Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox
Picture for Mario Lucic

Mario Lucic

Underspecification Presents Challenges for Credibility in Modern Machine Learning


Nov 06, 2020
Alexander D'Amour, Katherine Heller, Dan Moldovan, Ben Adlam, Babak Alipanahi, Alex Beutel, Christina Chen, Jonathan Deaton, Jacob Eisenstein, Matthew D. Hoffman, Farhad Hormozdiari, Neil Houlsby, Shaobo Hou, Ghassen Jerfel, Alan Karthikesalingam, Mario Lucic, Yian Ma, Cory McLean, Diana Mincu, Akinori Mitani, Andrea Montanari, Zachary Nado, Vivek Natarajan, Christopher Nielson, Thomas F. Osborne, Rajiv Raman, Kim Ramasamy, Rory Sayres, Jessica Schrouff, Martin Seneviratne, Shannon Sequeira, Harini Suresh, Victor Veitch, Max Vladymyrov, Xuezhi Wang, Kellie Webster, Steve Yadlowsky, Taedong Yun, Xiaohua Zhai, D. Sculley

Figure 1 for Underspecification Presents Challenges for Credibility in Modern Machine Learning
Figure 2 for Underspecification Presents Challenges for Credibility in Modern Machine Learning
Figure 3 for Underspecification Presents Challenges for Credibility in Modern Machine Learning
Figure 4 for Underspecification Presents Challenges for Credibility in Modern Machine Learning

  Access Paper or Ask Questions

A Sober Look at the Unsupervised Learning of Disentangled Representations and their Evaluation


Oct 27, 2020
Francesco Locatello, Stefan Bauer, Mario Lucic, Gunnar Rätsch, Sylvain Gelly, Bernhard Schölkopf, Olivier Bachem

Figure 1 for A Sober Look at the Unsupervised Learning of Disentangled Representations and their Evaluation
Figure 2 for A Sober Look at the Unsupervised Learning of Disentangled Representations and their Evaluation
Figure 3 for A Sober Look at the Unsupervised Learning of Disentangled Representations and their Evaluation
Figure 4 for A Sober Look at the Unsupervised Learning of Disentangled Representations and their Evaluation
* Journal of Machine Learning Research 2020, Volume 21, Number 209 
 * arXiv admin note: substantial text overlap with arXiv:1811.12359 
  Access Paper or Ask Questions

Which Model to Transfer? Finding the Needle in the Growing Haystack


Oct 13, 2020
Cedric Renggli, André Susano Pinto, Luka Rimanic, Joan Puigcerver, Carlos Riquelme, Ce Zhang, Mario Lucic

Figure 1 for Which Model to Transfer? Finding the Needle in the Growing Haystack
Figure 2 for Which Model to Transfer? Finding the Needle in the Growing Haystack
Figure 3 for Which Model to Transfer? Finding the Needle in the Growing Haystack
Figure 4 for Which Model to Transfer? Finding the Needle in the Growing Haystack

  Access Paper or Ask Questions

Representation learning from videos in-the-wild: An object-centric approach


Oct 06, 2020
Rob Romijnders, Aravindh Mahendran, Michael Tschannen, Josip Djolonga, Marvin Ritter, Neil Houlsby, Mario Lucic

Figure 1 for Representation learning from videos in-the-wild: An object-centric approach
Figure 2 for Representation learning from videos in-the-wild: An object-centric approach
Figure 3 for Representation learning from videos in-the-wild: An object-centric approach
Figure 4 for Representation learning from videos in-the-wild: An object-centric approach

  Access Paper or Ask Questions

A Commentary on the Unsupervised Learning of Disentangled Representations


Jul 28, 2020
Francesco Locatello, Stefan Bauer, Mario Lucic, Gunnar Rätsch, Sylvain Gelly, Bernhard Schölkopf, Olivier Bachem

Figure 1 for A Commentary on the Unsupervised Learning of Disentangled Representations
Figure 2 for A Commentary on the Unsupervised Learning of Disentangled Representations
* The Thirty-Fourth AAAI Conference on Artificial Intelligence 2020 (AAAI-20) 
  Access Paper or Ask Questions

On Robustness and Transferability of Convolutional Neural Networks


Jul 16, 2020
Josip Djolonga, Jessica Yung, Michael Tschannen, Rob Romijnders, Lucas Beyer, Alexander Kolesnikov, Joan Puigcerver, Matthias Minderer, Alexander D'Amour, Dan Moldovan, Sylvan Gelly, Neil Houlsby, Xiaohua Zhai, Mario Lucic

Figure 1 for On Robustness and Transferability of Convolutional Neural Networks
Figure 2 for On Robustness and Transferability of Convolutional Neural Networks
Figure 3 for On Robustness and Transferability of Convolutional Neural Networks
Figure 4 for On Robustness and Transferability of Convolutional Neural Networks

  Access Paper or Ask Questions

Self-Supervised Learning of Video-Induced Visual Invariances


Dec 05, 2019
Michael Tschannen, Josip Djolonga, Marvin Ritter, Aravindh Mahendran, Neil Houlsby, Sylvain Gelly, Mario Lucic

Figure 1 for Self-Supervised Learning of Video-Induced Visual Invariances
Figure 2 for Self-Supervised Learning of Video-Induced Visual Invariances
Figure 3 for Self-Supervised Learning of Video-Induced Visual Invariances
Figure 4 for Self-Supervised Learning of Video-Induced Visual Invariances

  Access Paper or Ask Questions

Semantic Bottleneck Scene Generation


Nov 26, 2019
Samaneh Azadi, Michael Tschannen, Eric Tzeng, Sylvain Gelly, Trevor Darrell, Mario Lucic

Figure 1 for Semantic Bottleneck Scene Generation
Figure 2 for Semantic Bottleneck Scene Generation
Figure 3 for Semantic Bottleneck Scene Generation
Figure 4 for Semantic Bottleneck Scene Generation

  Access Paper or Ask Questions

The Visual Task Adaptation Benchmark


Oct 01, 2019
Xiaohua Zhai, Joan Puigcerver, Alexander Kolesnikov, Pierre Ruyssen, Carlos Riquelme, Mario Lucic, Josip Djolonga, Andre Susano Pinto, Maxim Neumann, Alexey Dosovitskiy, Lucas Beyer, Olivier Bachem, Michael Tschannen, Marcin Michalski, Olivier Bousquet, Sylvain Gelly, Neil Houlsby

Figure 1 for The Visual Task Adaptation Benchmark
Figure 2 for The Visual Task Adaptation Benchmark
Figure 3 for The Visual Task Adaptation Benchmark
Figure 4 for The Visual Task Adaptation Benchmark

  Access Paper or Ask Questions

On Mutual Information Maximization for Representation Learning


Jul 31, 2019
Michael Tschannen, Josip Djolonga, Paul K. Rubenstein, Sylvain Gelly, Mario Lucic

Figure 1 for On Mutual Information Maximization for Representation Learning
Figure 2 for On Mutual Information Maximization for Representation Learning
Figure 3 for On Mutual Information Maximization for Representation Learning
Figure 4 for On Mutual Information Maximization for Representation Learning
* Michael Tschannen and Josip Djolonga contributed equally 
  Access Paper or Ask Questions

Evaluating Generative Models Using Divergence Frontiers


May 26, 2019
Josip Djolonga, Mario Lucic, Marco Cuturi, Olivier Bachem, Olivier Bousquet, Sylvain Gelly

Figure 1 for Evaluating Generative Models Using Divergence Frontiers
Figure 2 for Evaluating Generative Models Using Divergence Frontiers
Figure 3 for Evaluating Generative Models Using Divergence Frontiers
Figure 4 for Evaluating Generative Models Using Divergence Frontiers

  Access Paper or Ask Questions

High-Fidelity Image Generation With Fewer Labels


Mar 06, 2019
Mario Lucic, Michael Tschannen, Marvin Ritter, Xiaohua Zhai, Olivier Bachem, Sylvain Gelly

Figure 1 for High-Fidelity Image Generation With Fewer Labels
Figure 2 for High-Fidelity Image Generation With Fewer Labels
Figure 3 for High-Fidelity Image Generation With Fewer Labels
Figure 4 for High-Fidelity Image Generation With Fewer Labels
* Mario Lucic, Michael Tschannen, and Marvin Ritter contributed equally to this work 
  Access Paper or Ask Questions

Recent Advances in Autoencoder-Based Representation Learning


Dec 12, 2018
Michael Tschannen, Olivier Bachem, Mario Lucic

Figure 1 for Recent Advances in Autoencoder-Based Representation Learning
Figure 2 for Recent Advances in Autoencoder-Based Representation Learning
Figure 3 for Recent Advances in Autoencoder-Based Representation Learning
Figure 4 for Recent Advances in Autoencoder-Based Representation Learning
* Presented at the third workshop on Bayesian Deep Learning (NeurIPS 2018) 
  Access Paper or Ask Questions

Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations


Dec 02, 2018
Francesco Locatello, Stefan Bauer, Mario Lucic, Sylvain Gelly, Bernhard Schölkopf, Olivier Bachem

Figure 1 for Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations
Figure 2 for Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations
Figure 3 for Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations
Figure 4 for Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations
* This is a preliminary preprint based on our initial experimental results 
  Access Paper or Ask Questions

Self-Supervised Generative Adversarial Networks


Nov 27, 2018
Ting Chen, Xiaohua Zhai, Marvin Ritter, Mario Lucic, Neil Houlsby

Figure 1 for Self-Supervised Generative Adversarial Networks
Figure 2 for Self-Supervised Generative Adversarial Networks
Figure 3 for Self-Supervised Generative Adversarial Networks
Figure 4 for Self-Supervised Generative Adversarial Networks

  Access Paper or Ask Questions

Are GANs Created Equal? A Large-Scale Study


Oct 29, 2018
Mario Lucic, Karol Kurach, Marcin Michalski, Sylvain Gelly, Olivier Bousquet

Figure 1 for Are GANs Created Equal? A Large-Scale Study
Figure 2 for Are GANs Created Equal? A Large-Scale Study
Figure 3 for Are GANs Created Equal? A Large-Scale Study
Figure 4 for Are GANs Created Equal? A Large-Scale Study
* NIPS'18: Added a section on the limitations of the study and additional empirical results 
  Access Paper or Ask Questions

Deep Generative Models for Distribution-Preserving Lossy Compression


Oct 28, 2018
Michael Tschannen, Eirikur Agustsson, Mario Lucic

Figure 1 for Deep Generative Models for Distribution-Preserving Lossy Compression
Figure 2 for Deep Generative Models for Distribution-Preserving Lossy Compression
Figure 3 for Deep Generative Models for Distribution-Preserving Lossy Compression
Figure 4 for Deep Generative Models for Distribution-Preserving Lossy Compression
* NIPS 2018. Code: https://github.com/mitscha/dplc . Changes w.r.t. v1: Some clarifications in the text and additional numerical results 
  Access Paper or Ask Questions

Assessing Generative Models via Precision and Recall


Oct 28, 2018
Mehdi S. M. Sajjadi, Olivier Bachem, Mario Lucic, Olivier Bousquet, Sylvain Gelly

Figure 1 for Assessing Generative Models via Precision and Recall
Figure 2 for Assessing Generative Models via Precision and Recall
Figure 3 for Assessing Generative Models via Precision and Recall
* NIPS 2018 
  Access Paper or Ask Questions

The GAN Landscape: Losses, Architectures, Regularization, and Normalization


Oct 26, 2018
Karol Kurach, Mario Lucic, Xiaohua Zhai, Marcin Michalski, Sylvain Gelly

Figure 1 for The GAN Landscape: Losses, Architectures, Regularization, and Normalization
Figure 2 for The GAN Landscape: Losses, Architectures, Regularization, and Normalization
Figure 3 for The GAN Landscape: Losses, Architectures, Regularization, and Normalization
Figure 4 for The GAN Landscape: Losses, Architectures, Regularization, and Normalization
* Changed formatting from ICML workshop to ICLR. We added additional resnet ablation studies, hinge loss, and an empirical comparison between KID and FID 
  Access Paper or Ask Questions

On Self Modulation for Generative Adversarial Networks


Oct 02, 2018
Ting Chen, Mario Lucic, Neil Houlsby, Sylvain Gelly

Figure 1 for On Self Modulation for Generative Adversarial Networks
Figure 2 for On Self Modulation for Generative Adversarial Networks
Figure 3 for On Self Modulation for Generative Adversarial Networks
Figure 4 for On Self Modulation for Generative Adversarial Networks

  Access Paper or Ask Questions

Scalable k-Means Clustering via Lightweight Coresets


Jun 06, 2018
Olivier Bachem, Mario Lucic, Andreas Krause

Figure 1 for Scalable k-Means Clustering via Lightweight Coresets
Figure 2 for Scalable k-Means Clustering via Lightweight Coresets
Figure 3 for Scalable k-Means Clustering via Lightweight Coresets
* To appear in the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD) 
  Access Paper or Ask Questions

One-Shot Coresets: The Case of k-Clustering


Feb 20, 2018
Olivier Bachem, Mario Lucic, Silvio Lattanzi

* To Appear In AISTATS 2018 
  Access Paper or Ask Questions

Training Gaussian Mixture Models at Scale via Coresets


Jan 15, 2018
Mario Lucic, Matthew Faulkner, Andreas Krause, Dan Feldman

Figure 1 for Training Gaussian Mixture Models at Scale via Coresets
Figure 2 for Training Gaussian Mixture Models at Scale via Coresets
Figure 3 for Training Gaussian Mixture Models at Scale via Coresets
Figure 4 for Training Gaussian Mixture Models at Scale via Coresets

  Access Paper or Ask Questions

Stochastic Submodular Maximization: The Case of Coverage Functions


Nov 05, 2017
Mohammad Reza Karimi, Mario Lucic, Hamed Hassani, Andreas Krause

Figure 1 for Stochastic Submodular Maximization: The Case of Coverage Functions
Figure 2 for Stochastic Submodular Maximization: The Case of Coverage Functions
* 31st Conference on Neural Information Processing Systems (NIPS 2017) 
  Access Paper or Ask Questions

Practical Coreset Constructions for Machine Learning


Jun 04, 2017
Olivier Bachem, Mario Lucic, Andreas Krause

Figure 1 for Practical Coreset Constructions for Machine Learning
Figure 2 for Practical Coreset Constructions for Machine Learning

  Access Paper or Ask Questions

Uniform Deviation Bounds for Unbounded Loss Functions like k-Means


Feb 27, 2017
Olivier Bachem, Mario Lucic, S. Hamed Hassani, Andreas Krause


  Access Paper or Ask Questions

Horizontally Scalable Submodular Maximization


May 31, 2016
Mario Lucic, Olivier Bachem, Morteza Zadimoghaddam, Andreas Krause

Figure 1 for Horizontally Scalable Submodular Maximization
Figure 2 for Horizontally Scalable Submodular Maximization
Figure 3 for Horizontally Scalable Submodular Maximization
Figure 4 for Horizontally Scalable Submodular Maximization

  Access Paper or Ask Questions

Tradeoffs for Space, Time, Data and Risk in Unsupervised Learning


May 02, 2016
Mario Lucic, Mesrob I. Ohannessian, Amin Karbasi, Andreas Krause

Figure 1 for Tradeoffs for Space, Time, Data and Risk in Unsupervised Learning
Figure 2 for Tradeoffs for Space, Time, Data and Risk in Unsupervised Learning
* Conference version 
  Access Paper or Ask Questions

Linear-time Outlier Detection via Sensitivity


May 02, 2016
Mario Lucic, Olivier Bachem, Andreas Krause

Figure 1 for Linear-time Outlier Detection via Sensitivity
Figure 2 for Linear-time Outlier Detection via Sensitivity
Figure 3 for Linear-time Outlier Detection via Sensitivity
Figure 4 for Linear-time Outlier Detection via Sensitivity
* 8 pages 
  Access Paper or Ask Questions

Strong Coresets for Hard and Soft Bregman Clustering with Applications to Exponential Family Mixtures


May 02, 2016
Mario Lucic, Olivier Bachem, Andreas Krause

Figure 1 for Strong Coresets for Hard and Soft Bregman Clustering with Applications to Exponential Family Mixtures
Figure 2 for Strong Coresets for Hard and Soft Bregman Clustering with Applications to Exponential Family Mixtures
Figure 3 for Strong Coresets for Hard and Soft Bregman Clustering with Applications to Exponential Family Mixtures
Figure 4 for Strong Coresets for Hard and Soft Bregman Clustering with Applications to Exponential Family Mixtures
* 14 pages, camera ready version 
  Access Paper or Ask Questions