Provably Efficient Personalized Multi-Objective Bandits with Proactive Conversational Queries
Quick summary
arXiv:2606.08410v2 Announce Type: replace-cross Abstract: Personalized decision-making in multi-objective bandits requires learning user-specific trade-offs among competing objectives. Since arm utility depends on both unknown rewards and unknown preferences, existing methods infer preferences only from utility feedback, entangling preference learning with reward exploration. In practice, however, users often reveal their priorities through proactive conversational queries (e.g., "cheap and clean hotel"), yet this structured signal is not leveraged. We formalize a proactive query-based framewo
Key takeaways
- arXiv:2606.08410v2 Announce Type: replace-cross Abstract: Personalized decision-making in multi-objective bandits requires learning user-specific trade-offs among competing objectives.
- Since arm utility depends on both unknown rewards and unknown preferences, existing methods infer preferences only from utility feedback, entangling preference learning with reward exploration.
- In practice, however, users often reveal their priorities through proactive conversational queries (e.g., "cheap and clean hotel"), yet this structured signal is not leveraged.
Why it matters
The importance of “Provably Efficient Personalized Multi-Objective Bandits with Proactive Conversational Queries” will be measured by what changes in practice. User behavior, access conditions, verifiable performance and responsible-use outcomes are the signals worth following.
Member comments