PEML: Parameter-efficient Multi-Task Learning with Optimized Continuous Prompts

Presents PEML, a parameter-efficient multi-task learning approach using optimized continuous prompts to fine-tune a single LLM across multiple tasks while reducing resource consumption compared to deploying individual models.