Part I: Background on LMs
Conference Room A
- Introduction to LMs
- Root Causes of Bias in LMs
- Root Causes of Bias in LMs
Overview
Language Models (LMs) have demonstrated remarkable success across various domains over the years. However, despite their promising performance on various real world tasks, most of these algorithms lack fairness considerations, potentially leading to discriminatory outcomes against marginalized demographic groups and individuals. Many recent publications have explored ways to mitigate bias in LMs. Nevertheless, a comprehensive understanding of the root causes of bias, their effects, and possible limitations of LMs from the perspective of fairness is still in its early stages. To bridge this gap, this tutorial provides a systematic overview of recent advances in fair LMs, beginning with real-world case studies, followed by an analysis of bias causes. We then explore fairness concepts specific to LMs, summarizing bias evaluation strategies and algorithms designed to promote fairness. Finally, we analyze bias in LM datasets and discuss current research challenges and open questions in the field.
Our tutorial is structured into five key parts:
- Background on LMs
- Quantifying Bias in LMs
- Mitigating Bias in LMs
- Resources for Evaluating Bias
- Future Directions
Check out our survey on fairness in large language models.
This tutorial is grounded in our surveys and established benchmarks, all available as open-source resources.
Speakers
Agenda
12:00 - 12:30
12:30 - 13:00
Part II: Quantifying Bias in LMs
Conference Room A
- Demographic representation
- Stereotypical association
- Counterfactual fairness
- Performance disparities
- Stereotypical association
- Counterfactual fairness
- Performance disparities
Coffee Break
Room A5
Coffee and cakes lorem ipsum dolor sit amet, consectetur adipiscing elit.
13:30 - 14:00
Part III: Mitigating Bias in LMs
Conference Room A
- Pre-processing
- In-training
- Intra-processing
- Post-processing
- In-training
- Intra-processing
- Post-processing
14:00 - 14:30
Part IV: Resources for Evaluating Bias
Conference Room B
- Toolkits
- Datasets
- Datasets
14:30 - 15:00
Part V:Future Directions
Conference Room C
- Rational Counterfactual Data Augmentation
- Balancing Performance and Fairness in LLMs
- Fulfilling Multiple Types of Fairness
- Developing More and Tailored Datasets