Can Transformers Learn Sequential Function Classes In Context?

Autor: Campbell, Ryan, Guo, Emma, Hu, Evan, Vir, Reya, Hsiao, Ethan

In-context learning (ICL) has revolutionized the capabilities of transformer models in NLP. In our project, we extend the understanding of the mechanisms underpinning ICL by exploring whether transformers can learn from sequential, non-textual functi

Externí odkaz: http://arxiv.org/abs/2312.12655

Tento výsledek nelze pro nepřihlášené uživatele zobrazit.
K zobrazení výsledku je třeba se přihlásit.

JBrowse Jupyter: a Python interface to JBrowse 2.

Autor: De Jesus Martinez T; Department of Bioengineering, Stanley Hall, University of California, Berkeley, CA 94720, USA., Hershberg EA; Department of Bioengineering, Stanley Hall, University of California, Berkeley, CA 94720, USA., Guo E; Department of Bioengineering, Stanley Hall, University of California, Berkeley, CA 94720, USA., Stevens GJ; Department of Bioengineering, Stanley Hall, University of California, Berkeley, CA 94720, USA., Diesh C; Department of Bioengineering, Stanley Hall, University of California, Berkeley, CA 94720, USA., Xie P; Department of Bioengineering, Stanley Hall, University of California, Berkeley, CA 94720, USA., Bridge C; Adaptive Oncology, Ontario Institute for Cancer Research, MaRS Centre, 661 University Avenue, Suite 510, Toronto, ON M5G 0A3, Canada., Cain S; Adaptive Oncology, Ontario Institute for Cancer Research, MaRS Centre, 661 University Avenue, Suite 510, Toronto, ON M5G 0A3, Canada., Haw R; Adaptive Oncology, Ontario Institute for Cancer Research, MaRS Centre, 661 University Avenue, Suite 510, Toronto, ON M5G 0A3, Canada., Buels RM; Department of Bioengineering, Stanley Hall, University of California, Berkeley, CA 94720, USA., Stein LD; Adaptive Oncology, Ontario Institute for Cancer Research, MaRS Centre, 661 University Avenue, Suite 510, Toronto, ON M5G 0A3, Canada., Holmes IH; Department of Bioengineering, Stanley Hall, University of California, Berkeley, CA 94720, USA.

Publikováno v: Bioinformatics (Oxford, England) [Bioinformatics] 2023 Jan 01; Vol. 39 (1).