Friday, 5th July - 11am Luca Soldaini : Seminar

Title:  OLMo: Accelerating the Science of Open Language Models

Abstract:

Recently, we have seen tremendous pace in the field of language models (LMs), with the release of many open models and closed API systems. However, fewer and fewer disclose how they are created: Which corpora are do they use? How are they trained? How much energy they consume? In this talk, I am going to provide an overview of OLMo (https://allenai.org/olmo), an initiative at AI2 to create transparent LMs that advance the science of LLMs. I will discuss current releases, such as Tulu, and Dolma, and OLMo-7b, goals, ethical, and legal considerations in this initiative, as well as what’s coming next. 

Bio:

Luca Soldaini is a Senior Applied Research Scientist at the Allen Institute for AI in the Semantic Scholar & OLMo teams. Their current research focuses on data-centric NLP, information retrieval, and use of LMs for scientific applications. Prior to joining AI2 in 2022, Luca was a Senior Applied Scientist at Amazon Alexa, where they worked on Open Domain Question Answering. Luca obtained their PhD from Georgetown University in 2018.