Collaborative Machine Learning Research: Open Data, Models, Tools

Abstract: In this session, Suzana will talk about BigScience – a one-year long collaborative research initiative on large multilingual datasets and large language models with hundreds of contributors from all over the world. She’ll show how we can work toward accessibility, transparency, and responsible AI development through a decentralized community-driven research approach and collaboratively designed release strategies, while the Hugging Face ecosystem provides the open source infrastructure for data, models and tools.

BigScience launched the training of an open source multilingual 176B parameter LLM. You can follow @BigScienceW on Twitter for live training updates. More info: [https://bigscience.huggingface.co/](https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fbigscience.huggingface.co%2F&data=05%7C01%7Clinqing.liu.20%40ucl.ac.uk%7C5388c44f8c3e48e6e87f08da2a883667%7C1faf88fea9984c5b93c9210a11d9a5c2%7C0%7C0%7C637869061762655206%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=%2Bc1kq1iURD4WCc1gFcz6juSk8oCQNDekdX9gGuEA%2FEU%3D&reserved=0)

Bio: Suzana is a Technical Program Manager at Hugging Face, where she is co-chairing the BigScience organization working group. She is also leading MLT, a machine learning nonprofit organization. Previously she worked as a Computational Linguist on NLP research and product in London and Tokyo.