Musketeer: Joint Training for Multi-task Vision Language Model with Task Explanation Prompts
This addresses the challenge of building a single, unified model for multiple vision-language tasks, which is incremental as it builds on existing multi-task learning approaches.
The researchers tackled the problem of multi-task vision-language model training where heterogeneous tasks interfere with each other, and achieved results comparable to or better than strong single-task baselines across multiple tasks.
We present a vision-language model whose parameters are jointly trained on all tasks and fully shared among multiple heterogeneous tasks which may interfere with each other, resulting in a single model which we named Musketeer. The integration of knowledge across heterogeneous tasks is enabled by a novel feature called Task Explanation Prompt (TEP). With rich and structured information such as task input/output format, TEP reduces interference among tasks, allowing the model to focus on their shared structure. With a single model, Musketeer achieves results comparable to or better than strong baselines trained on single tasks, almost uniformly across multiple tasks.