Advancing Sequential Numerical Prediction in Autoregressive Models

摘要

Autoregressive models have become the de facto choice for sequence generation tasks, but standard approaches treat digits as independent tokens and apply cross-entropy loss, overlooking the coherent structure of numerical sequences. This paper introduces \textit{\textbf{N}umerical \textbf{T}oken \textbf{I}ntegrity \textbf{Loss} (NTIL)} to address this gap. NTIL operates at two levels: (1) token-level, where it extends the Earth Mover’s Distance (EMD) to preserve ordinal relationships between numerical values, and (2) sequence-level, where it penalizes the overall discrepancy between the predicted and actual sequences. This dual approach improves numerical prediction and integrates effectively with LLMs/MLLMs. Extensive experiments show significant performance improvements with NTIL.