Market Decompression: Learning Latent-State Inverse Maps for Cold-Start Financial Prediction

Introduces market decompression, which pretrains a model to recover hidden volatility, jump, and regime states from simulated financial market data, then uses a small number of real outcomes for forecasting. Tests are conducted in simulated settings and five real markets; results confirm improved performance in three real markets. Single-author paper accepted to the 7th ACM International Conference on AI in Finance (ICAIF 2026). This upload is the author-identified manuscript, not the final ACM version of record.