Stanford AI Lab Papers and Talks at NeurIPS 2020

The Neural Information Processing Systems (NeurIPS) 2020 conference is being hosted virtually from Dec 6th – Dec 12th. We’re excited to share all the work from SAIL that’s being presented, and you’ll find links to papers, videos and blogs below. Feel free to reach out to the contact authors directly to learn more about the work that’s happening at Stanford!

List of Accepted Papers


Provably Efficient Reward-Agnostic Navigation with Linear Value Iteration


Authors: Andrea Zanette, Alessandro Lazaric, Mykel Kochenderfer, Emma Brunskill

Contact: zanette@stanford.edu

Keywords: reinforcement learning, function approximation, exploration


Acceleration with a Ball Optimization Oracle


Authors: Yair Carmon, Arun Jambulapati, Qijia Jiang, Yujia Jin, Yin Tat Lee, Aaron Sidford, Kevin Tian

Contact: kjtian@stanford.edu

Award nominations: Oral presentation

Links: Paper

Keywords: convex optimization, local search, trust region methods


BanditPAM: Almost Linear Time k-Medoids Clustering via Multi-Armed Bandits


Authors: Mo Tiwari, Martin Jinye Zhang, James Mayclin, Sebastian Thrun, Chris Piech, Ilan Shomorony

Contact: Motiwari@stanford.edu

Links: Paper | Video

Keywords: clustering, k-means, k-medoids, multi-armed bandits


CaSPR: Learning Canonical Spatiotemporal Point Cloud Representations


Authors: Davis Rempe, Tolga Birdal, Yongheng Zhao, Zan Gojcic, Srinath Sridhar, Leonidas J. Guibas

Contact: drempe@stanford.edu

Links: Paper | Video | Website

Keywords: 3d vision, dynamic point clouds, representation learning


Compositional Explanations of Neurons


Authors: Jesse Mu, Jacob Andreas

Contact: muj@stanford.edu

Award nominations: oral

Links: Paper

Keywords: interpretability, explanation, deep learning, computer vision, natural language processing, adversarial examples


Continuous Meta-Learning without Tasks


Authors: James Harrison, Apoorva Sharma, Chelsea Finn, Marco Pavone

Contact: jharrison@stanford.edu

Links: Paper

Keywords: meta-learning, continuous learning, changepoint detection


Deep learning versus kernel learning: an empirical study of loss landscape geometry and the time evolution of the Neural Tangent Kernel


Authors: Stanislav Fort, Gintare Karolina Dziugaite, Mansheej Paul, Sepideh Kharaghani, Daniel M. Roy, Surya Ganguli

Contact: sfort1@stanford.edu

Links: Paper

Keywords: loss landscape, neural tangent kernel, linearization, taylorization, basin, nonlinear advantage


Diversity can be Transferred: Output Diversification for White- and Black-box Attacks


Authors: Yusuke Tashiro, Yang Song, Stefano Ermon

Contact: ytashiro@stanford.edu

Links: Paper | Website

Keywords: adversarial examples, deep learning, robustness


Evidential Sparsification of Multimodal Latent Spaces in Conditional Variational Autoencoders


Authors: Masha Itkina, Boris Ivanovic, Ransalu Senanayake, Mykel J. Kochenderfer, and Marco Pavone

Contact: mitkina@stanford.edu

Links: Paper | Website

Keywords: sparse distributions, generative models, discrete latent spaces, behavior prediction, image generation


Federated Accelerated Stochastic Gradient Descent


Authors: Honglin Yuan, Tengyu Ma

Contact: yuanhl@stanford.edu

Award nominations: Best Paper Award of Federated Learning for User Privacy and Data Confidentiality in Conjunction with ICML 2020 (FL-ICML’20)

Links: Paper | Website

Keywords: federated learning, local sgd, acceleration, fedac


Fourier-transform-based attribution priors improve the interpretability and stability of deep learning models for genomics


Authors: Alex Michael Tseng, Avanti Shrikumar, Anshul Kundaje

Contact: amtseng@stanford.edu

Links: Paper | Website

Keywords: deep learning, interpretability, attribution prior, computational biology, genomics


From Trees to Continuous Embeddings and Back: Hyperbolic Hierarchical Clustering


Authors: Ines Chami, Albert Gu, Vaggos Chatziafratis, Christopher Re

Contact: chami@stanford.edu

Links: Paper | Video | Website

Keywords: hierarchical clustering, hyperbolic embeddings


FrugalML: How to Use ML Prediction APIs More Accurately and Cheaply


Authors: Lingjiao Chen; Matei Zaharia; James Zou

Contact: lingjiao@stanford.edu

Links: Paper | Blog Post | Website

Keywords: machine learning as a service, ensemble learning, meta learning, systems for machine learning


Generative 3D Part Assembly via Dynamic Graph Learning


Authors: Jialei Huang, Guanqi Zhan, Qingnan Fan, Kaichun Mo, Lin Shao, Baoquan Chen, Leonidas Guibas, Hao Dong

Contact: fqnchina@gmail.com

Links: Paper

Keywords: 3d part assembly, dynamic graph learning


Generative 3D Part Assembly via Dynamic Graph Learning


Authors: Jialei Huang*, Guanqi Zhan*, Qingnan Fan, Kaichun Mo, Lin Shao, Baoquan Chen, Leonidas J. Guibas, Hao Dong

Contact: kaichun@cs.stanford.edu

Links: Paper | Website

Keywords: 3d part assembly, graph neural network


Gradient Surgery for Multi-Task Learning


Authors: Tianhe Yu, Saurabh Kumar, Abhishek Gupta, Sergey Levine, Karol Hausman, Chelsea Finn

Contact: tianheyu@cs.stanford.edu

Links: Paper | Website

Keywords: multi-task learning, deep reinforcement learning


HiPPO: Recurrent Memory with Optimal Polynomial Projections


Authors: Albert Gu*, Tri Dao*, Stefano Ermon, Atri Rudra, Chris Ré

Contact: albertgu@stanford.edu, trid@stanford.edu

Links: Paper | Blog Post

Keywords: representation learning, time series, recurrent neural networks, lstm, orthogonal polynomials


Identifying Learning Rules From Neural Network Observables


Authors: Aran Nayebi, Sanjana Srivastava, Surya Ganguli, Daniel L.K. Yamins

Contact: anayebi@stanford.edu

Award nominations: Spotlight Presentation

Links: Paper | Website

Keywords: computational neuroscience, learning rule, deep networks


Improved Techniques for Training Score-Based Generative Models


Authors: Yang Song, Stefano Ermon

Contact: songyang@stanford.edu

Links: Paper

Keywords: score-based generative modeling, score matching, deep generative models


Language Through a Prism: A Spectral Approach for Multiscale Language Representations


Authors: Alex Tamkin, Dan Jurafsky, Noah Goodman

Contact: atamkin@stanford.edu

Links: Paper

Keywords: bert, signal processing, self-supervised learning, interpretability, multiscale


Large-Scale Methods for Distributionally Robust Optimization


Authors: Daniel Levy, Yair Carmon, John Duchi, Aaron Sidford

Contact: danilevy@stanford.edu

Links: Paper

Keywords: robustness dro optimization large-scale optimal


Learning Physical Graph Representations from Visual Scenes


Authors: Daniel Bear, Chaofei Fan, Damian Mrowca, Yunzhu Li, Seth Alter, Aran Nayebi, Jeremy Schwartz, Li F. Fei-Fei, Jiajun Wu, Josh Tenenbaum, Daniel L. Yamins

Contact: dbear@stanford.edu

Links: Paper | Blog Post | Website

Keywords: structure learning, graph learning, visual scene representations, unsupervised learning, unsupervised segmentation, object-centric representation, intuitive physics


MOPO: Model-based Offline Policy Optimization


Authors: Tianhe Yu*, Garrett Thomas*, Lantao Yu, Stefano Ermon, James Zou, Sergey Levine, Chelsea Finn, Tengyu Ma

Contact: tianheyu@cs.stanford.edu

Links: Paper | Website

Keywords: offline reinforcement learning, model-based reinforcement learning


MOPO: Model-based Offline Policy Optimization


Authors: Tianhe Yu, Garrett Thomas, Lantao Yu, Stefano Ermon, James Zou, Sergey Levine, Chelsea Finn, Tengyu Ma

Contact: tianheyu@cs.stanford.edu,gwthomas@stanford.edu

Links: Paper

Keywords: model-based rl, offline rl, batch rl


Measuring Robustness to Natural Distribution Shifts in Image Classification


Authors: Rohan Taori, Achal Dave, Vaishaal Shankar, Nicholas Carlini, Benjamin Recht, Ludwig Schmidt

Contact: rtaori@stanford.edu

Award nominations: Spotlight

Links: Paper | Website

Keywords: machine learning, robustness, image classification


Minibatch Stochastic Approximate Proximal Point Methods


Authors: Hilal Asi, Karan Chadha, Gary Cheng, John Duchi

Contact: chenggar@stanford.edu

Award nominations: Spotlight talk

Links: Paper

Keywords: stochastic optimization, sgd, aprox


Model-based Adversarial Meta-Reinforcement Learning


Authors: Zichuan Lin, Garrett Thomas, Guangwen Yang, Tengyu Ma

Contact: lzcthu12@gmail.com,gwthomas@stanford.edu

Links: Paper

Keywords: model-based rl, meta-rl, minimax


Multi-Plane Program Induction with 3D Box Priors


Authors: Yikai Li, Jiayuan Mao, Xiuming Zhang, William T. Freeman, Joshua B. Tenenbaum, Noah Snavely, Jiajun Wu

Contact: jiajunwu@cs.stanford.edu

Links: Paper | Video | Website

Keywords: visual program induction, 3d vision, image editing


Multi-label Contrastive Predictive Coding


Authors: Jiaming Song, Stefano Ermon

Contact: jiaming.tsong@gmail.com

Links: Paper

Keywords: representation learning, mutual information


Neuron Shapley: Discovering the Responsible Neurons


Authors: Amirata Ghorbani, James Zou

Contact: amiratag@stanford.edu

Links: Paper

Keywords: interpretability, deep learning, shapley value


No Subclass Left Behind: Fine-Grained Robustness in Coarse-Grained Classification Problems


Authors: Nimit Sharad Sohoni, Jared Alexander Dunnmon, Geoffrey Angus, Albert Gu, Christopher Ré

Contact: nims@stanford.edu

Links: Paper | Blog Post | Video

Keywords: classification, robustness, clustering, neural feature representations


Off-policy Policy Evaluation For Sequential Decisions Under Unobserved Confounding


Authors: Hongseok Namkoong, Ramtin Keramati, Steve Yadlowsky, Emma Brunskill

Contact: keramati@stanford.edu

Links: Paper

Keywords: off-policy policy evaluation, unobserved confounding, reinforcement learning


One Solution is Not All You Need: Few-Shot Extrapolation via Structured MaxEnt RL


Authors: Saurabh Kumar, Aviral Kumar, Sergey Levine, Chelsea Finn

Contact: szk@stanford.edu

Links: Paper

Keywords: robustness, diversity, reinforcement learning


Point process models for sequence detection in high-dimensional neural spike trains


Authors: Alex H. Williams, Anthony Degleris, Yixin Wang, Scott W. Linderman

Contact: ahwillia@stanford.edu

Award nominations: Selected for Oral Presentation

Links: Paper | Website

Keywords: bayesian nonparametrics, unsupervised learning


Predictive coding in balanced neural networks with noise, chaos and delays


Authors: Jonathan Kadmon, Jonathan Timcheck, Surya Ganguli

Contact: kadmonj@stanford.edu

Links: Paper

Keywords: neuroscience, predictive coding, chaos


Probabilistic Circuits for Variational Inference in Discrete Graphical Models


Authors: Andy Shih, Stefano Ermon

Contact: andyshih@stanford.edu

Links: Paper

Keywords: variational inference, discrete, high-dimensions, sum product networks, probabilistic circuits, graphical models


Provably Good Batch Off-Policy Reinforcement Learning Without Great Exploration


Authors: Yao Liu, Adith Swaminathan, Alekh Agarwal, Emma Brunskill.

Contact: yaoliu@stanford.edu

Links: Paper

Keywords: reinforcement leanring, off-policy, batch reinforcement learning


Pruning neural networks without any data by iteratively conserving synaptic flow


Authors: Hidenori Tanaka, Daniel Kunin, Daniel L. K. Yamins, Surya Ganguli

Contact: kunin@stanford.edu

Links: Paper | Video | Website

Keywords: network pruning, sparse initialization, lottery ticket


Robust Sub-Gaussian Principal Component Analysis and Width-Independent Schatten Packing


Authors: Arun Jambulapati, Jerry Li, Kevin Tian

Contact: kjtian@stanford.edu

Award nominations: Spotlight presentation

Links: Paper

Keywords: robust statistics, principal component analysis, positive semidefinite programming


Self-training Avoids Using Spurious Features Under Domain Shift


Authors: Yining Chen*, Colin Wei*, Ananya Kumar, Tengyu Ma (*equal contribution)

Contact: cynnjjs@stanford.edu

Links: Paper

Keywords: self-training, pseudo-labeling, domain shift, robustness


Wasserstein Distances for Stereo Disparity Estimation


Authors: Divyansh Garg, Yan Wang, Bharath Hariharan, Mark Campbell, Kilian Q. Weinberger, Wei-Lun Chao

Contact: divgarg@stanford.edu

Award nominations: Spotlight

Links: Paper | Video | Website

Keywords: depth estimation, disparity estimation, autonomous driving, 3d object detection, statistical learning


We look forward to seeing you at NeurIPS2020!

Read More