Zobrazeno 1 - 10
of 1 158
pro vyhledávání: '"Sherborne, A."'
Autor:
Na, Clara, Magnusson, Ian, Jha, Ananya Harsh, Sherborne, Tom, Strubell, Emma, Dodge, Jesse, Dasigi, Pradeep
Training data compositions for Large Language Models (LLMs) can significantly affect their downstream performance. However, a thorough data ablation study exploring large sets of candidate data mixtures is typically prohibitively expensive since the
Externí odkaz:
http://arxiv.org/abs/2410.15661
Autor:
Matton, Alexandre, Sherborne, Tom, Aumiller, Dennis, Tommasone, Elena, Alizadeh, Milad, He, Jingyi, Ma, Raymond, Voisin, Maxime, Gilsenan-McMahon, Ellen, Gallé, Matthias
In this paper, we consider contamination by code generation test sets, in particular in their use in modern large language models. We discuss three possible sources of such contamination and show findings supporting each of them: (i) direct data leak
Externí odkaz:
http://arxiv.org/abs/2407.07565
Sharpness-aware minimization (SAM) reports improving domain generalization by reducing the loss surface curvature in the parameter space. However, generalization during fine-tuning is often more dependent on the transferability of representations in
Externí odkaz:
http://arxiv.org/abs/2310.03646
Autor:
Peng, Hao, Cao, Qingqing, Dodge, Jesse, Peters, Matthew E., Fernandez, Jared, Sherborne, Tom, Lo, Kyle, Skjonsberg, Sam, Strubell, Emma, Plessas, Darrell, Beltagy, Iz, Walsh, Evan Pete, Smith, Noah A., Hajishirzi, Hannaneh
Rising computational demands of modern natural language processing (NLP) systems have increased the barrier to entry for cutting-edge research while posing serious environmental concerns. Yet, progress on model efficiency has been impeded by practica
Externí odkaz:
http://arxiv.org/abs/2307.09701
Cross-lingual semantic parsing transfers parsing capability from a high-resource language (e.g., English) to low-resource languages with scarce training data. Previous work has primarily considered silver-standard data augmentation or zero-shot metho
Externí odkaz:
http://arxiv.org/abs/2307.04096
Autor:
Jha, Ananya Harsh, Sherborne, Tom, Walsh, Evan Pete, Groeneveld, Dirk, Strubell, Emma, Beltagy, Iz
Large language models (LLMs) enable unparalleled few- and zero-shot reasoning capabilities but at a high computational footprint. A growing assortment of methods for compression promises to reduce the computational burden of LLMs in deployment, but s
Externí odkaz:
http://arxiv.org/abs/2305.14864
Automatic machine translation (MT) metrics are widely used to distinguish the translation qualities of machine translation systems across relatively large test sets (system-level evaluation). However, it is unclear if automatic metrics are reliable a
Externí odkaz:
http://arxiv.org/abs/2212.10297
Autor:
Sherborne, Tom, Lapata, Mirella
Localizing a semantic parser to support new languages requires effective cross-lingual generalization. Recent work has found success with machine-translation or zero-shot methods although these approaches can struggle to model how native speakers ask
Externí odkaz:
http://arxiv.org/abs/2209.12577
Autor:
Forbes, Matthew S., Marx, Samuel K., Cohen, Tim J., Sherborne-Higgins, Bryce, Francke, Alexander, Peleckis, Germanas, Jones, Brian G., Dosseto, Anthony, Cadd, Haidee, Swallow, Elizabeth, Raven, Mark, Cendón, Dioni I., Peterson, Mark A.
Publikováno v:
In Applied Geochemistry October 2024 172
Autor:
Arun Chandramohan, Hubert Josien, Tsz Ying Yuen, Ruchia Duggal, Diana Spiegelberg, Lin Yan, Yu-Chi Angela Juang, Lan Ge, Pietro G. Aronica, Hung Yi Kristal Kaan, Yee Hwee Lim, Andrea Peier, Brad Sherborne, Jerome Hochman, Songnian Lin, Kaustav Biswas, Marika Nestor, Chandra S. Verma, David P. Lane, Tomi K. Sawyer, Robert Garbaccio, Brian Henry, Srinivasaraghavan Kannan, Christopher J. Brown, Charles W. Johannes, Anthony W. Partridge
Publikováno v:
Nature Communications, Vol 15, Iss 1, Pp 1-19 (2024)
Abstract Although stapled α-helical peptides can address challenging targets, their advancement is impeded by poor understandings for making them cell permeable while avoiding off-target toxicities. By synthesizing >350 molecules, we present workflo
Externí odkaz:
https://doaj.org/article/244b34e3aef44b508696598d0927724f