Week Ending 10.11.2020
RESEARCH WATCH: 10.11.2020
This week was extremely active for "Computer Science - Artificial Intelligence", with 256 new papers.
The paper discussed most in the news over the past week was by a team at DeepMind: "Assessing Game Balance with AlphaZero: Exploring Alternative Rule Sets in Chess" by Nenad Tomašev et al (Sep 2020), which was referenced 10 times, including in the article AlphaZero/Kramnik: Torpedo pawns in ChessBase. The paper author, Vladimir Kramnik, was quoted saying "For quite a number of games on the highest level, half of the game—sometimes a full game—is played out of memory. You don't even play your own preparation; you play your computer's preparation." The paper got social media traction with 299 shares. The investigators use AlphaZero to creatively explore and design new chess variants. A user, @debarghya_das, tweeted "1/5 This chess paper from DeepMind and has absolutely consumed my mind in the last few days. They answered a question many chess players have dreamed of - how fair is chess? If you change the rules, does that change?".
Leading researcher Yoshua Bengio (Université de Montréal) came out with "RNNLogic: Learning Logic Rules for Reasoning on Knowledge Graphs".
The paper shared the most on social media this week is by a team at University of California, Davis: "Energy-based Out-of-distribution Detection" by Weitang Liu et al (Oct 2020) with 335 shares. @gastronomy (Tanat Tonguthaisri) tweeted "> Determining whether inputs are out-of-distribution (OOD) is an essential building block for safely deploying machine learning models in the open world. However, previou".
This week was active for "Computer Science - Computer Vision and Pattern Recognition", with 241 new papers.
The paper discussed most in the news over the past week was by a team at Massachusetts Institute of Technology: "Joint Modeling of Chest Radiographs and Radiology Reports for Pulmonary Edema Assessment" by Geeticka Chauhan et al (Aug 2020), which was referenced 11 times, including in the article MIT Develops AI Tool That Warns of Heart Attack by Reading Lung X-Rays in IBTimes Singapore. The paper author, Geeticka Chauhan (Massachusetts Institute of Technology), was quoted saying "Our model can turn both images and text into compact numerical abstractions from which an interpretation can be derived. We trained it to minimize the difference between the representations of the X-ray images and the text of the radiology reports, using the reports to improve the image interpretation". The paper got social media traction with 10 shares. A Twitter user, @rayruizhiliao, said "and I are presenting this work at in Oct, jointly with You can check out our preprint at Our code is also available at".
Leading researcher Pieter Abbeel (UC Berkeley) came out with "LaND: Learning to Navigate from Disengagements".
The paper shared the most on social media this week is by a team at Google: "Gradient Flow in Sparse Neural Networks and How Lottery Tickets Win" by Utku Evci et al (Oct 2020) with 65 shares. The authors attempt to answer : (1) why training unstructured sparse networks from random initialization performs poorly and; (2) what makes LTs and DST the exceptions?. @ml_collective (ML Collective) tweeted "Congrats on publishing! 🥳*This* and RigL will be discussed in details in a few hours at our weekly reading group ".
This week was active for "Computer Science - Computers and Society", with 33 new papers.
The paper discussed most in the news over the past week was "The Grey Hoodie Project: Big Tobacco, Big Tech, and the threat on academic integrity" by Mohamed Abdalla et al (Sep 2020), which was referenced 2 times, including in the article Many Top AI Researchers Get Financial Backing From Big Tech in Wired News. The paper author, Mohamed Abdalla, was quoted saying "There are very few people that don't have some sort of connection to Big Tech". The paper got social media traction with 31 shares. A Twitter user, @VaroonMathur, observed "Re: the WIRED article on AI Research - the original paper really hones in on the parallels of Big Tech and Big Tobacco and those comparisons are...interesting for sure. I'd say there is at least one key difference that complicates this view though".
The paper shared the most on social media this week is "Metrics and methods for a systematic comparison of fairness-aware machine learning algorithms" by Gareth P. Jones et al (Oct 2020) with 68 shares.
This week was very active for "Computer Science - Human-Computer Interaction", with 55 new papers.
This week was extremely active for "Computer Science - Learning", with 557 new papers.
The paper discussed most in the news over the past week was "Its Not Just Size That Matters: Small Language Models Are Also Few-Shot Learners" by Timo Schick et al (Sep 2020), which was referenced 4 times, including in the article AI Training Method Exceeds GPT-3 Performance with 99.9% Fewer Parameters in InfoQ. The paper got social media traction with 354 shares. The authors show that performance similar to GPT-3 can be obtained with language models whose parameter count is several orders of magnitude smaller. On Twitter, @timo_schick commented "🎉 New paper 🎉 We show that language models are few-shot learners even if they have far less than 175B parameters. Our method performs similar to GPT-3 on SuperGLUE after training on 32 examples with just 0.1% of its parameter count: #NLProc".
Leading researcher Yoshua Bengio (Université de Montréal) came out with "CausalWorld: A Robotic Manipulation Benchmark for Causal Structure and Transfer Learning".
The paper shared the most on social media this week is by a team at University of California, Davis: "Energy-based Out-of-distribution Detection" by Weitang Liu et al (Oct 2020)
Over the past week, ten new papers were published in "Computer Science - Multiagent Systems".
Over the past week, 35 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 Massachusetts Institute of Technology: "Conditional Generative Adversarial Networks to Model Urban Outdoor Air Pollution" by Jamal Toutouh (Oct 2020) with 63 shares.
This week was active for "Computer Science - Robotics", with 48 new papers.
The paper discussed most in the news over the past week was by a team at Massachusetts Institute of Technology: "Integrated Benchmarking and Design for Reproducible and Accessible Evaluation of Robotic Agents" by Jacopo Tani et al (Sep 2020), which was referenced 4 times, including in the article Hands-On AI: Duckietown Foundation Offering Free edX Robotics Course Powered by NVIDIA Jetson Nano 2GB in NVIDIA Newsroom. The paper author, Emilio Frazzoli (Massachusetts Institute of Technology), was quoted saying "The Duckietown educational platform provides a hands-on, scaled down, accessible version of real world autonomous systems". The paper got social media traction with 24 shares. The investigators describe a new concept for reproducible robotics research that integrates development and benchmarking, so that reproducibility is obtained by design from the beginning of the research/development processes. A user, @MontrealRobots, tweeted "Check out new work from our group, to be presented at IROS 2020. With "Integrated benchmarking and design for benchmarking and accessible evaluation of robotic agents"".
Leading researcher Yoshua Bengio (Université de Montréal) published "CausalWorld: A Robotic Manipulation Benchmark for Causal Structure and Transfer Learning".