| Mlp-mixer: An all-mlp architecture for vision IO Tolstikhin, N Houlsby, A Kolesnikov, L Beyer, X Zhai, T Unterthiner, ... Advances in neural information processing systems 34, 24261-24272, 2021 | 2808 | 2021 |
| Wasserstein auto-encoders I Tolstikhin, O Bousquet, S Gelly, B Schoelkopf arXiv preprint arXiv:1711.01558, 426-433, 2017 | 1330 | 2017 |
| Adagan: Boosting generative models IO Tolstikhin, S Gelly, O Bousquet, CJ Simon-Gabriel, B Schölkopf Advances in neural information processing systems 30, 2017 | 282 | 2017 |
| Towards a learning theory of cause-effect inference D Lopez-Paz, K Muandet, B Schölkopf, I Tolstikhin International Conference on Machine Learning, 1452-1461, 2015 | 215 | 2015 |
| From optimal transport to generative modeling: the VEGAN cookbook O Bousquet, S Gelly, I Tolstikhin, CJ Simon-Gabriel, B Schoelkopf URL http://arxiv. org/abs/1705.07642, 2017 | 168 | 2017 |
| Minimax estimation of maximum mean discrepancy with radial kernels IO Tolstikhin, BK Sriperumbudur, B Schölkopf Advances in Neural Information Processing Systems 29, 2016 | 140 | 2016 |
| Predicting neural network accuracy from weights T Unterthiner, D Keysers, S Gelly, O Bousquet, I Tolstikhin arXiv preprint arXiv:2002.11448, 2020 | 95 | 2020 |
| PAC-Bayes-empirical-Bernstein inequality IO Tolstikhin, Y Seldin Advances in Neural Information Processing Systems 26, 2013 | 95 | 2013 |
| Minimax estimation of kernel mean embeddings I Tolstikhin, BK Sriperumbudur, K Mu Journal of Machine Learning Research 18 (86), 1-47, 2017 | 85 | 2017 |
| What do neural networks learn when trained with random labels? H Maennel, IM Alabdulmohsin, IO Tolstikhin, R Baldock, O Bousquet, ... Advances in Neural Information Processing Systems 33, 19693-19704, 2020 | 81 | 2020 |
| On the latent space of wasserstein auto-encoders PK Rubenstein, B Schoelkopf, I Tolstikhin arXiv preprint arXiv:1802.03761, 2018 | 56 | 2018 |
| Practical and consistent estimation of f-divergences P Rubenstein, O Bousquet, J Djolonga, C Riquelme, IO Tolstikhin Advances in Neural Information Processing Systems 32, 2019 | 49 | 2019 |
| Differentially private database release via kernel mean embeddings M Balog, I Tolstikhin, B Schölkopf International Conference on Machine Learning, 414-422, 2018 | 46 | 2018 |
| When can unlabeled data improve the learning rate? C Göpfert, S Ben-David, O Bousquet, S Gelly, I Tolstikhin, R Urner Conference on Learning Theory, 1500-1518, 2019 | 30 | 2019 |
| Learning disentangled representations with wasserstein auto-encoders PK Rubenstein, B Schölkopf, I Tolstikhin | 28 | 2018 |
| Genet: Deep representations for metagenomics M Rojas-Carulla, I Tolstikhin, G Luque, N Youngblut, R Ley, B Schölkopf arXiv preprint arXiv:1901.11015, 2019 | 25 | 2019 |
| Competitive training of mixtures of independent deep generative models F Locatello, D Vincent, I Tolstikhin, G Rätsch, S Gelly, B Schölkopf arXiv preprint arXiv:1804.11130, 2018 | 23 | 2018 |
| Localized complexities for transductive learning I Tolstikhin, G Blanchard, M Kloft Conference on Learning Theory, 857-884, 2014 | 21 | 2014 |
| Permutational Rademacher complexity: a new complexity measure for transductive learning I Tolstikhin, N Zhivotovskiy, G Blanchard Algorithmic Learning Theory: 26th International Conference, ALT 2015, Banff …, 2015 | 15 | 2015 |
| Wasserstein auto-encoders: Latent dimensionality and random encoders PK Rubenstein, B Schölkopf, I Tolstikhin | 13 | 2018 |
Bernhard SchölkopfDirector, Max Planck Institute for Intelligent Systems & ELLIS Institute Tübingen; Professor at ETHVerified email at tuebingen.mpg.de