We're thrilled to announce that we've had two research papers accepted to EMNLP, Positive Impact Track 2024 in Miami, Florida! EMNLP (Conference on Empirical Methods in Natural Language Processing) is one of the top-tier international NLP conferences and is a standard venue submitted to by universities and tech companies alike.
Congratulations to our first group, Abhay Gupta, Philip Meng, and Ece Yurtseven! Their research, titled “Detecting LLM Biases on NLU Tasks in AAVE via a Novel Benchmark,” addresses a critical issue in equitable NLP by developing a novel benchmark to evaluate biases in language models, specifically focusing on African American Vernacular English (AAVE).
Read their paper: https://www.arxiv.org/pdf/2408.14845
Learn more about their project: https://aavenue.live/
Congratulations to our second group, Rajat Rawat, Hudson McBride, Rajarshi Ghosh, Dhiyaan, Dhruv Alamuri, and Jong Moon! Their research, titled “DiversityMedQA: Assessing Demographic Biases in Medical Diagnosis using Large Language Models, explores critical issues in the intersection of medical diagnosis and demographic biases, showcasing innovative methods to assess and mitigate these biases using large language models.
Read their paper: https://arxiv.org/pdf/2409.01497
For our students to have their work accepted to Ph.D-level conferences is an extraordinary achievement! We're thrilled about their research and look forward to their presentation at the conference in November, located in Miami, Florida!