Learning What Matters Now: Dynamic Preference Inference under Contextual Shifts

1. 1Model the preference weights as unobserved latent variables
2. 2Incorporate contextual information to capture shifts in priorities
3. 3Develop algorithms for dynamic preference inference and sequential decision-making
4. 4Evaluate the performance of the proposed approach in various scenarios

This research benefits AI engineers and ML researchers working on multi-objective reinforcement learning and decision-making systems, as it provides a framework for adapting to changing priorities and preferences

💡 Preference weights can drift with context and should be modeled as unobserved latent variables

🤖 Dynamic preference inference for sequential decision-making under contextual shifts