Right-Sizing Communication and Recommendation Set Size in AI-Assisted Search

In three linesTheoretical paper on optimizing user-AI recommendation system interaction. Models communication cost (precision of user message) and search cost (size of recommendation set). For large d, characterizes how optimal message precision and recommendation set size depend on cost parameters under two sampling schemes: posterior belief and optimized tilted distribution.

Summary generated by Claude — human-verified