Overview
Explore the critical issue of social biases in text representations and strategies for their mitigation in this 43-minute conference talk by Dr. Danushka Bollegala, an NLP and ML expert from the University of Liverpool. Delve into topics such as gender bias, biases in word embeddings and masked language models, and multi-lingual bias evaluation. Gain insights from Bollegala's extensive experience as head of the ML Research Group and Amazon Scholar, and learn about the latest advancements in addressing AI bias. Participate in an interactive session featuring a quiz, detailed explanations of machine learning model anatomy, and a Q&A segment to deepen your understanding of this crucial aspect of responsible AI development.
Syllabus
- Introductions
- Quiz Time
- How about AI?
- Anatomy of Machine Learning Models
- Gender Bias
- Bias in word embeddings
- Bias in Masked Language Models
- Multi-lingual Bias Evaluation
- Q & A
Taught by
Open Data Science