Schedule & Accepted Papers
Invited Speakers
This year's theme is the use of deep learning uncertainty in real-world applications, with speakers working on various problems:- Sergey Levine (Berkeley, reinforcement learning)
- Debora Marks (Harvard Medical School, genetics)
- Frank Wood (UBC, probabilistic programming)
- Yarin Gal (University of Oxford, autonomous driving)
- Dmitry Vetrov (Samsung AI centre, model compression)
- Harri Valpola (Curious AI Company, model-based control)
- Christian Leibig (Tuebingen, medical)
- Yashar Hezaveh (Kavli Institute for Particle Astrophysics and Cosmology at Stanford, astrophysics)
- Balaji Lakshminarayanan (DeepMind, medical application with ensemble uncertainty)
- David Sontag (MIT, healthcare)
- Tim Genewein (Bosch Center for AI, model compression)
Schedule
8.00 - 8.05 | Opening remarks | Yarin Gal (Oxford) | |
8.05 - 8.25 | Invited talk | Frank Wood (UBC) | Challenges at the confluence of deep learning and probabilistic programming |
8.25 - 8.45 | Invited talk | Dmitry Vetrov (Samsung AI centre) | (Semi-)Implicit Modeling as New Deep Tool for Approximate Bayesian Inference |
8.45 - 9.00 | Contributed talk | Matthias Bauer and Andriy Mnih | Resampled Priors for Variational Autoencoders |
9.00 - 9.20 | Invited talk | Debora Marks (Harvard Medical School) | Generative deep models for challenging biological problems |
9.20 - 9.40 | Invited talk | Harri Valpola (Curious AI Company) | Estimating uncertainty for model-based reinforcement learning |
9.40 - 9.55 | Poster spotlights | ||
9.55 - 10.55 | Discussion over coffee and poster session | ||
10.55 - 11.15 | Invited talk | Christian Leibig (Tuebingen) | Leveraging (Bayesian) uncertainty information: opportunities and failure modes |
11.15 - 11.30 | Contributed talk | Dan Rosenbaum, Frederic Besse, Fabio Viola, Danilo Rezende and Ali Eslami | Learning models for visual 3D localization with implicit mapping |
11.30 - 11.50 | Invited talk | Balaji Lakshminarayanan (DeepMind) | Probabilistic model ensembles for predictive uncertainty estimation |
11.50 - 13.20 | Lunch | ||
13.20 - 13.40 | Invited talk | Sergey Levine (Berkeley) | Control as Inference: a Connection Between Reinforcement Learning and Graphical Models |
13.40 - 13.55 | Contributed talk | Frank Soboczenski, Michael D. Himes, Molly D. O'Beirne, Simone Zorzan, Atılım Günes Baydin, Adam D. Cobb, Daniel Angerhausen, Giada N. Arney and Shawn D. Domagal-Goldman | Bayesian Deep Learning for Exoplanet Atmospheric Retrieval |
13.55 - 14.10 | Invited talk | Yashar Hezaveh (Stanford) | Mapping the most distant galaxies of the universe with Bayesian neural networks |
14.10 - 14.30 | Invited talk | Tim Genewein (Bosch Center for AI) | A Bayesian view on neural network compression |
14.30 - 15.30 | Discussion over coffee and poster session | ||
15.30 - 15.50 | Invited talk | David Sontag (MIT) | Bayesian deep learning for healthcare |
15.50 - 16.05 | Contributed talk | Hyunjik Kim, Andriy Mnih, Jonathan Schwarz, Marta Garnelo, Ali Eslami, Dan Rosenbaum, Oriol Vinyals and Yee Whye Teh | Attentive Neural Processes |
16.05 - 16.25 | Invited talk | Yarin Gal (Oxford) | Bayesian Deep Learning in Self-Driving Cars (and more) |
16.30 - 17.30 | Panel Session | Panellists: Sergey Levine Debora Marks Frank Wood Yarin Gal Harri Valpola Christian Leibig Yashar Hezaveh Balaji Lakshminarayanan David Sontag Moderator: Neil Lawrence |
|
17.30 - 19.30 | Poster session |
Accepted Abstracts
Authors | Title |
Matthias Bauer and Andriy Mnih | Resampled Priors for Variational Autoencoders [paper] |
Ben Poole, Sherjil Ozair, Aaron van den Oord, Alexander Alemi and George Tucker | On variational lower bounds of mutual information [paper] |
Adrià Garriga-Alonso, Carl E. Rasmussen and Laurence Aitchison | Deep Convolutional Networks as shallow Gaussian Processes [paper] |
Gintare Karolina Dziugaite, Gabriel Arpino and Daniel Roy | Towards generalization guarantees for SGD: Data-dependent PAC-Bayes priors [paper] |
Bo Dai, Hanjun Dai, Niao He, Arthur Gretton, Le Song and Dale Schuurmans | Exponential Family Estimation via Dynamics Embedding [paper] |
Eric Nalisnick, Akihiro Matsukawa, Yee Whye Teh, Dilan Gorur and Balaji Lakshminarayanan | Do Deep Generative Models Know What They Don’t Know? [paper] |
Ke Li and Jitendra Malik | Implicit Maximum Likelihood Estimation [paper] |
Emile Mathieu, Tom Rainforth, Siddharth Narayanaswamy and Yee Whye Teh | Disentangling Disentanglement [paper] |
George Tucker, Dieterich Lawson, Shane Gu and Chris J. Maddison | Doubly Reparameterized Gradient Estimators for Monte Carlo Objectives [paper] |
Chen Zeno, Itay Golan, Elad Hoffer and Daniel Soudry | Task Agnostic Continual Learning Using Online Variational Bayes [paper] |
Xu Hu, Pablo Garcia Moreno, Neil Lawrence and Andreas Damianou | β-BNN: A Rate-Distortion Perspective on Bayesian Neural Networks [paper] |
Mariia Vladimirova, Julyan Arbel and Pablo Mesejo | Bayesian neural networks become heavier-tailed with depth [paper] |
Hyunjik Kim, Andriy Mnih, Jonathan Schwarz, Marta Garnelo, Ali Eslami, Dan Rosenbaum, Oriol Vinyals and Yee Whye Teh | Attentive Neural Processes [paper] |
Salvator Lombardo, Jun Han, Christopher Schroers and Stephan Mandt | Video Compression through Deep Bayesian Learning [paper] |
Emily Fertig, Aryan Arbabi and Alex Alemi | β-VAEs can retain label information even at high compression [paper] |
Hamid Eghbal-Zadeh, Werner Zellinger and Gerhard Widmer | Mixture Density Generative Adversarial Networks [paper] |
Pengyu Cheng, Chang Liu, Chunyuan Li, Dinghan Shen, Ricardo Henao and Lawrence Carin | Straight-Through Estimator as Projected Wasserstein Gradient Flow [paper] |
Fabio Viola and Danilo Rezende | Generalized ELBO with Constrained Optimization, GECO [paper] |
Chao Ma, Yingzhen Li and Jose Miguel Hernandez Lobato | Variational Implicit Processes [paper] |
Tom Ryder, Dennis Prangle, Andy Golightly and Stephen McGough | Black-Box Autoregressive Density Estimation for State-Space Models [paper] |
Mingzhang Yin | Semi-implicit generative model [paper] |
Conor Durkan, George Papamakarios and Iain Murray | Sequential Neural Methods for Likelihood-free Inference [paper] |
Florian Wenzel, Alexander Buchholz and Stephan Mandt | Quasi-Monte Carlo Flows [paper] |
Kento Nozawa and Issei Sato | PAC-Bayes Analysis of Transferred Sentence Vectors [paper] |
Iryna Korshunova, Yarin Gal, Joni Dambre and Arthur Gretton | Conditional BRUNO: A Deep Recurrent Process for Exchangeable Labelled Data [paper] |
Ranganath Krishnan, Mahesh Subedar and Omesh Tickoo | BAR: Bayesian Activity Recognition using variational inference [paper] |
Jonathan Gordon, John Bronskill, Matthias Bauer, Sebastian Nowozin and Richard Turner | Versa: Versatile and Efficient Few-shot Learning [paper] |
Kimin Lee, Sukmin Yun, Kibok Lee, Honglak Lee, Bo Li and Jinwoo Shin | Robust Determinantal Generative Classifier for Noisy Labels and Adversarial Attacks [paper] |
Dejiao Zhang, Tianchen Zhao and Laura Balzano | Information Maximization Auto-Encoding [paper] |
Anusha Lalitha, Shubhanshu Shekhar, Tara Javidi and Farinaz Koushanfar | Fully Decentralized Federated Learning [paper] |
Maxime Wabartha, Audrey Durand, Vincent François-Lavet and Joelle Pineau | Sampling diverse neural networks for exploration in reinforcement learning [paper] |
Yao-Hung Hubert Tsai, Paul Pu Liang, Amir Zadeh, Louis-Philippe Morency and Ruslan Salakhutdinov | Learning Multimodal Representations using Factorized Deep Generative Models [paper] |
Samuel Smith, Daniel Duckworth, Semon Rezchikov, Quoc Le and Jascha Sohl-Dickstein | Stochastic natural gradient descent draws posterior samples in function space [paper] |
Chao Ma, Jose Miguel Hernandez Lobato, Noam Koenigstein, Sebastian Nowozin and Cheng Zhang | Partial VAE for Hybrid Recommender System [paper] |
Pranav Shyam, Wojciech Jaśkowski and Faustino Gomez | Model-Based Active Exploration [paper] |
Roman Novak, Lechao Xiao, Yasaman Bahri, Jaehoon Lee, Greg Yang, Daniel Abolafia, Jeffrey Pennington and Jascha Sohl-Dickstein | Bayesian Deep Convolutional Networks with Many Channels are Gaussian Processes [paper] |
Fabio Viola and Danilo Rezende | On the properties of high-capacity VAEs [paper] |
Jishnu Mukhoti, Pontus Stenetorp and Yarin Gal | On the Importance of Strong Baselines in Bayesian Deep Learning [paper] |
Pierre-Alexandre Mattei and Jes Frellsen | Refit your Encoder when New Data Comes by [paper] |
Xinyu Hu, Paul Szerlip, Theofanis Karaletsos and Rohit Singh | Applying SVGD to Bayesian Neural Networks for Cyclical Time-Series Prediction and Inference [paper] |
Ananya Kumar, Ali Eslami, Danilo Rezende, Marta Garnelo, Fabio Viola, Edward Lockhart and Murray Shanahan | Consistent Jumpy Predictions for Videos and Scenes [paper] |
Christian Henning, Johannes von Oswald, Joao Sacramento, Jean-Pascal Pfister and Benjamin F. Grewe | Approximating the Predictive Distribution via Adversarially-Trained Hypernetworks [paper] |
Ivan Ovinnikov | Poincaré Wasserstein Autoencoder [paper] |
Shengyang Sun, Guodong Zhang, Jiaxin Shi and Roger Grosse | Functional Variational Bayesian Neural Networks [paper] |
Da Tang, Dawen Liang and Tony Jebara | Correlated Variational Auto-Encoders [paper] |
Jiawei He, Yu Gong, Joseph Marino, Greg Mori and Andreas Lehrmann | Variational Latent Dependency Learning [paper] |
Tim Pearce | Bayesian Neural Network Ensembles [paper] |
Artur Bekasov and Iain Murray | Bayesian Adversarial Spheres: Bayesian Inference and Adversarial Examples in a Noiseless Setting [paper] |
Eric Nalisnick, Akihiro Matsukawa, Yee Whye Teh, Dilan Gorur and Balaji Lakshminarayanan | Hybrid Models with Deep and Invertible Features [paper] |
Andrey Malinin and Mark Gales | Prior Networks for Detection of Adversarial Attacks [paper] |
Seb Farquhar and Yarin Gal | A Unifying View of Bayesian Continual Learning [paper] |
Micha Livne and David Fleet | TzK Flow - Conditional Generative Model [paper] |
Jason Ramapuram, Alexandros Kalousis, Russ Webb and Maurits Diephuis | Variational Saccading: Efficient Inference for Large Resolution Images [paper] |
Stanislav Fort | Machine learning approach to detection and characterization of X-ray cavities in clusters of galaxies [paper] |
Yibo Yang and Paris Perdikaris | Physics-informed deep generative models [paper] |
Kumar Sricharan and Ashok Srivastava | Building robust classifiers through generation of confident out of distribution examples [paper] |
Benjamin Bloem-Reddy and Yee Whye Teh | Neural network models of exchangeable sequences [paper] |
Ghassen Jerfel, Erin Grant, Thomas L. Griffiths and Katherine A. Heller | Stochastic Gradient-Based Mixture Models for Transfer Modulation in Meta-Learning [paper] |
Kurt Cutajar, Mark Pullin, Andreas Damianou, Neil Lawrence and Javier Gonzalez | Deep Gaussian Processes for Multi-fidelity Modeling [paper] |
Eric Nalisnick and José Miguel Hérnandez-Lobato | Automatic Depth Determination for Bayesian ResNets [paper] |
Artem Sobolev and Dmitry Vetrov | Importance Weighted Hierarchical Variational Inference [paper] |
Kashyap Chitta, Jose M. Alvarez and Adam Lesnikowski | Deep Probabilistic Ensembles: Approximate Variational Inference through KL Regularization [paper] |
Dan Rosenbaum, Frederic Besse, Fabio Viola, Danilo Rezende and Ali Eslami | Learning models for visual 3D localization with implicit mapping [paper] |
Dieterich Lawson, George Tucker, Christian Naesseth, Chris Maddison, Ryan Adams and Yee Whye Teh | Twisted Variational Sequential Monte Carlo [paper] |
Mahdi Karami, Laurent Dinh, Daniel Duckworth, Jascha Sohl-Dickstein and Dale Schuurmans | Generative Convolutional Flow for Density Estimation [paper] |
Adam Kortylewski, Mario Wieser, Andreas Morel-Forster, Aleksander Wieczorek, Sonali Parbhoo, Volker Roth and Thomas Vetter | Informed MCMC with Bayesian Neural Networks for Facial Image Analysis [paper] |
Patrick Dallaire and Francois Laviolette | Bayesian Nonparametric Deep Learning [paper] |
Zihao Zhang, Stefan Zohren and Stephen Roberts | BDLOB: Bayesian Deep Convolutional Neural Networks for Limit Order Books [paper] |
Pierre-Alexandre Mattei and Jes Frellsen | missIWAE: Deep Generative Modelling and Imputation of Incomplete Data Sets [paper] |
Danielle Maddix, Yuyang Wang and Alex Smola | Deep Factors with Gaussian Processes for Forecasting [paper] |
Prasanna Sattigeri, Soumya Ghosh, Abhishek Kumar, Karthikeyan Ramamurthy, Samuel Hoffman, Youssef Drissi and Inkit Padhi | Probabilistic Mixture of Model-Agnostic Meta-Learners [paper] |
Frank Soboczenski, Michael D. Himes, Molly D. O'Beirne, Simone Zorzan, Atılım Günes Baydin, Adam D. Cobb, Daniel Angerhausen, Giada N. Arney and Shawn D. Domagal-Goldman | Bayesian Deep Learning for Exoplanet Atmospheric Retrieval [paper] |
Albert Shaw, Bo Dai, Weiyang Liu and Le Song | Bayesian Meta-network Architecture Learning [paper] |
Marcel Nassar, Xin Wang and Evren Tumer | Conditional Graph Neural Processes: A Functional Autoencoder Approach [paper] |
Victor Gallego and David Rios | Stochastic Gradient MCMC with Repulsive Forces [paper] |
Daniel Flam-Shepherd, James Requiema and David Duvenaud | Characterizing and Warping the Function space of Bayesian Neural Networks [paper] |
Qiang Zhang, Shangsong Liang and Emine Yilmaz | Variational Self-attention Model for Sentence Representation [paper] |
Weiwei Pan, Melanie Fernandez Pradier, Jiayu Yao, Finale Doshi-Velez and Soumya Ghosh | Projected BNNs: Avoiding weight-space pathologies by projecting neural network weights [paper] |
Buu Phan, Rick Salay, Krzysztof Czarnecki, Vahdat Abdelzad, Taylor Denouden and Sachin Vernekar | Calibrating Uncertainties in Object Localization Task [paper] |
Tim Georg Johann Rudner, Vincent Fortuin, Yee Whye Teh and Yarin Gal | On the Connection between Neural Processes and Gaussian Processes with Deep Kernels [paper] |
Kyle Cranmer, Stefan Gadatsch, Aishik Ghosh, Tobias Golling, Gilles Louppe, David Rousseau, Dalila Salamani and Graeme Stewart | Deep generative models for fast shower simulation in ATLAS [paper] |
Ziyin Liu, Junxiang Chen, Paul Pu Liang and Masahito Ueda | Relational Attention Networks via Fully-Connected Conditional Random Fields [paper] |
Adam Foster, Martin Jankowiak, Eli Bingham, Yee Whye Teh, Tom Rainforth and Noah Goodman | Variational Optimal Experiment Design: Efficient Automation of Adaptive Experiments [paper] |
Natasa Tagasovska and David Lopez-Paz | Frequentist uncertainty estimates for deep learning [paper] |
Valery Kharitonov, Dmitry Molchanov and Dmitry Vetrov | Variational Dropout via Empirical Bayes [paper] |
Thang Bui, Cuong Nguyen, Siddharth Swaroop and Richard Turner | Partitioned Variational Inference for Federated Bayesian Deep Learning [paper] |
Gregory Gundersen, Bianca Dumitrascu, Jordan Ash and Barbara Engelhardt | End-to-end training of deep probabilistic CCA for joint modeling of paired biomedical observations [paper] |
Beliz Gokkaya, Jessica Ai, Michael Tingley, Yonglong Zhang, Ning Dong, Thomas Jiang, Anitha Kubendran and Arun Kumar | Bayesian Neural Networks using HackPPL with Application to User Location State Prediction [paper] |
J. Jon Ryu, Young-Han Kim, Yoojin Choi, Mostafa El-Khamy and Jungwon Lee | Variational Inference via a Joint Latent Variable Model with Common Information Extraction [paper] |
Ziyin Liu, Hubert Tsai Yao-Hung Tsai, Makoto Yamada and Ruslan Salakhutdinov | Semi-Supervised Pairing via Basis-Sharing Wasserstein Matching Auto-Encoder [paper] |
Arunesh Mittal, Paul Sajda and John Paisley | Deep Bayesian Nonparametric Factor Analysis [paper] |
Sophie Burkhardt, Julia Siekiera and Stefan Kramer | Semi-Supervised Bayesian Active Learning for Text Classification [paper] |
Tal Kachman, Michal Moshkovitz and Michal Rosen-Zvi | Novel Uncertainty Framework for Deep Learning Ensembles [paper] |
Xuechen Li and Will Grathwohl | Training Glow with Constant Memory Cost [paper] |
Martin Jankowiak | Closed Form Variational Objectives For Bayesian Neural Networks with a Single Hidden Layer [paper] |
Belhal Karimi and Eric Moulines | MISSO: Minimization by Incremental Stochastic Surrogate for large-scale nonconvex Optimization [paper] |
Pashupati Hegde, Markus Heinonen, Harri Lähdesmäki and Samuel Kaski | Deep learning with differential Gaussian process flows [paper] |
Ryan Turner, Jane Hung, Jason Yosinski and Yunus Saatci | Metropolis-Hastings GANs [paper] |
Juhan Bae, Guodong Zhang and Roger Grosse | Eigenvalue Corrected Noisy Natural Gradient [paper] |
Cusuh Ham, Amit Raj, Vincent Cartillier and Irfan Essa | Variational Image Inpainting [paper] |
Jaehoon Lee, Lechao Xiao, Jascha Sohl-Dickstein and Jeffrey Pennington | Gaussian Predictions from Gradient Descent Training of Wide Neural Networks [paper] |
Eugene Golikov and Maksim Kretov | Embedding-reparameterization trick for manifold-valued latent variables in generative models [paper] |
Fábio Perez, Rémi Lebret and Karl Aberer | Cluster-Based Active Learning [paper] |
Maksim Kuznetsov, Daniil Polykovskiy, Dmitry Vetrov and Alexander Zhebrak | Subset-Conditioned Generation Using Variational Autoencoder With A Learnable Tensor-Train Induced Prior [paper] |
Remus Pop and Patric Fulop | Deep Ensemble Bayesian Active Learning [paper] |
Tuan Anh Le, Hyunjik Kim, Marta Garnelo, Dan Rosenbaum, Jonathan Schwarz and Yee Whye Teh | Empirical Evaluation of Neural Process Objectives [paper] |
Bin Dai and David Wipf | Diagnosing and Enhancing Gaussian VAE Models [paper] |
Jiaming Zeng, Adam Lesnikowski and Jose Alvarez | The Relevance of Bayesian Layer Positioning to Model Uncertainty in Deep Bayesian Active Learning [paper] |
Rui Zhao and Qiang Ji | An Empirical Evaluation of Bayesian Inference Methods for Bayesian Neural Networks [paper] |
Maithra Raghu, Katy Blumer, Rory Sayres, Ziad Obermeyer, Sendhil Mullainathan and Jon Kleinberg | Direct Uncertainty Prediction for Medical Second Opinions [paper] |
Ershad Banijamali, Amir-Hossein Karimi and Ali Ghodsi | Deep Variational Sufficient Dimensionality Reduction [paper] |
Chanwoo Park, Jae Myung Kim, Seok Hyeon Ha and Jungwoo Lee | A simple method for predictive uncertainty estimation using gradient uncertainty [paper] |
Luca Ambrogioni, Umut Guclu, Yagmur Gucluturk and Marcel van Gerven | Wasserstein Variational Gradient Descent: From Semi-Discrete Optimal Transport to Ensemble Variational Inference [paper] |
Kumar Sricharan, Kumar Saketh and Ashok Srivastava | Improving robustness of classifiers by training against live traffic [paper] |
Daniel Flam-Shepherd, Yuxiang Gao and Zhaoyu Guo | Stick-Breaking Neural Latent Variable Models [paper] |
Lavanya Sita Tekumalla, Priyanka Agrawal and Indrajit Bhattacharya | Deep Nested Hierarchical Dirichlet Processes [paper] |
Matt Benatan and Edward Pyzer-Knapp | Practical Considerations for Probabilistic Backpropagation [paper] |
Noah Weber, Janez Starc, Arpit Mittal, Roi Blanco and Lluis Marquez | Optimizing over a Bayesian Last Layer [paper] |
Mahmoud Elnaggar, Kamin Whitehouse and Cody Fleming | Bayesian Wireless Channel Prediction for Safety-Critical Connected Autonomous Vehicles [paper] |
Siddhartha Jain, Ge Liu and Jonas Mueller | Maximizing Overall Diversity to Control Out-of-Distribution Behavior of Deep Ensembles [paper] |
Siddhartha Jain and Nathan Hunt | Approximate Mutual Information-based Acquisition for General Models in Bayesian Optimization [paper] |
Navneet Madhu Kumar | Empowerment-driven Exploration using Mutual Information Estimation [paper] |
Yifeng Li and Xiaodan Zhu | Capsule Restricted Boltzmann Machine [paper] |
Matthew Willetts, Aiden Doherty, Stephen J. Roberts and Christopher C. Holmes | Semi-unsupervised Learning using Deep Generative Models [paper] |
Danil Kuzin, Olga Isupova and Lyudmila Mihaylova | Uncertainty propagation in neural networks for sparse coding [paper] |
Jovana Mitrovic, Peter Wirnsberger, Charles Blundell, Dino Sejdinovic and Yee Whye Teh | Infinitely Deep Infinite-Width Networks [paper] |
Daniel Park, Samuel Smith, Jascha Sohl-Dickstein and Quoc Le | Optimal SGD Hyperparameters for Fully Connected Networks [paper] |
Sambarta Dasgupta, Kumar Sricharan and Ashok Srivastava | Finite Rank Deep Kernel Learning [paper] |
Dustin Tran, Mike Dusenberry, Mark van der Wilk and Danijar Hafner | Bayesian Layers [paper] |
Michael Tschannen, Mario Lucic and Olivier Bachem | Recent Advances in Autoencoder-Based Representation Learning [paper] |
Mahmoud Hossam, Trung Le, Viet Huynh and Dinh Phung | Text Generation with Deep Variational GAN [paper] |
Lisha Chen and Qiang Ji | Kernel Density Network for Quantifying Uncertainty in Face Alignment [paper] |
Alexander Sagel and Martin Gottwald | Never Mind the Density, Here's the Level Set [paper] |
Yilun Du and Igor Mordatch | Implicit Generation and Representation Learning with Energy Based Models [paper] |
Brian Trippe, Jonathan Huggins and Tamara Broderick | Fast Bayesian Inference in GLMs with Low Rank Data Approximations [paper] |
Rita Kuznetsova, Oleg Bakhteev and Alexander Ogaltsov | Variational learning across domains with triplet information [paper] |