{"about":{"non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","site":"https://codewithpapers.app","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/paper/mmrl-parameter-efficient-and-interaction","title":"MMRL++: Parameter-Efficient and Interaction-Aware Representation Learning for Vision-Language Models","arxiv_id":"2505.10088","date":"2025-05-15","proceeding":null,"authors":["Yuncheng Guo","Xiaodong Gu"],"abstract":"Large-scale pre-trained Vision-Language Models (VLMs) have significantly advanced transfer learning across diverse tasks. However, adapting these models with limited few-shot data often leads to overfitting, undermining their ability to generalize to new tasks. To address this, we propose Multi-Modal Representation Learning (MMRL), which introduces a shared, learnable, modality-agnostic representation space. MMRL generates space tokens projected into both text and image encoders as representation tokens, enabling more effective cross-modal interactions. Unlike prior methods that mainly optimize class token features, MMRL inserts representation tokens into higher encoder layers--where task-specific features are more prominent--while preserving general knowledge in the lower layers. During training, both class and representation features are jointly optimized: a trainable projection layer is applied to representation tokens for task adaptation, while the projection layer for class token remains frozen to retain pre-trained knowledge. To further promote generalization, we introduce a regularization term aligning class and text features with the frozen VLM's zero-shot features. At inference, a decoupling strategy uses both class and representation features for base tasks, but only class features for novel tasks due to their stronger generalization. Building upon this, we propose MMRL++, a parameter-efficient and interaction-aware extension that significantly reduces trainable parameters and enhances intra-modal interactions--particularly across the layers of representation tokens--allowing gradient sharing and instance-specific information to propagate more effectively through the network. Extensive experiments on 15 datasets demonstrate that MMRL and MMRL++ consistently outperform state-of-the-art methods, achieving a strong balance between task-specific adaptation and generalization.","url_abs":"https://arxiv.org/abs/2505.10088v1","url_pdf":"https://arxiv.org/pdf/2505.10088v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"mmrl-parameter-efficient-and-interaction","repo_url":"https://github.com/yunncheng/MMRL","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"general-knowledge","task_name":"General Knowledge"},{"task_slug":"prompt-engineering","task_name":"Prompt Engineering"},{"task_slug":"representation-learning","task_name":"Representation Learning"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"}],"methods":[{"method_slug":"align","method_name":"ALIGN"},{"method_slug":"base","method_name":"BASE"},{"method_slug":"clip","method_name":"CLIP"},{"method_slug":"coop","method_name":"CoOp"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/prompt-engineering-on-caltech-101","task":"Prompt Engineering","dataset":"Caltech-101","model":"MMRL++","rank_in_archive_order":3,"of":14,"metrics":{"Harmonic mean":"96.75"},"uses_additional_data":false},{"leaderboard":"/sota/prompt-engineering-on-dtd","task":"Prompt Engineering","dataset":"DTD","model":"MMRL++","rank_in_archive_order":2,"of":14,"metrics":{"Harmonic mean":"74.46"},"uses_additional_data":false},{"leaderboard":"/sota/prompt-engineering-on-eurosat","task":"Prompt Engineering","dataset":"EuroSAT","model":"MMRL++","rank_in_archive_order":1,"of":14,"metrics":{"Harmonic mean":"91.94"},"uses_additional_data":false},{"leaderboard":"/sota/prompt-engineering-on-fgvc-aircraft","task":"Prompt Engineering","dataset":"FGVC-Aircraft","model":"MMRL++","rank_in_archive_order":2,"of":14,"metrics":{"Harmonic mean":"42.24"},"uses_additional_data":false},{"leaderboard":"/sota/prompt-engineering-on-food-101","task":"Prompt Engineering","dataset":"Food-101","model":"MMRL++","rank_in_archive_order":8,"of":13,"metrics":{"Harmonic mean":"91.1"},"uses_additional_data":false},{"leaderboard":"/sota/prompt-engineering-on-imagenet","task":"Prompt Engineering","dataset":"ImageNet","model":"MMRL++","rank_in_archive_order":4,"of":15,"metrics":{"Harmonic mean":"74.44"},"uses_additional_data":false},{"leaderboard":"/sota/prompt-engineering-on-oxford-102-flower","task":"Prompt Engineering","dataset":"Oxford 102 Flower","model":"MMRL++","rank_in_archive_order":3,"of":14,"metrics":{"Harmonic mean":"87.01"},"uses_additional_data":false},{"leaderboard":"/sota/prompt-engineering-on-oxford-iiit-pet-dataset","task":"Prompt Engineering","dataset":"Oxford-IIIT Pet Dataset","model":"MMRL++","rank_in_archive_order":7,"of":14,"metrics":{"Harmonic mean":"96.51"},"uses_additional_data":false},{"leaderboard":"/sota/prompt-engineering-on-sun397","task":"Prompt Engineering","dataset":"SUN397","model":"MMRL++","rank_in_archive_order":3,"of":14,"metrics":{"Harmonic mean":"81.28"},"uses_additional_data":false},{"leaderboard":"/sota/prompt-engineering-on-stanford-cars-1","task":"Prompt Engineering","dataset":"Stanford Cars","model":"MMRL++","rank_in_archive_order":2,"of":14,"metrics":{"Harmonic mean":"78.18"},"uses_additional_data":false},{"leaderboard":"/sota/prompt-engineering-on-ucf101","task":"Prompt Engineering","dataset":"UCF101","model":"MMRL++","rank_in_archive_order":4,"of":14,"metrics":{"Harmonic mean":"83.81"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers","held":null,"reason":"arxiv_id_not_in_atlas","checked_against":{"atlas_ids":549449,"atlas_ids_sha256":"afd72399dafd9906e9062927f70655952385a3d75488558e1f0484e950ab09b7","read_at":"2026-10-05T01:46:21Z"},"by_repo":[{"repo":"yunncheng/MMRL","kind":"official","samples":0,"ran":0,"instrument":0,"unverified":0,"harvested_for_other_papers":1,"harvested_for":["2503.08497"],"harvested_for_calls":[{"arxiv_id":"2503.08497","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2503.08497"},"call":"get_harvested_code_for_paper(arxiv_id=\"2503.08497\")"}]}],"by_repo_harvested_for_is":"harvested_for: the arXiv ids (as Syntology's graph holds them) of every paper whose samples Syntology's graph links to this repository, in arXiv-id order, complete; the page names the first 3 and says how many more. A link only: it does not say the papers are the same work, that one copied the other, or that the code implements this paper's method; harvested_for_calls gives the get_harvested_code_for_paper call for each arXiv-shaped id among them","by_repo_is":"one row per repository in the archive's code links, Syntology's graph links or the samples; samples 0 means no sample is linked to this paper, a statement about Syntology's coverage, not about the repository: with harvested_for_other_papers present, Syntology harvested the repository for that many other papers, and without it, Syntology harvested nothing from it; placed_by_identical_code counts samples whose link records no repository, placed here because identical code was harvested from this repository for another paper; ran counts executions on a synthesized input and instrument counts failures of Syntology's instrument, not of the code"},"code_links_note":null,"arxiv_metadata":null,"syntology_extracted_results":null,"record_sha256":"d7f904d838f5d63e4b618c4f9b5c35bb0f30045847f430f23a52d11a85e01638","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}