{"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/information-theoretic-representation","title":"Information Theoretic Representation Distillation","arxiv_id":"2112.00459","date":"2021-12-01","proceeding":null,"authors":["Roy Miles","Adrian Lopez Rodriguez","Krystian Mikolajczyk"],"abstract":"Despite the empirical success of knowledge distillation, current state-of-the-art methods are computationally expensive to train, which makes them difficult to adopt in practice. To address this problem, we introduce two distinct complementary losses inspired by a cheap entropy-like estimator. These losses aim to maximise the correlation and mutual information between the student and teacher representations. Our method incurs significantly less training overheads than other approaches and achieves competitive performance to the state-of-the-art on the knowledge distillation and cross-model transfer tasks. We further demonstrate the effectiveness of our method on a binary distillation task, whereby it leads to a new state-of-the-art for binary quantisation and approaches the performance of a full precision model. Code: www.github.com/roymiles/ITRD","url_abs":"https://arxiv.org/abs/2112.00459v3","url_pdf":"https://arxiv.org/pdf/2112.00459v3.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":"information-theoretic-representation","repo_url":"https://github.com/roymiles/ITRD","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"classification-with-binary-weight-network","task_name":"Classification with Binary Weight Network"},{"task_slug":"knowledge-distillation","task_name":"Knowledge Distillation"},{"task_slug":"question-answering","task_name":"Question Answering"}],"methods":[{"method_slug":"knowledge-distillation","method_name":"Knowledge Distillation"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/knowledge-distillation-on-cifar-100","task":"Knowledge Distillation","dataset":"CIFAR-100","model":"resnet8x4 (T: resnet32x4 S: resnet8x4)","rank_in_archive_order":9,"of":27,"metrics":{"Top-1 Accuracy (%)":"76.68"},"uses_additional_data":false},{"leaderboard":"/sota/knowledge-distillation-on-cifar-100","task":"Knowledge Distillation","dataset":"CIFAR-100","model":"vgg8 (T:vgg13 S:vgg8)","rank_in_archive_order":16,"of":27,"metrics":{"Top-1 Accuracy (%)":"74.93"},"uses_additional_data":false},{"leaderboard":"/sota/knowledge-distillation-on-cifar-100","task":"Knowledge Distillation","dataset":"CIFAR-100","model":"resnet110 (T:resnet110 S:resnet20)","rank_in_archive_order":23,"of":27,"metrics":{"Top-1 Accuracy (%)":"71.99"},"uses_additional_data":false},{"leaderboard":"/sota/knowledge-distillation-on-imagenet","task":"Knowledge Distillation","dataset":"ImageNet","model":"ITRD (T: ResNet-34 S:ResNet-18)","rank_in_archive_order":41,"of":52,"metrics":{"CRD training setting":"✓","Top-1 accuracy %":"71.68","model size":"11.69M"},"uses_additional_data":false},{"leaderboard":"/sota/question-answering-on-squad11","task":"Question Answering","dataset":"SQuAD1.1","model":"BERT - 6 Layers","rank_in_archive_order":50,"of":213,"metrics":{"EM":"81.5","F1":"88.5"},"uses_additional_data":false},{"leaderboard":"/sota/question-answering-on-squad11","task":"Question Answering","dataset":"SQuAD1.1","model":"BERT - 3 Layers","rank_in_archive_order":97,"of":213,"metrics":{"EM":"77.7","F1":"85.8"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2112.00459","atlas_url":"https://app.syntology.ai/?focus=2112.00459","mcp":null,"developers":"https://syntology.ai/developers","held":true,"reason":"arxiv_id_in_atlas","checked_against":{"atlas_ids":549449,"atlas_ids_sha256":"afd72399dafd9906e9062927f70655952385a3d75488558e1f0484e950ab09b7","read_at":"2026-10-05T01:46:21Z"},"agent_calls":[{"tool":"get_citation_path","arguments":{"paper_1":"2112.00459"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"2112.00459\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"2112.00459"},"call":"get_concepts_for_paper(arxiv_id=\"2112.00459\")"}],"by_repo":[{"repo":"roymiles/ITRD","kind":"official","samples":0,"ran":0,"instrument":0,"unverified":0}],"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":"9e5ceee0b094ae295518362799f8d514a9b3f74d6beefa654480da676e495df3","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}