New bandit algorithms incorporate risk-sensitive Cumulative Prospect Theory

Researchers have developed new algorithms for non-parametric and kernel bandits that incorporate Cumulative Prospect Theory (CPT) for risk-sensitive decision-making. These algorithms, named Non-Parametric CPT Bandits (NPCB) and Kernel-CPT bandits, aim to optimize for CPT-based regret rather than traditional expected reward. The proposed methods achieve regret bounds comparable to existing non-parametric bandit algorithms, indicating that incorporating CPT risk sensitivity does not incur additional order-wise regret costs. AI

IMPACT Introduces novel algorithms for risk-sensitive decision-making in machine learning, potentially impacting areas where risk aversion is critical.

RANK_REASON The cluster contains a research paper detailing new algorithms in machine learning. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• arXiv
• Cumulative prospect theory
• Hugging Face
• Kernel-CPT bandits
• Non-Parametric CPT Bandits

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →