Zobrazeno 1 - 10
of 806
pro vyhledávání: '"Jaskowski P"'
Autor:
Powalski, Rafał, Klockiewicz, Bazyli, Jaśkowski, Maciej, Topolski, Bartosz, Dąbrowski-Tumański, Paweł, Wiśniewski, Maciej, Kuciński, Łukasz, Miłoś, Piotr, Plewczynski, Dariusz
Accelerating molecular docking -- the process of predicting how molecules bind to protein targets -- could boost small-molecule drug discovery and revolutionize medicine. Unfortunately, current molecular docking tools are too slow to screen potential
Externí odkaz:
http://arxiv.org/abs/2411.00004
The BERTScore metric is commonly used to evaluate automatic text simplification systems. However, current implementations of the metric fail to provide complete visibility into all information the metric can produce. Notably, the specific token match
Externí odkaz:
http://arxiv.org/abs/2409.17160
Autor:
Borchmann, Łukasz, Pietruszka, Michał, Jaśkowski, Wojciech, Jurkiewicz, Dawid, Halama, Piotr, Józiak, Paweł, Garncarek, Łukasz, Liskowski, Paweł, Szyndler, Karolina, Gretkowski, Andrzej, Ołtusek, Julita, Nowakowska, Gabriela, Zawłocki, Artur, Duhr, Łukasz, Dyda, Paweł, Turski, Michał
The vast portion of workloads employing LLMs involves answering questions grounded on PDF or scan content. We introduce the Arctic-TILT achieving accuracy on par with models 1000$\times$ its size on these use cases. It can be fine-tuned and deployed
Externí odkaz:
http://arxiv.org/abs/2408.04632
Publikováno v:
Archives of Civil Engineering, Vol 65, Iss 4, Pp 295-307 (2019)
Most construction projects involve subcontracting some work packages. A subcontractor is employed on the basis of their bid as well as according to their availability. A viable schedule must account for resource availability constraints. These resour
Externí odkaz:
https://doaj.org/article/41a39dbc5dec453da417bcffc372cabf
Publikováno v:
Archives of Civil Engineering, Vol vol. 70, Iss No 3, Pp 607-618 (2024)
Offsite construction technologies are developed to reduce project cost and duration. To make the most of the potential offered by prefabrication the planner should consider the whole supply chain. A failure to coordinate the off-site production with
Externí odkaz:
https://doaj.org/article/5ce4653cb10e45c79d6435987b40b85f
Autor:
Chrzanowicz Marcin, Jaskowski Piotr
Publikováno v:
MATEC Web of Conferences, Vol 231, p 04002 (2018)
The lighting parameters of advertising media observed from the road leading drivers are one of the basic factors which should be taken into account in the process of determining the influence of the carrier on the risk related to the possibility of a
Externí odkaz:
https://doaj.org/article/0ca854e4b67c453e8a6b6ee28ff9dd95
Publikováno v:
Archives of Civil Engineering, Vol vol. 69, Iss No 4, Pp 89-104 (2023)
The construction contractor is concerned with reducing the cost of the project, including reducing unnecessary downtime. This is achieved when resources are fully utilized; this means the crews work continuously moving without interruption from one l
Externí odkaz:
https://doaj.org/article/75a0d8b615e049e8a2382989011fe0fe
Autor:
Bonnie Jaskowski Huge, Caitlin M. Kerr, Sacheela Wanigasinghe, Matthew M. Champion, Norman J. Dovichi
Publikováno v:
Scientific Reports, Vol 13, Iss 1, Pp 1-11 (2023)
Abstract Capillary zone electrophoresis (CZE) is a powerful tool for high resolution chemical separations. Applying CZE to microbial samples may facilitate a deeper understanding of bacterial physiology and behavior. However, the study of complex mic
Externí odkaz:
https://doaj.org/article/a00868557b8148e58d12c658cfd12ca6
Autor:
D'Oro, Pierluca, Jaśkowski, Wojciech
Deterministic-policy actor-critic algorithms for continuous control improve the actor by plugging its actions into the critic and ascending the action-value gradient, which is obtained by chaining the actor's Jacobian matrix with the gradient of the
Externí odkaz:
http://arxiv.org/abs/2004.14309
Autor:
Srivastava, Rupesh Kumar, Shyam, Pranav, Mutz, Filipe, Jaśkowski, Wojciech, Schmidhuber, Jürgen
We develop Upside-Down Reinforcement Learning (UDRL), a method for learning to act using only supervised learning techniques. Unlike traditional algorithms, UDRL does not use reward prediction or search for an optimal policy. Instead, it trains agent
Externí odkaz:
http://arxiv.org/abs/1912.02877