Khalil Mrini

Mrini's Image


Consumer Health Applications in the age of LLMs & ChatGPT


16:20 - 17:05


LLMs have pushed the state of the art in question answering, natural language understanding, and a range of other NLP tasks. ChatGPT pushed NLP to the mainstream. In this talk, I introduce my two publications at COLING 2022 and ACL 2022. I compare the models I developed just two years ago for biomedical applications to ChatGPT's zero-shot and retrieval abilities. In these papers, I develop methods for text understanding and question answering for consumer health applications, or users of medical language technology systems. First, I bring together answer selection and question understanding to design a system for medical Question Understanding and Answering. The proposed system takes a long, user-written medical question as input, and selects the best answer from a medical knowledge base using self-supervised losses. Then, I study text understanding through the lens of entity linking for utterances written by users on social media. I find that reformulating tasks as language generation can help in teaching models the conceptual vocabulary gap between doctors and patients. Finally, I outline the progress and limitations of current LLMs in knowledge-hungry domains like the biomedical field.


Dr. Khalil Mrini is a Senior Research Scientist at TikTok, working on Generative AI, LLMs and Multimodality. Before joining TikTok, he worked at Meta as AI Research Scientist on Generative NLP problems ranging from hashtag recommendation to pre-training for LLMs. Dr. Mrini graduated with a PhD in Computer Science from UC San Diego in 2022, and won his department's Best Doctoral Dissertation Award.

He has published first-author conference papers at top-tier venues like ACL and EMNLP, served as Area Chair for ACL and NAACL, and has won a Best Student Paper Award at a NAACL workshop. During his PhD, Khalil interned at Adobe Research, Amazon Alexa, Facebook AI and Google Brain. He has won an Amazon Research Award and numerous Adobe Research Gifts to fund his PhD research. He holds Bachelor's and Master's Degrees in Computer Science from EPFL in Switzerland, and spent a year abroad at NTU in Singapore and on internship at Infosys in Bangalore, India. Khalil was also a co-founder of North Africans in NLP, an affinity group for North African researchers in language technology, and a co-organizer of the ArabicNLP conference.

Go Back