{"paper":{"title":"PEML: Parameter-efficient Multi-Task Learning with Optimized Continuous Prompts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"PEML jointly optimizes continuous prompts via neural architecture engineering and low-rank model adaptation to improve multi-task LLM performance.","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Anjir Ahmed Chowdhury, Feng Yan, Syed Zawad, Xiaolong Ma, Xu Dong","submitted_at":"2026-05-13T19:25:56Z","abstract_excerpt":"Parameter-Efficient Fine-Tuning (PEFT) is widely used for adapting Large Language Models (LLMs) for various tasks. Recently, there has been an increasing demand for fine-tuning a single LLM for multiple tasks because it requires overall less data for fine-tuning thanks to the common features shared among tasks. More importantly, LLMs are resource demanding and deploying a single model for multiple tasks facilitates resource consolidation and consumes significantly less resources compared to deploying individual large model for each task. Existing PEFT methods like LoRA and Prefix Tuning are de"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"The evaluation results present an average accuracy improvement of up to 6.67%, with individual tasks showing peak gains of up to 10.75%.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that the proposed neural architecture for prompt optimization combined with low-rank adaptation will consistently outperform existing methods across diverse tasks without introducing new overfitting risks or requiring extensive hyperparameter tuning.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"PEML co-optimizes continuous prompts and low-rank adaptations to deliver up to 6.67% average accuracy gains over existing multi-task PEFT methods on GLUE, SuperGLUE, and other benchmarks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"PEML jointly optimizes continuous prompts via neural architecture engineering and low-rank model adaptation to improve multi-task LLM performance.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"df9cd44af316d65fbf179600d7ebaa0f01eddd148925f9a9fae86dbab98a077f"},"source":{"id":"2605.14055","kind":"arxiv","version":1},"verdict":{"id":"a9e0e923-90d6-40d8-b48a-872150052678","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T05:22:17.529353Z","strongest_claim":"The evaluation results present an average accuracy improvement of up to 6.67%, with individual tasks showing peak gains of up to 10.75%.","one_line_summary":"PEML co-optimizes continuous prompts and low-rank adaptations to deliver up to 6.67% average accuracy gains over existing multi-task PEFT methods on GLUE, SuperGLUE, and other benchmarks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that the proposed neural architecture for prompt optimization combined with low-rank adaptation will consistently outperform existing methods across diverse tasks without introducing new overfitting risks or requiring extensive hyperparameter tuning.","pith_extraction_headline":"PEML jointly optimizes continuous prompts via neural architecture engineering and low-rank model adaptation to improve multi-task LLM performance."},"references":{"count":97,"sample":[{"doi":"","year":2022,"title":"International conference on machine learning , pages=","work_id":"b0afdc1a-933a-4632-8573-7d9369733159","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Parameter-efficient multi-task fine-tuning for transformers via shared hypernetworks , author=. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th Int","work_id":"c30f71f5-c2f5-4aa0-ab94-964798d41ffd","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=","work_id":"78866b2d-199c-4f39-bbee-b29ec1cae08d","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.18653/v1/2024.findings-emnlp.956","year":2024,"title":"H yper L o RA : Efficient Cross-task Generalization via Constrained Low-Rank Adapters Generation","work_id":"960c86ea-b25e-4898-9d4c-823c9a6372ad","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Lorahub: Efficient cross-task generalization via dynamic lora composition","work_id":"66ee831a-2b33-423d-87ec-ba52b341a21b","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":97,"snapshot_sha256":"ea3e51eef5d7f44d5ae903f106e7696f708240b9a538c7084dcab1ae53ee4b81","internal_anchors":31},"formal_canon":{"evidence_count":2,"snapshot_sha256":"87cfab8c17c29dd93f60d454e8c0829daa5b8c9c64dc2ce7056e308a25643a2d"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}