Week Ending 11.15.2020
RESEARCH WATCH: 11.15.2020
This week was active for "Computer Science", with 1,205 new papers.
The paper discussed most in the news over the past week was "How Did That Get In My Phone? Unwanted App Distribution on Android Devices" by Platon Kotzias et al (Oct 2020), which was referenced 16 times, including in the article Play Store identified as main distribution vector for most Android malware in ZDNet. The paper was shared 4 times in social media. On Twitter, @Michal_Jarski observed "Recent research shows 67% of the malicious app installs come from the official Google Play Store".
Leading researcher Kyunghyun Cho (New York University) came out with "Learned Equivariant Rendering without Transformation Supervision".
The paper shared the most on social media this week is by a team at Google: "Long Range Arena: A Benchmark for Efficient Transformers" by Yi Tay et al (Nov 2020) with 101 shares.
This week was very active for "Computer Science - Artificial Intelligence", with 178 new papers.
The paper discussed most in the news over the past week was by a team at University of North Carolina at Chapel Hill: "Vokenization: Improving Language Understanding with Contextualized, Visual-Grounded Supervision" by Hao Tan et al (Oct 2020), which was referenced 2 times, including in the article Vokenizing: A new way to give AI language models much-needed common sense in BiometricUpdate.com. The paper got social media traction with 254 shares. On Twitter, @andre_niyongabo posted "Vokenization🎉🎉🎉This is very cool!!!", while @mohitban47 said ""Vokens" = Visually-grounded-tokens (contextual) to imprv lang-pretraining & engl NLU tasks (imp divergence/grounding ratio issues, extrapolates frm small dataset)! pdf: Full code: ➡️Hao is on job market🙂".
Leading researcher Sergey Levine (University of California, Berkeley) published "Reinforcement Learning with Videos: Combining Offline Observations with Interaction" The investigators consider the question : can they perform reinforcement learning directly on experience collected by humans?. @gastronomy tweeted "> Reinforcement learning is a powerful framework for robots to acquire skills from experience, but often requires a substantial a".
The paper shared the most on social media this week is by a team at Google: "Long Range Arena: A Benchmark for Efficient Transformers" by Yi Tay et al (Nov 2020)
This week was active for "Computer Science - Computer Vision and Pattern Recognition", with 229 new papers.
The paper discussed most in the news over the past week was by a team at University of Texas at San Antonio: "Zoom on the Keystrokes: Exploiting Video Calls for Keystroke Inference Attacks" by Mohd Sabra et al (Oct 2020), which was referenced 8 times, including in the article Study: Tiny Visual Cues Can Give Hackers Your Password On Video Calls in MyTechDecisions. The paper author, Murtuza Jadliwala (University of Texas at San Antonio), was quoted saying "For someone to carry out this attack [today], they’d need a lot of experience and expertise". The paper got social media traction with 27 shares. The authors design and evaluate an attack framework to infer one type of such private information from the video stream of a call -- keystrokes, i.e., text typed during the call. On Twitter, @paultaylorbrain posted "Crazy. Guessing what you type during Zoom from arm movements. Accuracy 18% at guessing passwords (already slightly scary) goes up beyond 90% at saying whether or not you are typing something specific like your username. Harder if wearing sleeves!".
Leading researcher Kyunghyun Cho (New York University) published "Learned Equivariant Rendering without Transformation Supervision".
The paper shared the most on social media this week is by a team at Google: "Long Range Arena: A Benchmark for Efficient Transformers" by Yi Tay et al (Nov 2020)
Over the past week, 24 new papers were published in "Computer Science - Computers and Society".
The paper discussed most in the news over the past week was "The De-democratization of AI: Deep Learning and the Compute Divide in Artificial Intelligence Research" by Nur Ahmed et al (Oct 2020), which was referenced 3 times, including in the article AI research finds a ‘compute divide’ concentrates power and accelerates inequality in the era of deep learning in Venturebeat. The paper got social media traction with 243 shares. A user, @hyounpark, tweeted "Great thread by on the need for bias awareness in AI research. The more we learn about #AI, the more we realize that it reifies status quo hierarchies without active governance", while @joftius posted "This paper is about a really important problem. One cause it mentions that I would emphasize more: the role of proprietary data. Most of the value of "AI" comes from the users and content moderators on massive platforms that generate/curate data".
Over the past week, 20 new papers were published in "Computer Science - Human-Computer Interaction".
The paper discussed most in the news over the past week was by a team at University of Texas at San Antonio: "Zoom on the Keystrokes: Exploiting Video Calls for Keystroke Inference Attacks" by Mohd Sabra et al (Oct 2020)
This week was extremely active for "Computer Science - Learning", with 453 new papers.
The paper discussed most in the news over the past week was by a team at University of Waterloo: "Less Than One-Shot Learning: Learning N Classes From MThat's so cool! Research "Less than one-shot learning can teach a model to identify more objects than the number of examples it is trained on." Article: Paper".
Leading researcher Sergey Levine (University of California, Berkeley) published "Continual Learning of Control Primitives: Skill Discovery via Reset-Games" @iandanforth tweeted "Ok I didn't get this at first. "reset skills" means "the skills necessary to reset the environment to a point where you can attempt the primary task." So if you fail you need to be able to clean up after yourself to try again. Makes a lot of sense!".
The paper shared the most on social media this week is by a team at Google: "Long Range Arena: A Benchmark for Efficient Transformers" by Yi Tay et al (Nov 2020)
Over the past week, ten new papers were published in "Computer Science - Multiagent Systems".
The paper discussed most in the news over the past week was "HAMLET: A Hierarchical Agent-based Machine Learning Platform" by Ahmad Esmaeili et al (Oct 2020), which was referenced 1 time, including in the article HAMLET: A platform to simplify AI research and development in Tech Xplore. The paper author, Ahmad Esmaeili (Researchers), was quoted saying "Organizing and keeping track of the machine learning algorithms and datasets has always been a major challenge for us, as well for as many other researchers in the field". The paper got social media traction with 14 shares. The researchers introduce HAMLET (Hierarchical Agent - based Machine LEarning plaTform), a platform based on hierarchical multi - agent systems, to facilitate the research and democratization of machine learning entities distributed geographically or locally. On Twitter, @cdossman said "HAMLET: A hierarchical agent-based #machine_learning platform Read more", while @arXiv__ml observed "#machinelearning Hierarchical Multi-Agent Systems provide a convenient and relevant way to analyze, model, and simulate complex systems in which a".
Over the past week, 26 new papers were published in "Computer Science - Neural and Evolutionary Computing".
The paper discussed most in the news over the past week was by a team at Google: "Reverse engineering learned optimizers reveals known and novel mechanisms" by Niru Maheswaranathan et al (Nov 2020), which was referenced 3 times, including in the article Google Brain Paper Demystifies Learned Optimizers in SyncedReview.com. The paper got social media traction with 9 shares. On Twitter, @memming commented "Loved the talk by on interpreting learned optimization!", while @permutans said "Nice talk from Niru on dynamical systems interpretation of learned optimisers".
This week was extremely active for "Computer Science - Robotics", with 156 new papers.
The paper discussed most in the news over the past week was by a team at Politecnico di Torino Interdepartmental Centre: "DeepWay: a Deep Learning Estimator for Unmanned Ground Vehicle Global Path Planning" by Vittorio Mazzia et al (Oct 2020), which was referenced 2 times, including in the article Deep-Way: A Neural Network Architecture for Unmanned Ground Vehicle Path Planning — A Review in Medium.com. The paper got social media traction with 8 shares.
Leading researcher Sergey Levine (University of California, Berkeley) came out with "Reinforcement Learning with Videos: Combining Offline Observations with Interaction" The investigators consider the question : can they perform reinforcement learning directly on experience collected by humans?. @gastronomy tweeted "> Reinforcement learning is a powerful framework for robots to acquire skills from experience, but often requires a substantial a".