MARVIS: Modality Adaptive Reasoning over VISualizations
This addresses the flexibility gap between specialized and foundation models for scientific applications, offering a training-free solution that avoids exposing personally identifiable information.
The paper tackles the problem of enabling small vision-language models to predict any data modality with high accuracy, achieving competitive performance across vision, audio, biological, and tabular domains, beating Gemini by 16% on average without domain-specific training.
Scientific applications of machine learning often rely on small, specialized models tuned to particular domains. Such models often achieve excellent performance, but lack flexibility. Foundation models offer versatility, but typically underperform specialized approaches, especially on non-traditional modalities and long-tail domains. We propose MARVIS (Modality Adaptive Reasoning over VISualizations), a training-free method that enables even small vision-language models to predict any data modality with high accuracy. MARVIS transforms latent embedding spaces into visual representations and then leverages the spatial and fine-grained reasoning skills of VLMs to successfully interpret and utilize them. MARVIS achieves competitive performance on vision, audio, biological, and tabular domains using a single 3B parameter model, achieving results that beat Gemini by 16\% on average and approach specialized methods, without exposing personally identifiable information (P.I.I.) or requiring any domain-specific training. We open source our code and datasets at https://github.com/penfever/marvis