Direct Preference Optimization With Unobserved Preference Heterogeneity
Keywords: Pluralistic Alignment, RLHF, Fairness
Abstract: RLHF has emerged as a pivotal step in aligning language models with human objectives and values. It typically involves learning a reward model from human preference data and then using reinforcement learning to update the generative model accordingly. Conversely, Direct Preference Optimization (DPO) directly optimizes the generative model with preference data, skipping reinforcement learning. However, both RLHF and DPO assume uniform preferences, overlooking the reality of diverse human annotators. This paper presents a new method to align generative models with varied human preferences. We propose an Expectation-Maximization adaptation to DPO, generating a mixture of models based on latent preference types of the annotators. We then introduce a min-max regret ensemble learning model to produce a single generative method to minimize worst-case regret among annotator subgroups with similar latent factors. Our algorithms leverage the simplicity of DPO while accommodating diverse preferences. Experimental results validate the effectiveness of our approach in producing equitable generative policies.
Archival: Non-archival Track
Submission Number: 18
Loading