Work

Staged Continual Adaptation of Multimodal Foundation Models for Japanese Financial Documents

Year: 2026

Type: preprint

Abstract: Staged post-training of multimodal foundation models chains vision-language alignment, reasoning distillation, and domain tuning, yet how each phase trades capabilities against the others remains under-characterised. Tracking an 8.4B-parameter model from an un-fine-tuned baseline through three training phases on Japanese financial documents (evaluated on five text-input benchmarks), we find the phases interact non-additively: math-only distillation overwrites the financial knowledge acquired earlier even as procedural math skill survives, and the domain phase repairs more than it installs. A skip-phase ablation is best for domain knowledge but forfeits numerical-reasoning gains, so per-phase contributions are not decomposable and the best checkpoint is task-dependent, with implications for checkpoint retention and scheduling rather than added stages.

Source: arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.27904

Authors: Kakimoto, Genshin, Yanagisawa, Atsushi

Language: English

Topic: Multimodal Machine Learning Applications

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

Funders: National Institute of Advanced Industrial Science and Technology, Ministry of Education, Culture, Sports, Science and Technology

OpenAlex ID: W7223084839

Cites: 10

Cited by:

Related to: 10