Presented By: University Career Center
Salesforce Research Speaker Series: Neural Language Modeling for Protein Design
Join Ali Madani in this weeks speaker series as he shares his work on Neural Language Modeling for Protein Design.
Abstract:
Transformer architectures have proven to learn useful representations for tasks in natural language processing (NLP). Taking a sequence-first perspective,we examine how transformer architectures can leverage the exponentially growing set of protein sequences that lack costly, structural annotations.In ProGen, we train a 1.2B-parameter language model that can generate relevant protein sequences as determined by a series of computational metrics. In ProVis, we analyze the inner workings of transformers to explore how the model discerns structural and functional properties.
RSVP Here: https://salesforce.recsolu.com/external/events/MbXRnk-K_Hco5y_AmHCYtw
Abstract:
Transformer architectures have proven to learn useful representations for tasks in natural language processing (NLP). Taking a sequence-first perspective,we examine how transformer architectures can leverage the exponentially growing set of protein sequences that lack costly, structural annotations.In ProGen, we train a 1.2B-parameter language model that can generate relevant protein sequences as determined by a series of computational metrics. In ProVis, we analyze the inner workings of transformers to explore how the model discerns structural and functional properties.
RSVP Here: https://salesforce.recsolu.com/external/events/MbXRnk-K_Hco5y_AmHCYtw
Explore Similar Events
-
Loading Similar Events...
