Week Ending 12.26.2021
RESEARCH WATCH: 12.26.2021
This week was active for "Computer Science - Artificial Intelligence", with 111 new papers.
The paper discussed most in the news over the past week was by a team at Massachusetts Institute of Technology: "3DP3: 3D Scene Perception via Probabilistic Programming" by Nishad Gothoskar et al (Oct 2021), which was referenced 12 times, including in the article Probabilistically Programmed Systems See More Like Humans Do in Photonics. The paper author, Vikash K. Mansinghka (Massachusetts Institute of Technology), was quoted saying "I found it surprising to see how large the errors from deep learning could sometimes be – producing scene representations where objects really didn’t match with what people would perceive. I also found it surprising that only a little bit of model-based inference in our causal probabilistic program was enough to detect and fix these errors. Of course, there is still a long way to go to make it fast and robust enough for challenging real-time vision systems – but for the first time, we’re seeing probabilistic programming and structured causal models improving robustness over deep learning on hard 3D vision benchmarks". The paper was shared 3 times in social media. A Twitter user, @summarizedml, commented "3DP3, a framework for inverse graphics that uses inference in a structured generative model of objects, scenes, and images. 📄".
Leading researcher Sergey Levine (University of California, Berkeley) published "RvS: What is Essential for Offline RL via Supervised Learning?" @HochreiterSepp tweeted "ArXiv Experimental study of using supervised learning for offline RL. Maximizing likelihood with a two-layer MLP is competitive with TD and transformers for sequences".
The paper shared the most on social media this week is by a team at Google: "Few-shot Learning with Multilingual Language Models" by Xi Victoria Lin et al (Dec 2021) with 131 shares. The researchers train multilingual autoregressive language models on a balanced corpus covering a diverse set of languages, and study their few- and zero - shot learning capabilities in a wide range of tasks. @ak92501 (AK) tweeted "Few-shot Learning with Multilingual Language Models abs: 7.5B parameter model, sota in few-shot learning in more than 20 representative languages, outperforming GPT-3 of comparable size in multilingual commonsense reasoning and natural language inference".
The most influential Twitter user discussing papers is 彩恵りり🧚♀️サイエンスライター✨おしごと募集中 who shared "Pokemon: Protected Logic Qubit Derived from the 0-$\pi$ Qubit" by J. Q. You et al (Dec 2021) and said: "Pokemon量子ビット...🤔 Pokemon: Protected Logic Qubit Derived from the 0-π Qubit".
This week was active for "Computer Science - Computer Vision and Pattern Recognition", with 232 new papers.
The paper discussed most in the news over the past week was by a team at Massachusetts Institute of Technology: "3DP3: 3D Scene Perception via Probabilistic Programming" by Nishad Gothoskar et al (Oct 2021)
Leading researcher Ilya Sutskever (OpenAI) published "GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models", which had 30 shares over the past 5 days. @aicrumb tweeted "1000 steps/model (base/upscale) "steampunk style oil painting of a mountain with canons poking out of it's sides"". This paper was also shared the most on social media with 1,877 tweets. @aicrumb (crumb) tweeted "1000 steps/model (base/upscale) "steampunk style oil painting of a mountain with canons poking out of it's sides"".
Over the past week, 21 new papers were published in "Computer Science - Computers and Society".
This week was active for "Computer Science - Human-Computer Interaction", with 25 new papers.
The paper discussed most in the news over the past week was by a team at Institute of Information Security: "Phishing in Organizations: Findings from a Large-Scale and Long-Term Study" by Daniele Lain et al (Dec 2021), which was referenced 5 times, including in the article Research: Simulated Phishing Tests Make Organizations Less Secure in Security Week. The paper also got the most social media traction with 329 shares. The authors present findings from a large - scale and long - term phishing experiment that they have conducted in collaboration with a partner company. A Twitter user, @SrdjanCapkun, observed "15 months, more than 14,000 participants How susceptible are employees of a large organization to phishing? Does training help? Can employees collectively detect phishing? Answers in this great work by (to appear in IEEE S&P 2022)".
This week was very active for "Computer Science - Learning", with 301 new papers.
The paper discussed most in the news over the past week was by a team at DeepMind: "Player of Games" by Martin Schmid et al (Dec 2021), which was referenced 6 times, including in the article Interesting Innovations From DeepMind In 2021 in Analytics India Magazine. The paper author, Schmid, was quoted saying "[O]ne would expect that the applications that benefited from AlphaZero might also benefit from Player of Games". The paper got social media traction with 392 shares. A user, @PatrickPilarski, tweeted "Excellent new research from our Edmonton DeepMind office, showing an agent that can learn to skillfully engage in perfect and imperfect information games, including Scotland Yard. Great work folks!", while @MichaelHBowling said "Really excited that this work is finally coming out: seeing search, learning, and game theory really demonstrate its generality. So glad that I get to work with this great team".
Leading researcher Sergey Levine (University of California, Berkeley) published "RvS: What is Essential for Offline RL via Supervised Learning?" @HochreiterSepp tweeted "ArXiv
The paper shared the most on social media this week is by a team at OpenAI: "GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models" by Alex Nichol et al (Dec 2021) with 1877 shares. @aicrumb (crumb) tweeted "1000 steps/model (base/upscale) "steampunk style oil painting of a mountain with canons poking out of it's sides"".
The most influential Twitter user discussing papers is 彩恵りり🧚♀️サイエンスライター✨おしごと募集中 who shared "Pokemon: Protected Logic Qubit Derived from the 0-$\pi$ Qubit" by J. Q. You et al (Dec 2021) and said: "Pokemon量子ビット...🤔 Pokemon: Protected Logic Qubit Derived from the 0-π Qubit".
Over the past week, nine new papers were published in "Computer Science - Multiagent Systems".
Over the past week, 14 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: "Collective Intelligence for Deep Learning: A Survey of Recent Developments" by David Ha et al (Nov 2021), which was referenced 1 time, including in the article Advancing Deep Learning With Collective Intelligence: Google Brain Surveys Recent Developments in SyncedReview.com. The paper also got the most social media traction with 329 shares. A Twitter user, @EdHenry_, commented "I see a Foerster et al. reference but I've not seen much work in the learning to communicate space, which I view as an emergent behavior, the last few years. This, of course, could be a bias in my sampling of papers these days. Maybe this will change!".
Over the past week, 26 new papers were published in "Computer Science - Robotics".