{"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/rmm-reinforced-memory-management-for-class-1","title":"RMM: Reinforced Memory Management for Class-Incremental Learning","arxiv_id":"2301.05792","date":"2023-01-14","proceeding":"NeurIPS 2021 12","authors":["Yaoyao Liu","Bernt Schiele","Qianru Sun"],"abstract":"Class-Incremental Learning (CIL) [40] trains classifiers under a strict memory budget: in each incremental phase, learning is done for new data, most of which is abandoned to free space for the next phase. The preserved data are exemplars used for replaying. However, existing methods use a static and ad hoc strategy for memory allocation, which is often sub-optimal. In this work, we propose a dynamic memory management strategy that is optimized for the incremental phases and different object classes. We call our method reinforced memory management (RMM), leveraging reinforcement learning. RMM training is not naturally compatible with CIL as the past, and future data are strictly non-accessible during the incremental phases. We solve this by training the policy function of RMM on pseudo CIL tasks, e.g., the tasks built on the data of the 0-th phase, and then applying it to target tasks. RMM propagates two levels of actions: Level-1 determines how to split the memory between old and new classes, and Level-2 allocates memory for each specific class. In essence, it is an optimizable and general method for memory management that can be used in any replaying-based CIL method. For evaluation, we plug RMM into two top-performing baselines (LUCIR+AANets and POD+AANets [30]) and conduct experiments on three benchmarks (CIFAR-100, ImageNet-Subset, and ImageNet-Full). Our results show clear improvements, e.g., boosting POD+AANets by 3.6%, 4.4%, and 1.9% in the 25-Phase settings of the above benchmarks, respectively.","url_abs":"https://arxiv.org/abs/2301.05792v1","url_pdf":"https://arxiv.org/pdf/2301.05792v1.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":"rmm-reinforced-memory-management-for-class-1","repo_url":"https://gitlab.mpi-klsb.mpg.de/yaoyaoliu/rmm","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null},{"paper_slug":"rmm-reinforced-memory-management-for-class-1","repo_url":"https://gitlab.mpi-klsb.mpg.de/yaoyaoliu/rmm/","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null},{"paper_slug":"rmm-reinforced-memory-management-for-class-1","repo_url":"https://github.com/aimagelab/mammoth","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"rmm-reinforced-memory-management-for-class-1","repo_url":"https://github.com/g-u-n/pycil","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"class-incremental-learning","task_name":"Class Incremental Learning"},{"task_slug":"incremental-learning","task_name":"Incremental Learning"},{"task_slug":"management","task_name":"Management"},{"task_slug":"class-incremental-learning-1","task_name":"class-incremental learning"}],"methods":[{"method_slug":"hoc","method_name":"HOC"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/incremental-learning-on-cifar-100-50-classes-2","task":"Incremental Learning","dataset":"CIFAR-100 - 50 classes + 10 steps of 5 classes","model":"RMM (Modified ResNet-32)","rank_in_archive_order":6,"of":13,"metrics":{"Average Incremental Accuracy":"67.61"},"uses_additional_data":false},{"leaderboard":"/sota/incremental-learning-on-cifar-100-50-classes-1","task":"Incremental Learning","dataset":"CIFAR-100 - 50 classes + 25 steps of 2 classes","model":"RMM (Modified ResNet-32)","rank_in_archive_order":2,"of":5,"metrics":{"Average Incremental Accuracy":"66.21"},"uses_additional_data":false},{"leaderboard":"/sota/incremental-learning-on-cifar-100-50-classes-3","task":"Incremental Learning","dataset":"CIFAR-100 - 50 classes + 5 steps of 10 classes","model":"RMM (Modified ResNet-32)","rank_in_archive_order":6,"of":15,"metrics":{"Average Incremental Accuracy":"68.86"},"uses_additional_data":false},{"leaderboard":"/sota/incremental-learning-on-imagenet-10-steps","task":"Incremental Learning","dataset":"ImageNet - 10 steps","model":"RMM (ResNet-18)","rank_in_archive_order":5,"of":10,"metrics":{"Average Incremental Accuracy":"67.45"},"uses_additional_data":false},{"leaderboard":"/sota/incremental-learning-on-imagenet-500-classes-2","task":"Incremental Learning","dataset":"ImageNet - 500 classes + 10 steps of 50 classes","model":"RMM (ResNet-18)","rank_in_archive_order":1,"of":4,"metrics":{"Average Incremental Accuracy":"67.45"},"uses_additional_data":false},{"leaderboard":"/sota/incremental-learning-on-imagenet-500-classes-3","task":"Incremental Learning","dataset":"ImageNet - 500 classes + 25 steps of 20 classes","model":"RMM (ResNet-18)","rank_in_archive_order":1,"of":1,"metrics":{"Average Incremental Accuracy":"63.93"},"uses_additional_data":false},{"leaderboard":"/sota/incremental-learning-on-imagenet-500-classes-1","task":"Incremental Learning","dataset":"ImageNet - 500 classes + 5 steps of 100 classes","model":"RMM (ResNet-18)","rank_in_archive_order":1,"of":4,"metrics":{"Average Incremental Accuracy":"69.21"},"uses_additional_data":false},{"leaderboard":"/sota/incremental-learning-on-imagenet-100-50-2","task":"Incremental Learning","dataset":"ImageNet-100 - 50 classes + 10 steps of 5 classes","model":"RMM (ResNet-18)","rank_in_archive_order":1,"of":5,"metrics":{"Average Incremental Accuracy":"78.47"},"uses_additional_data":false},{"leaderboard":"/sota/incremental-learning-on-imagenet-100-50-1","task":"Incremental Learning","dataset":"ImageNet-100 - 50 classes + 25 steps of 2 classes","model":"RMM (ResNet-18)","rank_in_archive_order":1,"of":3,"metrics":{"Average Incremental Accuracy":"76.54"},"uses_additional_data":false},{"leaderboard":"/sota/incremental-learning-on-imagenet-100-50-3","task":"Incremental Learning","dataset":"ImageNet-100 - 50 classes + 5 steps of 10 classes","model":"RMM (ResNet-18)","rank_in_archive_order":2,"of":5,"metrics":{"Average Incremental Accuracy":"79.52"},"uses_additional_data":false},{"leaderboard":"/sota/incremental-learning-on-imagenet100-10-steps","task":"Incremental Learning","dataset":"ImageNet100 - 10 steps","model":"RMM (ResNet-18)","rank_in_archive_order":2,"of":13,"metrics":{"Average Incremental Accuracy":"78.47"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2301.05792","atlas_url":"https://app.syntology.ai/?focus=2301.05792","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2301.05792"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+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":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/aimagelab/mammoth","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://gitlab.mpi-klsb.mpg.de/yaoyaoliu/rmm","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://gitlab.mpi-klsb.mpg.de/yaoyaoliu/rmm/","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/g-u-n/pycil","reach":null}],"summary":{"ran_draft_wrong":1,"ran_honours":1,"unverified":2},"by_repo_kind":{"listed":{"samples":2,"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":4,"samples":[{"code_sha256_prefix":"fa5bbac75c929c9f","entry":"accuracy","repo":null,"repo_kind":null,"path":null,"file_url":null,"link_basis":"identical_code_first_harvested_elsewhere","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":null,"inline_ok":false,"mcp_get_code":{"code_sha256":"fa5bbac75c929c9f"}},{"code_sha256_prefix":"929c76a2fa2e8c42","entry":"tensor2numpy","repo":null,"repo_kind":null,"path":null,"file_url":null,"link_basis":"identical_code_first_harvested_elsewhere","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"invariant","behaviour_fingerprint":false,"licence":null,"inline_ok":false,"mcp_get_code":{"code_sha256":"929c76a2fa2e8c42"}},{"code_sha256_prefix":"8347652eeeb36202","entry":"BaseLearner","repo":"g-u-n/pycil","repo_kind":"listed","path":"models/rmm.py","file_url":"https://github.com/g-u-n/pycil/blob/HEAD/models/rmm.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"8347652eeeb36202"}},{"code_sha256_prefix":"81b51807a56d10fa","entry":"RMMBase","repo":"g-u-n/pycil","repo_kind":"listed","path":"models/rmm.py","file_url":"https://github.com/g-u-n/pycil/blob/HEAD/models/rmm.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"81b51807a56d10fa"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}