One Model, Multiple Modalities: A Sparsely Activated Approach for Text, Sound, Image, Video and Code
This addresses the need for more efficient and interpretable multi-modal AI systems, offering a novel approach that could reduce computational costs and improve flexibility, though it appears incremental in combining sparsity with multi-modality.
The paper tackles the problem of AI systems processing only individual modalities by introducing SkillNet, a single model that sparsely activates specialized parameters for multiple modalities like text, sound, image, video, and code. It achieves comparable or better performance than modality-specific models, with higher accuracy in Chinese text-to-image retrieval than leading systems while using fewer activated parameters.
People perceive the world with multiple senses (e.g., through hearing sounds, reading words and seeing objects). However, most existing AI systems only process an individual modality. This paper presents an approach that excels at handling multiple modalities of information with a single model. In our "{SkillNet}" model, different parts of the parameters are specialized for processing different modalities. Unlike traditional dense models that always activate all the model parameters, our model sparsely activates parts of the parameters whose skills are relevant to the task. Such model design enables SkillNet to learn skills in a more interpretable way. We develop our model for five modalities including text, image, sound, video and code. Results show that, SkillNet performs comparably to five modality-specific fine-tuned models. Moreover, our model supports self-supervised pretraining with the same sparsely activated way, resulting in better initialized parameters for different modalities. We find that pretraining significantly improves the performance of SkillNet on five modalities, on par with or even better than baselines with modality-specific pretraining. On the task of Chinese text-to-image retrieval, our final system achieves higher accuracy than existing leading systems including Wukong{ViT-B} and Wenlan 2.0 while using less number of activated parameters.