Zobrazeno 1 - 10
of 16 117
pro vyhledávání: '"Leech, A"'
Evaluating modern ML models is hard. The strong incentive for researchers and companies to report a state-of-the-art result on some metric often leads to questionable research practices (QRPs): bad practices which fall short of outright research frau
Externí odkaz:
http://arxiv.org/abs/2407.12220
Autor:
Royer, Jessica, Paquola, Casey, Valk, Sofie L., Kirschner, Matthias, Hong, Seok-Jun, Park, Bo-yong, Bethlehem, Richard A. I., Leech, Robert, Yeo, B. T. Thomas, Jefferies, Elizabeth, Smallwood, Jonathan, Margulies, Daniel, Bernhardt, Boris C.
Multimodal neuroimaging grants a powerful in vivo window into the structure and function of the human brain. Recent methodological and conceptual advances have enabled investigations of the interplay between large-scale spatial trends, or gradients,
Externí odkaz:
http://arxiv.org/abs/2402.11055
We explore the AI2050 "hard problems" that block the promise of AI and cause AI risks: (1) developing general capabilities of the systems; (2) assuring the performance of AI systems and their training processes; (3) aligning system goals with human g
Externí odkaz:
http://arxiv.org/abs/2402.04464
Autor:
Samyogita Hardikar, Brontë McKeown, Adam Turnbull, Ting Xu, Sofie L. Valk, Boris C. Bernhardt, Daniel S. Margulies, Michael P. Milham, Elizabeth Jefferies, Robert Leech, Arno Villringer, Jonathan Smallwood
Publikováno v:
Communications Biology, Vol 7, Iss 1, Pp 1-8 (2024)
Abstract Human cognition supports complex behaviour across a range of situations, and traits (e.g. personality) influence how we react in these different contexts. Although viewing traits as situationally grounded is common in social sciences, often
Externí odkaz:
https://doaj.org/article/cbf1d1952d1144d69fd076386bc6182c
Autor:
Turner, Alexander Matt, Thiergart, Lisa, Leech, Gavin, Udell, David, Vazquez, Juan J., Mini, Ulisse, MacDiarmid, Monte
Prompt engineering and finetuning aim to maximize language model performance on a given metric (like toxicity reduction). However, these methods do not fully elicit a model's capabilities. To reduce this gap, we introduce activation engineering: the
Externí odkaz:
http://arxiv.org/abs/2308.10248
Autor:
Leech, Amy Joyce
Focusing on novels by three contemporary German authors and one multi-author theater text, “A Multidirectional Europe” investigates how their writing responds to post-1989 memory paradigms in which post-socialist memory, in relation to the Holoca
Publikováno v:
International Journal of Prison Health, 2024, Vol. 20, Issue 3, pp. 271-285.
Externí odkaz:
http://www.emeraldinsight.com/doi/10.1108/IJOPH-05-2023-0024
Reweighted wake-sleep (RWS) is a machine learning method for performing Bayesian inference in a very general class of models. RWS draws $K$ samples from an underlying approximate posterior, then uses importance weighting to provide a better estimate
Externí odkaz:
http://arxiv.org/abs/2305.11022
Autor:
Valerio Morlacci, Marco Milia, Jérémy Saiter, Irene Preet Bhela, Matthew C. Leech, Kevin Lam
Publikováno v:
JACS Au, Vol 4, Iss 11, Pp 4199-4205 (2024)
Externí odkaz:
https://doaj.org/article/aed068a1aec44b92bfa4ebf6ae9b321f