Zobrazeno 1 - 10
of 83
pro vyhledávání: '"Zając Michał"'
Autor:
Dąbrowski, Karol M., Nooraiepour, Mohammad, Masoudi, Mohammad, Zając, Michał, Kuczyński, Szymon, Smulski, Rafał, Barbacki, Jan, Hellevang, Helge, Nagy, Stanisław
Salt precipitation triggered by the evaporation of formation brine into injected supercritical CO2 can cause injectivity and containment issues in near-wellbore regions. Predicting the distribution of precipitated salts and their impact on near-wellb
Externí odkaz:
http://arxiv.org/abs/2410.04910
Autor:
Howe, Nikolaus, McKenzie, Ian, Hollinsworth, Oskar, Zajac, Michał, Tseng, Tom, Tucker, Aaron, Bacon, Pierre-Luc, Gleave, Adam
Language models exhibit scaling laws, whereby increasing model and dataset size yields predictable decreases in negative log likelihood, unlocking a dazzling array of capabilities. This phenomenon spurs many companies to train ever larger models in p
Externí odkaz:
http://arxiv.org/abs/2407.18213
Autor:
Wołczyk, Maciej, Cupiał, Bartłomiej, Ostaszewski, Mateusz, Bortkiewicz, Michał, Zając, Michał, Pascanu, Razvan, Kuciński, Łukasz, Miłoś, Piotr
Fine-tuning is a widespread technique that allows practitioners to transfer pre-trained capabilities, as recently showcased by the successful applications of foundation models. However, fine-tuning reinforcement learning (RL) models remains a challen
Externí odkaz:
http://arxiv.org/abs/2402.02868
Language model attacks typically assume one of two extreme threat models: full white-box access to model weights, or black-box access limited to a text generation API. However, real-world APIs are often more flexible than just text generation: these
Externí odkaz:
http://arxiv.org/abs/2312.14302
Publikováno v:
E3S Web of Conferences, Vol 173, p 03005 (2020)
One of waste heat recovery technologies are thermoelectric generators, which allow direct conversion heat to energy. In the face of shrinking fossil fuels resources, simultaneous increase of global energy demand and level of pollution it is becoming
Externí odkaz:
https://doaj.org/article/47a491b6b7bc4c9f8c3cbacfd4159058
Publikováno v:
E3S Web of Conferences, Vol 173, p 02002 (2020)
During last years, renewable energy sources (RES) find their way into the transportation industry. Among the units which may be powered directly with renewable energy, the UAVs (unmanned aerial vehicles) market is undergoing a rapid development. In t
Externí odkaz:
https://doaj.org/article/2fc6da8b732e4e7d9c637118a2bdcd6e
Class-incremental learning (CIL) is a particularly challenging variant of continual learning, where the goal is to learn to discriminate between all classes presented in an incremental fashion. Existing approaches often suffer from excessive forgetti
Externí odkaz:
http://arxiv.org/abs/2305.18806
Autor:
Zając, Michał, Deja, Kamil, Kuzina, Anna, Tomczak, Jakub M., Trzciński, Tomasz, Shkurti, Florian, Miłoś, Piotr
Diffusion models have achieved remarkable success in generating high-quality images thanks to their novel training procedures applied to unprecedented amounts of data. However, training a diffusion model from scratch is computationally expensive. Thi
Externí odkaz:
http://arxiv.org/abs/2303.15342
Autor:
Olko, Mateusz, Zając, Michał, Nowak, Aleksandra, Scherrer, Nino, Annadani, Yashas, Bauer, Stefan, Kuciński, Łukasz, Miłoś, Piotr
Inferring causal structure from data is a challenging task of fundamental importance in science. Observational data are often insufficient to identify a system's causal structure uniquely. While conducting interventions (i.e., experiments) can improv
Externí odkaz:
http://arxiv.org/abs/2211.13715
The ability of continual learning systems to transfer knowledge from previously seen tasks in order to maximize performance on new tasks is a significant challenge for the field, limiting the applicability of continual learning solutions to realistic
Externí odkaz:
http://arxiv.org/abs/2209.13900