Financial Context-Rich Strategy for Actor-Critic based Agents: A Deep Reinforcement Learning Approach for Stock Trading
Resumo indisponível. Consulte o paper original.
PAPER / ARXIV:2610.11732
Guanghao Wu, Zhuo Cai, Shoujin Wang
RESUMO
Large language model (LLM) agents for portfolio management learn from experience: they credit each experience in their memory with the outcome of the decisions that used it. In financial markets, however, this outcome mostly reflects the market move shared by all decisions on that date, so the credit tracks the market rather than the experience, and these agents often do worse than simply holding the equal-weight (1/N) portfolio. MemTrial drafts each decision with eight combinations of retrieved experiences, chosen by a fractional factorial design, and credits each experience with its Banzhaf value, the average of these differences, pooled across dates with a hierarchical Bayesian model that only acts on values once they predict unseen dates. On four benchmarks, MemTrial is the best of 15 methods on a semi-synthetic benchmark with known experience quality, limits losses to at most 2.2% below 1/N on PortBench and InvestorBench (versus 15-38% for the best experience-learning agent), and improves utility by 21.2% on average over five settings.
NO MESMO MAPA
Resumo indisponível. Consulte o paper original.
Resumo indisponível. Consulte o paper original.
Resumo indisponível. Consulte o paper original.
Resumo indisponível. Consulte o paper original.