Week Ending 07.07.19
RESEARCH WATCH: 07.07.19
Over the past week, 1,004 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 188 times, including in the article Here Are A Handful Of Incredibly Realistic Deepfake Videos Of Zuckerberg And Chris Christie in Daily Caller. 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 61456 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 "Learning the Arrow of Time".
The paper shared the most on social media this week is "The Ramanujan Machine: Automatically Generated Conjectures on Fundamental Constants" by Gal Raayoni et al (Jun 2019) with 203 shares. The researchers propose a novel and systematic approach that leverages algorithms for deriving new mathematical formulas for fundamental constants and help reveal their underlying structure. @octonion (Christopher D. Long) tweeted "That first relation on e can be restated as this process: r_0 = 1/e ceil(1/r_0) = 3, r_1 = 3 - r_0 ceil(1/r_1) = 4 r_2 = 4 - r_1 ceil(2/r_2) = 5 r_3 = 5 - 2*r_2 ceil(3/r_3) = 6 r_4 = 6 - 3*r_3".
Over the past week, 75 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 Université de Montréal: "Learning the Arrow of Time" by Nasim Rahaman et al (Jul 2019), which was referenced 3 times, including in the article Is this AI developing a sense of time? in ZDNet. The paper got social media traction with 24 shares. A user, @huisier, tweeted "A #NeuralNetwork that modifies #ReinforcementLearning,the pursuit of actions leading to goals. It computes the likelihood that once a given state of affairs leads to another, the process is not likely to be reversed to the earlier state".
The paper shared the most on social media this week is "The Ramanujan Machine: Automatically Generated Conjectures on Fundamental Constants" by Gal Raayoni et al (Jun 2019)
Over the past week, 195 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 Samsung: "Few-Shot Adversarial Learning of Realistic Neural Talking Head Models" by Egor Zakharov et al (May 2019)
Leading researcher Devi Parikh (Georgia Institute of Technology) published "Chasing Ghosts: Instruction Following as Bayesian State Tracking".
The paper shared the most on social media this week is by a team at DeepMind: "Sim2real transfer learning for 3D pose estimation: motion to the rescue" by Carl Doersch et al (Jul 2019) with 93 shares. @hirotomusiker (hiroto) tweeted "Bridging a sim2real gap - a network trained using: - optical flow and 2D keypoint motions of SYNTHETIC images as input - 3D pose data as ground-truth performs well for 3D pose estimation of REAL images. quoted from".
Over the past week, 23 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 106 times, including in the article Endless AI-generated spam risks clogging up Google’s search results in The Verge. 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 289 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".
The paper shared the most on social media this week is by a team at Microsoft: "Toward Fairness in AI for People with Disabilities: A Research Roadmap" by Anhong Guo et al (Jul 2019) with 76 shares. @marypcbuk (Mary Branscombe) tweeted "research roadmap for whether AI systems are fair to or exclude the disabled ; I'd love to see this cover which platforms the AI systems are available on and if that includes or excludes PWD".
This week was active for "Computer Science - Human-Computer Interaction", with 28 new papers.
The paper shared the most on social media this week is by a team at Microsoft: "Toward Fairness in AI for People with Disabilities: A Research Roadmap" by Anhong Guo et al (Jul 2019)
This week was very active for "Computer Science - Learning", with 353 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)
Leading researcher Yoshua Bengio (Université de Montréal)
The paper shared the most on social media this week is "The Ramanujan Machine: Automatically Generated Conjectures on Fundamental Constants" by Gal Raayoni et al (Jun 2019)
Over the past week, 11 new papers were published in "Computer Science - Multiagent Systems".
Over the past week, 22 new papers were published in "Computer Science - Neural and Evolutionary Computing".
Leading researcher Kyunghyun Cho (New York University) came out with "A Unified Framework of Online Learning Algorithms for Training Recurrent Neural Networks".
This week was active for "Computer Science - Robotics", with 46 new papers.
The paper discussed most in the news over the past week was by a team at University of Washington: "A laser-microfabricated electrohydrodynamic thruster for centimeter-scale aerial robots" by Elma Dedic et al (Jun 2019), which was referenced 3 times, including in the article An Itty-Bitty Robot That Lifts Off Like a Sci-Fi Spaceship in Wired News. The paper author, Igor Novosselov (University of Washington), was quoted saying "On the way from Point A to Point B, they have multiple collisions with neutral molecules, which is air—nitrogen, oxygen, a little bit of CO2 and water". The paper was shared 2 times in social media. A user, @maxsrinivas, tweeted "The #future of #drones, solid state propulsion. This paper discuses an alternative means of generating thrust that is not seen in biology: #Electrohydrodynamic (EHD) #thrust".