Paul K. Rubenstein
Max Planck Society
23 Papers
216 Citations
Paul K. Rubenstein is an academic researcher from Max Planck Society. The author has contributed to research in topics: Computer science & Causal model. The author has an hindex of 11, co-authored 20 publications.
Chat about Author
Papers
•Proceedings Article
On Mutual Information Maximization for Representation Learning
Michael Tschannen,Josip Djolonga,Paul K. Rubenstein,Sylvain Gelly,Mario Lucic +4 more
- 30 Apr 2020
TL;DR: This paper argues, and provides empirical evidence, that the success of these methods cannot be attributed to the properties of MI alone, and that they strongly depend on the inductive bias in both the choice of feature extractor architectures and the parametrization of the employed MI estimators.
•Posted Content
On Mutual Information Maximization for Representation Learning
TL;DR: The authors argue that the success of these methods cannot be attributed to the properties of MI alone, and that they strongly depend on the inductive bias in both the choice of feature extractor architectures and the parametrization of the employed MI estimators.
263
AudioPaLM: A Large Language Model That Can Speak and Listen
Paul K. Rubenstein,Chulayuth Asawaroengchai,DucDung Nguyen,Ankur Bapna,Zalán Borsos,Felix de Chaumont Quitry,Peter Chen,Dalia El Badawy,Weimin Huang,Eugene Kharitonov,Hannah Muckenhirn,Dirk R. Padfield,James Qin,Daniel Rozenberg,Tara N. Sainath,Johan Schalkwyk,Matthew Sharifi,Michelle D. Tadmor,Marco Tagliasacchi,Alexandru Tudor,Damien Vincent,Jiahui Yu,Yongqiang Wang,Vicky Zayats,Neil Zeghidour,Yu Zhang,Zhishuai Zhang,Lukas Zilka,Christian Frank +28 more
TL;DR: AudioPaLM as discussed by the authors fuses text-based and speech-based language models into a unified multimodal architecture that can process and generate text and speech with applications including speech recognition and speech to speech translation.
137
•Posted Content
Causal Consistency of Structural Equation Models
Paul K. Rubenstein,Sebastian Weichwald,Stephan Bongers,Joris M. Mooij,Dominik Janzing,Moritz Grosse-Wentrup,Bernhard Schölkopf +6 more
TL;DR: This work formalises the notion of consistency in the case of Structural Equation Models (SEMs) by introducing exact transformations between SEMs, which provides a general language to consider the different levels of description in the following three scenarios.
96
•Posted Content
On the Latent Space of Wasserstein Auto-Encoders.
TL;DR: It is argued that random encoders should be preferred over deterministicEncoders in Wasserstein auto-encoders with promising results on a benchmark disentanglement task.
52