RICE-Alpha uses a Multi-Tier Memory Layer to ground news interpretation in temporally eligible issuer-specific history, and a Typed Event Agent to construct event states whose successor relations are formed within issuers and pooled across firms only after valid local pairing.
Event graph corrections improve LLM agent stock forecasts by modeling historical reliability
Authors show that separating a history-aware base forecast from reliability-calibrated residual corrections derived from event transitions yields large gains in predictive accuracy and portfolio returns.
Research Lab
Tong Liu · Lanmiao Liu · Xiang Hu
Zircon Security · Utrecht University · The Max Planck Institute for Psycholinguistics · China Life R&D Center
Research Digest··2 min read
The authors present RICE-Alpha, a point-in-time stock-scoring framework that combines a history-aware multi-view Base Alpha with a reliability-weighted residual correction from event graph transitions.
Why this paper
From The Max Planck Institute for Psycholinguistics and 3 others
In one line
Historical event continuation adds incremental stock forecasting signal when temporally grounded, reliability-calibrated, and residualized against a multi-view base forecast.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ✓Reports numbers on named benchmarks (2 benchmarks)
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§