Week Ending 1.23.2022
RESEARCH WATCH: 1.23.2022
Over the past week, 951 new papers were published in "Computer Science".
The paper discussed most in the news over the past week was by a team at Google: "Measuring Attribution in Natural Language Generation Models" by Hannah Rashkin et al (Dec 2021), which was referenced 5 times, including in the article LaMDA: Towards Safe, Grounded, and High-Quality Dialog Models for Everything in Google AI Blog. The paper got social media traction with 51 shares. The authors present a new evaluation framework entitled Attributable to Identified Sources (AIS) for assessing the output of natural language generation models, when such output pertains to the external world. A user, @data_topology, tweeted "Commented preview: Google research w/ their New Attribution Framework to Identified Sources (AIS) in Natural Language Models. Watch ideas for the Future of XAI: ▶️ ◀️ #XAI #Attribution #NLG".
Leading researcher Luc Van Gool (Computer Vision Laboratory) published "Collapse by Conditioning: Training Class-conditional GANs with Limited Data" @summarizedml tweeted "Class-conditioning causes mode collapse in limited data settings, where unconditional learning leads to satisfactory generativeability. We propose a training strategy that effectively 📄".
The paper shared the most on social media this week is by a team at Google: "LaMDA: Language Models for Dialog Applications" by Romal Thoppilan et al (Jan 2022) with 406 shares. @NielsRogge (Niels Rogge) tweeted "So trained a Transformer for almost 60 days on 1024 TPUv3 chips. It can now count how many apples there are left".
This week was very active for "Computer Science - Artificial Intelligence", with 176 new papers.
The paper discussed most in the news over the past week was "AI and the Sense of Self" by Srinath Srinivasa et al (Jan 2022), which was referenced 2 times, including in the article Why giving AI ‘human ethics’ is probably a terrible idea in The Next Web. The paper was shared 3 times in social media. A Twitter user, @hybrid_vision, said "Formalised ethical constructs are deployed by clinicians in life/death scenarios every single day worldwide. These 'rules' are first enshrined in medical doctrine, followed and sometimes broken. Alignment is feasible. Researchers explore AI Sense of Self".
Leading researcher Luc Van Gool (Computer Vision Laboratory) came out with "Collapse by Conditioning: Training Class-conditional GANs with Limited Data" @summarizedml tweeted "Class-conditioning causes mode collapse in limited data settings, where unconditional learning leads to satisfactory generativeability. We propose a training strategy that effectively 📄".
The paper shared the most on social media this week is by a team at Google: "LaMDA: Language Models for Dialog Applications" by Romal Thoppilan et al (Jan 2022)
Over the past week, 187 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 Shandong University of Science and Technology: "Counteracting Dark Web Text-Based CAPTCHA with Generative Adversarial Learning for Proactive Cyber Threat Intelligence" by Ning Zhang et al (Jan 2022), which was referenced 2 times, including in the article Researchers Develop CAPTCHA Solver to Aid Dark Web Research in News-Primer.com. The paper got social media traction with 7 shares. On Twitter, @Adolfo_Hdez commented "Counteracting #DarkWeb text-based #CAPTCHA with generative adversarial learning for proactive #cyber #threatIntelligence: Researchers develop CAPTCHA solver that can overcome 94.4% of real challenges on darkwebsites. #cybersecurity".
Leading researcher Luc Van Gool (Computer Vision Laboratory) published "Collapse by Conditioning: Training Class-conditional GANs with Limited Data" @summarizedml tweeted "Class-conditioning causes mode collapse in limited data settings, where unconditional learning leads to satisfactory generativeability. We propose a training strategy that effectively 📄".
The paper shared the most on social media this week is by a team at Tel Aviv University: "Stitch it in Time: GAN-Based Facial Editing of Real Videos" by Rotem Tzaban et al (Jan 2022) with 176 shares. @HeadZonkStudios (Head-ZonkStudios) tweeted "Oh, this is going to be used to do some horrible things".
This week was active for "Computer Science - Computers and Society", with 38 new papers.
The paper shared the most on social media this week is by a team at University of Edinburgh: "SoK: Blockchain Governance" by Aggelos Kiayias et al (Jan 2022) with 86 shares. @bkschaffer (2 ₳lph₳) tweeted "Required reading for any serious and long-term Cardano holders. This right here is the real honey hole for long term prosperity, health and sustainability of the ecosystem. Managing a small town meeting is hard, governing this global beast is an absurd challenge. Get involved.👇🏻".
This week was active for "Computer Science - Human-Computer Interaction", with 29 new papers.
The paper discussed most in the news over the past week was "Obstacle avoidance for blind people using a 3D camera and a haptic feedback sleeve" by Manuel Zahn et al (Jan 2022), which was referenced 1 time, including in the article New Device Lets People Who Are Blind “See” in Infrared in Futurism. The paper got social media traction with 6 shares.
This week was very active for "Computer Science - Learning", with 303 new papers.
The paper discussed most in the news over the past week was by a team at OpenAI: "Grokking: Generalization Beyond Overfitting on Small Algorithmic Datasets" by Alethea Power et al (Jan 2022), which was referenced 2 times, including in the article What is grokking in machine learning? in Analytics India Magazine. The paper got social media traction with 193 shares. The investigators propose to study generalization of neural networks on small algorithmically generated datasets. On Twitter, @sebjrlee commented "I'm excited to give this a read! The fact that NN "generalization can happen well past the point of overfitting" is still pretty wild to me", while @JMGrange observed "This paper had a couple things that totally defied my own intution. Less data = more, not less optimization🤯".
The paper shared the most on social media this week is by a team at Tel Aviv University: "Stitch it in Time: GAN-Based Facial Editing of Real Videos" by Rotem Tzaban et al (Jan 2022)
Over the past week, 14 new papers were published in "Computer Science - Multiagent Systems".
The paper discussed most in the news over the past week was by a team at University of California, Irvine: "Anytime Optimal PSRO for Two-Player Zero-Sum Games" by Stephen McAleer et al (Jan 2022), which was referenced 1 time, including in the article UC Irvine & DeepMind’s Anytime Optimal PSRO: Guaranteed Convergence to a Nash Equilibrium With Decreased Exploitability in Two-Player Zero-Sum Games in SyncedReview.com. The paper was shared 3 times in social media. A user, @summarizedml, tweeted "Anytime Optimal Double Oracle (AODO), a tabular double oracle algorithm for 2-player zero-sum 📄".
Over the past week, 12 new papers were published in "Computer Science - Neural and Evolutionary Computing".
Over the past week, 39 new papers were published in "Computer Science - Robotics".
The paper discussed most in the news over the past week was by a team at NYU: "BiConMP: A Nonlinear Model Predictive Control Framework for Whole Body Motion Planning" by Avadesh Meduri et al (Jan 2022), which was referenced 1 time, including in the article Video Friday: An Agile Year in Spectrum Online.