{"about":{"site":"https://codewithpapers.app","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.","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"},"url":"/paper/arxiv-2601-05913","title":"Distilling Lightweight Domain Experts from Large ML Models by Identifying Relevant Subspaces","arxiv_id":"2601.05913","date":"2026-01-09","proceeding":null,"authors":["Pattarawat Chormai","Ali Hashemi","Klaus-Robert Müller","Grégoire Montavon"],"abstract":"Knowledge distillation involves transferring the predictive capabilities of large, high-performing AI models (teachers) to smaller models (students) that can operate in environments with limited computing power. In this paper, we address the scenario in which only a few classes and their associated intermediate concepts are relevant to distill. This scenario is common in practice, yet few existing distillation methods explicitly focus on the relevant subtask. To address this gap, we introduce 'SubDistill', a new distillation algorithm with improved numerical properties that only distills the relevant components of the teacher model at each layer. Experiments on CIFAR-100 and ImageNet with Convolutional and Transformer models demonstrate that SubDistill outperforms existing layer-wise distillation techniques on a representative set of subtasks. Our benchmark evaluations are complemented by Explainable AI analyses showing that our distilled student models more closely match the decision structure of the original teacher model.","url_abs":"https://arxiv.org/abs/2601.05913","url_pdf":"https://arxiv.org/pdf/2601.05913","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2601.05913","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2601.05913"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/p16i/subdistill","reach":{"status":"ok"}}],"summary":{"unverified":7},"by_repo_kind":{"found_in_text":{"samples":7,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":7,"samples":[{"code_sha256_prefix":"253d56ddbc191372","entry":"accuracy","repo":"p16i/subdistill","repo_kind":"found_in_text","path":"xaikd/metrics.py","file_url":"https://github.com/p16i/subdistill/blob/HEAD/xaikd/metrics.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"253d56ddbc191372"}},{"code_sha256_prefix":"6f1112fc90c9bbdc","entry":"attach_hook_intercept_module","repo":"p16i/subdistill","repo_kind":"found_in_text","path":"xaikd/interceptor.py","file_url":"https://github.com/p16i/subdistill/blob/HEAD/xaikd/interceptor.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"6f1112fc90c9bbdc"}},{"code_sha256_prefix":"9f4fc48300a3d171","entry":"collapse_dim","repo":"p16i/subdistill","repo_kind":"found_in_text","path":"xaikd/jacobian.py","file_url":"https://github.com/p16i/subdistill/blob/HEAD/xaikd/jacobian.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"9f4fc48300a3d171"}},{"code_sha256_prefix":"0b7310099b8bbd1b","entry":"compute","repo":"p16i/subdistill","repo_kind":"found_in_text","path":"xaikd/jacobian.py","file_url":"https://github.com/p16i/subdistill/blob/HEAD/xaikd/jacobian.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"0b7310099b8bbd1b"}},{"code_sha256_prefix":"41bd3f4d6fd427c2","entry":"get_logit_modifier","repo":"p16i/subdistill","repo_kind":"found_in_text","path":"xaikd/logit_modifiers.py","file_url":"https://github.com/p16i/subdistill/blob/HEAD/xaikd/logit_modifiers.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"41bd3f4d6fd427c2"}},{"code_sha256_prefix":"dfc3658a083cdb3f","entry":"register_explainer","repo":"p16i/subdistill","repo_kind":"found_in_text","path":"xaikd/explainers.py","file_url":"https://github.com/p16i/subdistill/blob/HEAD/xaikd/explainers.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"dfc3658a083cdb3f"}},{"code_sha256_prefix":"05275dc83f4eb5d3","entry":"resolve_lambda_layer","repo":"p16i/subdistill","repo_kind":"found_in_text","path":"xaikd/constants/lambda_layers.py","file_url":"https://github.com/p16i/subdistill/blob/HEAD/xaikd/constants/lambda_layers.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"05275dc83f4eb5d3"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.LG","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}