FinATOM uses a causal language model's token vocabulary to output stock forecasts and ETF allocations, and shows policy optimization improves out-of-sample Sharpe.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Financial Numerical Prediction and Allocation as Token Generation
FinATOM uses a causal language model's token vocabulary to output stock forecasts and ETF allocations, and shows policy optimization improves out-of-sample Sharpe.