also got the most social media traction with 565 shares. The authors introduce the Behaviour Suite for Reinforcement Learning, or bsuite for short. On Twitter, @MasterScrat commented "The intro from DeepMind's latest paper is highly relevant - can't be a coincidence, published a few days after their losses were announced".
Leading researcher Aaron Courville (Université de Montréal) published "VideoNavQA: Bridging the Gap between Visual and Embodied Question Answering".
The paper shared the most on social media this week is by a team at Australian National University: "Reward Tampering Problems and Solutions in Reinforcement Learning: A Causal Influence Diagram Perspective" by Tom Everitt et al (Aug 2019) with 164 shares. @v_maini (Vishal Maini) tweeted "another step towards developing a set of best practices for designing safe RL agents - in this case, by avoiding incentives for agents to tamper with their own reward function. great work, and team 🚀 🤖 ✅".
Over the past week, 150 new papers were published in "Computer Science - Computer Vision and Pattern Recognition".
The paper discussed most in the news over the past week was by a team at UC Berkeley: "Natural Adversarial Examples" by Dan Hendrycks et al (Jul 2019)
Leading researcher Aaron Courville (Université de Montréal) came out with "VideoNavQA: Bridging the Gap between Visual and Embodied Question Answering".
The paper shared the most on social media this week is by a team at University of Oxford: "DynaNet: Neural Kalman Dynamical Model for Motion Estimation and Prediction" by Changhao Chen et al (Aug 2019) with 158 shares.
Over the past week, nine new papers were published in "Computer Science - Computers and Society".
The paper discussed most in the news over the past week was "Anti-Money Laundering in Bitcoin: Experimenting with Graph Convolutional Networks for Financial Forensics"by Mark Weber et al (Jul 2019), which was referenced 1 time, including in the article The Elliptic Data Set - working with the community to combat financial crime in cryptocurrencies in Elliptic.co. The paper got social media traction with 14 shares. The authors motivate the opportunity to reconcile the cause of safety with that of financial inclusion.
Over the past week, 15 new papers were published in "Computer Science - Human-Computer Interaction".
The paper discussed most in the news over the past week was "Automated Corrosion Detection Using Crowd Sourced Training for Deep Learning" by W. T. Nash et al (Aug 2019), which was referenced 1 time, including in the article How State Politics Is Playing a Huge Role in Artificial Intelligence: Eye on A.I. in Fortune. The paper was shared 2 times in social media.
This week was active for "Computer Science - Learning", with 235 new papers.
The paper discussed most in the news over the past week was by a team at UC Berkeley: "Natural Adversarial Examples" by Dan Hendrycks et al (Jul 2019)
Leading researcher Kyunghyun Cho (New York University) came out with "Neural Text Generation with Unlikelihood Training" @evolvingstuff tweeted "Neural Text Generation with Unlikelihood Training "We propose a new objective, unlikelihood training, which forces unlikely generations to be assigned lower probability by the model."".
The paper shared the most on social media this week is "Einconv: Exploring Unexplored Tensor Decompositions for Convolutional Neural Networks" by Kohei Hayashi et al (Aug 2019)
Over the past week, 13 new papers were published in "Computer Science - Multiagent Systems".
The paper discussed most in the news over the past week was by a team at Rochester Institute of Technology: "A Framework for Monitoring Human Physiological Response during Human Robot Collaborative Task" by Celal Savur et al (Jul 2019), which was referenced 1 time, including in the article Monitoring human physiological responses to improve interactions with robots in PhysOrg.com. The paper author, Celal Savur, was quoted saying "In order to do this, a framework for a system that represents and records the robot motion and human physiological state concurrently was needed". The paper was shared 1 time in social media.
Over the past week, 18 new papers were published in "Computer Science - Neural and Evolutionary Computing".
The paper shared the most on social media this week is by a team at University of Liège: "Unconstrained Monotonic Neural Networks" by Antoine Wehenkel et al (Aug 2019) with 82 shares. @vincesitzmann (Vincent Sitzmann) tweeted "Cool work - I have never seen this way of integrating over the outputs of a neural net before, seems like this is generally a great trick to have in one’s repertoire!".
Over the past week, 33 new papers were published in "Computer Science - Robotics".
The paper discussed most in the news over the past week was "Learning to Solve a Rubiks Cube with a Dexterous Hand" by Tingguang Li et al (Jul 2019), which was referenced 4 times, including in the article Solving a Rubik's Cube with a dexterous hand in PhysOrg.com. The paper author, Tingguang Li (The Chinese University of Hong Kong), was quoted saying "Seeing some researchers use multi-fingered robot hands for tasks like reposing an object and manipulating a tool , we considered whether we could utilize a robot hand for more complicated tasks, such as solving a Rubik's Cube". The paper got social media traction with 8 shares.
The paper shared the most on social media this week is by a team at University of Oxford: "DynaNet: Neural Kalman Dynamical Model for Motion Estimation and Prediction" by Changhao Chen et al (Aug 2019)