26.09.2024Only affiliations of IDEAS NCBR is noted marked below:
1. Repurposing Language Models into Embedding Models: Finding the Compute-Optimal Recipe
https://arxiv.org/abs/2406.04165
Alicja Ziarko (IDEAS NCBR, University of Warsaw, Institute of Mathematics, Polish Academy of Sciences), Albert Q. Jiang, Bartosz Piotrowski (IDEAS NCBR), Wenda Li, Mateja Jamnik, Piotr Miłoś (IDEAS NCBR, University of Warsaw, Institute of Mathematics, Polish Academy of Sciences, deepsense.ai)
2. Bigger, Regularized, Optimistic: scaling for compute and sample-efficient continuous control
[the spotlight paper]
https://arxiv.org/abs/2405.16158
Michal Nauman (IDEAS NCBR, Faculty of Mathematics, Informatics, and Mechanics; University of Warsaw), Mateusz Ostaszewski, Krzysztof Jankowski, Piotr Miłoś (IDEAS NCBR, Faculty of Mathematics, Informatics, and Mechanics; University of Warsaw, Polish Academy of Sciences), Marek Cygan
3. Mixture of Tokens: Continuous MoE through Cross-Example Aggregation
https://arxiv.org/abs/2310.15961
Szymon Antoniak, Michał Krutul (IDEAS NCBR, University of Warsaw), Maciej Pióro (IDEAS NCBR, Polish Academy of Sciences), Jakub Krajewski (IDEAS NCBR, University of Warsaw), Jan Ludziejewski (IDEAS NCBR, University of Warsaw), Kamil Ciebiera, Krystian Król, Tomasz Odrzygóźdź, Marek Cygan, Sebastian Jaszczur (IDEAS NCBR, University of Warsaw)
4. Exploiting Activation Sparsity with Dense to Dynamic-k Mixture-of-Experts Conversion
https://arxiv.org/abs/2310.04361
Filip Szatkowski (IDEAS NCBR, Warsaw University of Technology), Bartosz Wójcik (IDEAS NCBR, Jagiellonian University), Mikołaj Piórczyński, Simone Scardapane
5. D-MiSo: Editing Dynamic 3D Scenes using Multi-Gaussians Soup
https://arxiv.org/abs/2405.14276
Joanna Waczyńska, Piotr Borycki, Joanna Kaleta, Sławomir Tadeja, Przemysław Spurek (IDEAS NCBR, Jagiellonian University)
6. Task-recency bias strikes back: Adapting covariances in Exemplar-Free Class Incremental Learning
https://arxiv.org/abs/2409.18265
Grzegorz Rypeść (Warsaw University of Technology), Sebastian Cygert (Gdańsk University of Technology), Tomasz Trzciński (Warsaw University of Technology, Tooploox), Bartłomiej Twardowski (Autonomous University of Barcelona, Computer Vision Center)
(all also IDEAS NCBR)
7. Benchmarking Structural Inference Methods for Interacting Dynamical Systems with Synthetic Data
https://openreview.net/forum?id=PCXvcULwiI
Aoran Wang, Tsz Pan Tong, Andrzej Mizera (IDEAS NCBR), Jun Pang
