🍪 Cookies talks

An informal journal club with sweet treats.

I organized the talks from Fall 2017 until Fall 2018. Here's a list of what was covered during that time:

• On October 9, 2018 Will Grathwohl presented Representation Learning with Contrastive Predictive Coding by van den Oord et al. (DeepMind).
• On October 2, 2018 Tingwu Wang presented Time-Agnostic Prediction: Predicting Predictable Video Frames by Jayaraman et al. (UC Berkeley).
• On September 18, 2018 Hannes Bretschneider presented Predicting the clinical impact of human mutation with deep neural networks by Sundaram et al.
• On September 11, 2018 Atef Chaudhury gave an overview of recent work in sentence representation.
• On September 4, 2018 Arvie Fryndenlund presented Using Morphological Knowledge in Open-Vocabulary Neural Language Models by Matthews et al. (CMU, DeepMind).
• On August 14, 2018 Kamyar Ghasemipour presented Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor by Haarnoja et al. (UC Berkeley).
• On August 7, 2018 Guodong Zhang presented Fast Approximate Natural Gradient Descent in a Kronecker-factored Eigenbasis by George et al. (MILA, Facebook).
• On July 31, 2018 Renjie Liao presented Learning unbelievable probabilities by Pitkow et al. (U. Rochester, Columbia U.).
• On July 24, 2018 William Saunders presented Concrete Problems in AI Safety by Amodei et al. (Google, Stanford, UC Berkeley, OpenAI).
• On July 10, 2018 Makarand Tapaswi presented Taskonomy: Disentangling Task Transfer Learning by Zamir et al. (Stanford, UC Berkeley).
• On July 3, 2018 Jinliang Wei presented Deep Gradient Compression: Reducing the Communication Bandwidth for Distributed Training by Lin et al. (Tsinghua, Beijing, Stanford, Google).
• On June 26, 2018 Elliot Creager (me) presented Fixing a Broken ELBO by Alemi et al. (Google).
• On June 19, 2018 Eric Langlois presented Reinforcement Learning and Control as Probabilistic Inference: Tutorial and Review by Levine (UC Berkeley).
• On June 12, 2018 Marc-Etienne Brunet presented Why Is My Classifier Discriminatory? by Chen et al. (MIT).
• On June 5, 2018 Ariel Herbert-Voss presented The Secret Sharer: Measuring Unintended Neural Network Memorization & Extracting Secrets by Carlini et al. (UC Berkeley, U. Singapore, Google).
• On April 26, 2018 Masha Shugrina presented Learning local shape descriptors from part correspondences with multi-view convolutional networks by Huang et al. (UMass Amherst, IIT Bombay, Adobe).
• On April 10, 2018 Mengye Ren presented Efficient Neural Architecture Search via Parameter Sharing by Pham et al. (Google, CMU, Stanford).
• On March 27, 2018 Matthew MacKay presented Improving GANs Using Optimal Transport by Salimans et al. (OpenAI, Rutgers U.).
• On March 20, 2018 Ekansh Sharma presented Auto-Encoding Sequential Monte Carlo by Le et al. (U. Oxford, U. Warwick).
• On March 13, 2018 Chris Cremer presented Variational Intrinsic Control by Gregor et al. (Google).
• On March 6, 2018 Atef Chaudhury presented Generating Wikipedia by Summarizing Long Sequences by Liu et al. (Google).
• On February 13, 2018 Shems Saleh presented Scalable Levy Process Priors for Spectral Kernel Learning by Jang et al. (Cornell U).
• On January 30, 2018 Paul Vicol presented Neural Speed Reading via Skim-RNN by Seo et al. (U Washington, Seoul Nat'l U, Allen Inst., xnor.ai).
• On January 16, 2018 Shenyang Sun presented Fast Kernel Learning for Multidimensional Pattern Extrapolation by Wilson et al. (CMU, Washington U, Columbia U).
• On December 19, 2017 Will Grathwohl presented Scalable Log Determinants for Gaussian Process Kernel Learning by Dong et al. (Cornell, Phillips Research).
• On October 24, 2017 Bowen Yu presented Markov Chain Monte Carlo and Variational Inference: Bridging the Gap by Salimans, Kingma and Welling (U Amsterdam)
• On October 31, 2017 Jason Li presented SampleRNN: An Unconditional End-to-End Neural Audio Generation Model by Mehri et al. (U Montreal, IIT Kanpur).
• On October 17, 2017 Lisa Zhang presented Variational Reasoning for Question Answering with Knowledge Graph by Zhang et al. (Georgia Teach, Amazon)
• On November 7, 2017 Stavros Tsogkas presented Look, Listen and Learn by Arandjelović and Zisserman (Deepmind, Oxford).
• On November 14, 2017 Eric Langlois presented Deep reinforcement learning from human preferences by Christiano et al. (OpenAI, DeepMind).
• On November 21, 2017 Eleni Triantafillou presented Meta-Learning with Temporal Convolutions by Mishra et al. (Berkeley, OpenAI).
• On November 28, 2017 Elliot Creager (me) presented Hierarchical Implicit Models and Likelihood-Free Variational Inference by Tran et al. (Columbia, Princeton).
• On December 12, 2017 Ioan Andrei Barsan presented Houdini: Fooling Deep Structured Visual and Speech Recognition Models with Adversarial Examples by Cisse et al. (Facebook, Bar-Ilan U).
• On October 10, 2017 David Madras presented On Calibration of Modern Neural Networks by Guo et al. (Cornell).
• On October 3, 2017 Amlan Kar presented Multi-view Supervision for Single-view Reconstruction via Differentiable Ray Consistency by Tulsiani et al. (UC Berkeley).
• On September 26, 2017 Renjie Liao presented Learning Combinatorial Optimization Algorithms over Graphs by Dai et al. (Georgia Tech).