Week Ending 06.16.19
RESEARCH WATCH: 06.16.19
Over the past week, 1,050 new papers were published in "Computer Science".
The paper discussed most in the news over the past week was by a team at Samsung: "Few-Shot Adversarial Learning of Realistic Neural Talking Head Models" by Egor Zakharov et al (May 2019), which was referenced 186 times, including in the article Samsung deepfake AI could fabricate a video of you from a single profile pic in Sott.net. The paper author, Egor Zakharov (Samsung), was quoted saying "Effectively, the learned model serves as a realistic avatar of a person". The paper also got the most social media traction with 61295 shares. On Twitter, @catovitch said "I wonder if this can/will be used with video compression. If it can be done in real time, the decoder could just be told "build a face our of this frame, then rotate it to these 3D positions for the next 5 frames"".
Leading researcher Yoshua Bengio (Université de Montréal) came out with "Learning Powerful Policies by Using Consistent Dynamics Model" @gastronomy tweeted "> Model-based Reinforcement Learning approaches have the promise of being sample efficient. Much of the progress in learning dynamics models in RL has".
The paper shared the most on social media this week is by a team at Google: "Tackling Climate Change with Machine Learning" by David Rolnick et al (Jun 2019) with 1075 shares. The authors describe how machine learning can be a powerful tool in reducing greenhouse gas emissions and helping society adapt to a changing climate. @ikdeepl (Ilia Karmanov) tweeted "Yikes “In 2006, at least two Scottish seafood firms flew hundreds of metric tons of shrimp from Scotland to China and Thailand for peeling, then back to Scotland for sale – because they could save on labor costs”".
Over the past week, 82 new papers were published in "Computer Science - Artificial Intelligence".
The paper discussed most in the news over the past week was "Automated Speech Generation from UN General Assembly Statements: Mapping Risks in AI Generated Texts"by Joseph Bullock et al (Jun 2019), which was referenced 99 times, including in the article Top AI researchers race to detect “deepfake” videos: “We are outgunned” in Washington Post. The paper got social media traction with 87 shares. A user, @naumenko_roman, tweeted "Some savings in budget thanks to the cloud. "The language model was trained in under 13 hours on NVIDIA K80 GPUs, costing as little as $7.80 on AWS spot instances."", while @lolitataub said "AI speech generator anyone? ... "matchethe style and cadence of real UN speeches roughly 90% of the time."".
Leading researcher Yoshua Bengio (Université de Montréal) published "Learning Powerful Policies by Using Consistent Dynamics Model" @gastronomy tweeted "> Model-based Reinforcement Learning approaches have the promise of being sample efficient. Much of the progress in learning dynamics models in RL has".
The paper shared the most on social media this week is by a team at Google: "Tackling Climate Change with Machine Learning" by David Rolnick et al (Jun 2019) with 1075 shares. The researchers describe how machine learning can be a powerful tool in reducing greenhouse gas emissions and helping society adapt to a changing climate. @ikdeepl (Ilia Karmanov) tweeted "Yikes “In 2006, at least two Scottish seafood firms flew hundreds of metric tons of shrimp from Scotland to China and Thailand for peeling, then back to Scotland for sale – because they could save on labor costs”".
This week was active for "Computer Science - Computer Vision and Pattern Recognition", with 231 new papers.
The paper discussed most in the news over the past week was by a team at Samsung: "Few-Shot Adversarial Learning of Realistic Neural Talking Head Models" by Egor Zakharov et al (May 2019), which was referenced 186 times, including in the article Samsung deepfake AI could fabricate a video of you from a single profile pic in Sott.net. The paper author, Egor Zakharov (Samsung), was quoted saying "Effectively, the learned model serves as a realistic avatar of a person". The paper also got the most social media traction with 61295 shares. A user, @catovitch, tweeted "I wonder if this can/will be used with video compression. If it can be done in real time, the decoder could just be told "build a face our of this frame, then rotate it to these 3D positions for the next 5 frames"".
Leading researcher Sergey Levine (University of California, Berkeley) published "Deep Reinforcement Learning for Industrial Insertion Tasks with Visual Inputs and Natural Rewards".
The paper shared the most on social media this week is "Large-scale Landmark Retrieval/Recognition under a Noisy and Diverse Dataset" by Kohei Ozaki et al (Jun 2019) with 120 shares. @smly (Standard ML/Yeah!) tweeted "Our 1st/3rd place solution is now available on arXiv. It will be presented at the Second Landmark Recognition Workshop at CVPR 2019 (Long Beach, CA)".
Over the past week, 16 new papers were published in "Computer Science - Computers and Society".
The paper discussed most in the news over the past week was by a team at University of Washington: "Defending Against Neural Fake News" by Rowan Zellers et al (May 2019), which was referenced 103 times, including in the article Top AI researchers race to detect “deepfake” videos: “We are outgunned” in Washington Post. The paper author, Rowan Zellers (University of Washington), was quoted saying "These models are not capable, we think right now, of inflicting serious harm. Maybe in a few years they will be, but not yet". The paper got social media traction with 273 shares. A user, @Thom_Wolf, tweeted "If you want a sneek-peek in and co-workers work on GROVER (a 1.5 billion param GPT-2-like model), check this live tweet 👇 Interesting hints, results, and analysis! Paper: Demo".
Leading researcher Yoshua Bengio (Université de Montréal) published "Tackling Climate Change with Machine Learning", which is getting attention with 160 shares this week. The authors describe how machine learning can be a powerful tool in reducing greenhouse gas emissions and helping society adapt to a changing climate. @ikdeepl tweeted "Yikes “In 2006, at least two Scottish seafood firms flew hundreds of metric tons of shrimp from Scotland to China and Thailand for peeling, then back to Scotland for sale – because they could save on labor costs”". This paper was also shared the most on social media with 1,075 tweets. @ikdeepl (Ilia Karmanov) tweeted "Yikes “In 2006, at least two Scottish seafood firms flew hundreds of metric tons of shrimp from Scotland to China and Thailand for peeling, then back to Scotland for sale – because they could save on labor costs”".
Over the past week, 17 new papers were published in "Computer Science - Human-Computer Interaction".
The paper discussed most in the news over the past week was "Making ethical decisions for the immersive web" by Diane Hosfelt (May 2019), which was referenced 1 time, including in the article How to Design Social VR Spaces in Medium.com. The paper got social media traction with 44 shares. The authors focus on how they can build an immersive web that encourages ethical development and usage. A Twitter user, @tylerzeph, said "Want a status report on where VR + ethics are at? Concerned about ethics and 'immersive media'? Generally concerned with technology, privacy, and ethics? If yes, read this paper".
This week was very active for "Computer Science - Learning", with 449 new papers.
The paper discussed most in the news over the past week was by a team at Samsung: "Few-Shot Adversarial Learning of Realistic Neural Talking Head Models" by Egor Zakharov et al (May 2019), which was referenced 186 times, including in the article Samsung deepfake AI could fabricate a video of you from a single profile pic in Sott.net. The paper author, Egor Zakharov (Samsung), was quoted saying "Effectively, the learned model serves as a realistic avatar of a person". The paper also got the most social media traction with 61295 shares. A Twitter user, @catovitch, observed "I wonder if this can/will be used with video compression. If it can be done in real time, the decoder could just be told "build a face our of this frame, then rotate it to these 3D positions for the next 5 frames"".
Leading researcher Yoshua Bengio (Université de Montréal) came out with "Tackling Climate Change with Machine Learning", which is getting buzz on social media with 160 shares this week. The authors describe how machine learning can be a powerful tool in reducing greenhouse gas emissions and helping society adapt to a changing climate. @ikdeepl tweeted "Yikes “In 2006, at least two Scottish seafood firms flew hundreds of metric tons of shrimp from Scotland to China and Thailand for peeling, then back to Scotland for sale – because they could save on labor costs”". This paper was also shared the most on social media with 1,076 tweets. @ikdeepl (Ilia Karmanov) tweeted "Yikes “In 2006, at least two Scottish seafood firms flew hundreds of metric tons of shrimp from Scotland to China and Thailand for peeling, then back to Scotland for sale – because they could save on labor costs”".
Over the past week, 11 new papers were published in "Computer Science - Multiagent Systems".
Over the past week, 30 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 University of Notre Dame: "Nonvolatile Spintronic Memory Cells for Neural Networks" by Andrew W. Stephan et al (May 2019), which was referenced 1 time, including in the article Spintronic memory cells for neural networks in PhysOrg.com. The paper was shared 1 time in social media.
Leading researcher Pieter Abbeel (University of California, Berkeley) published "Sub-policy Adaptation for Hierarchical Reinforcement Learning".
The paper shared the most on social media this week is "Weight Agnostic Neural Networks" by Adam Gaier et al (Jun 2019) with 790 shares. The investigators question to what extent neural network architectures alone, without learning any weight parameters, can encode solutions for a given task. @RahulAPanicker (Rahul Panicker) tweeted "Architecture over weights. 80% accuracy on MNIST with random weights. Very cool! On with connectionism".
Over the past week, 42 new papers were published in "Computer Science - Robotics".
The paper discussed most in the news over the past week was "EVDodge: Embodied AI For High-Speed Dodging On A Quadrotor Using Event Cameras" by Nitin J. Sanket et al (Jun 2019), which was referenced 3 times, including in the article AI helps drones dodge fast-moving objects in Venturebeat. The paper got social media traction with 5 shares.
Leading researcher Ruslan Salakhutdinov (Carnegie Mellon University) published "Efficient Exploration via State Marginal Matching" @rndmcnlly tweeted "sounds similar to your Targeted Trajectory Distribution stuff from 2006".
The paper shared the most on social media this week is by a team at Google: "Search on the Replay Buffer: Bridging Planning and Reinforcement Learning" by Benjamin Eysenbach et al (Jun 2019) with 73 shares.