{"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/neural-network-encapsulation","title":"Neural Network Encapsulation","arxiv_id":"1808.03749","date":"2018-08-11","proceeding":"ECCV 2018 9","authors":["Hongyang Li","Xiaoyang Guo","Bo Dai","Wanli Ouyang","Xiaogang Wang"],"abstract":"A capsule is a collection of neurons which represents different variants of a\npattern in the network. The routing scheme ensures only certain capsules which\nresemble lower counterparts in the higher layer should be activated. However,\nthe computational complexity becomes a bottleneck for scaling up to larger\nnetworks, as lower capsules need to correspond to each and every higher\ncapsule. To resolve this limitation, we approximate the routing process with\ntwo branches: a master branch which collects primary information from its\ndirect contact in the lower layer and an aide branch that replenishes master\nbased on pattern variants encoded in other lower capsules. Compared with\nprevious iterative and unsupervised routing scheme, these two branches are\ncommunicated in a fast, supervised and one-time pass fashion. The complexity\nand runtime of the model are therefore decreased by a large margin. Motivated\nby the routing to make higher capsule have agreement with lower capsule, we\nextend the mechanism as a compensation for the rapid loss of information in\nnearby layers. We devise a feedback agreement unit to send back higher capsules\nas feedback. It could be regarded as an additional regularization to the\nnetwork. The feedback agreement is achieved by comparing the optimal transport\ndivergence between two distributions (lower and higher capsules). Such an\nadd-on witnesses a unanimous gain in both capsule and vanilla networks. Our\nproposed EncapNet performs favorably better against previous state-of-the-arts\non CIFAR10/100, SVHN and a subset of ImageNet.","url_abs":"http://arxiv.org/abs/1808.03749v1","url_pdf":"http://arxiv.org/pdf/1808.03749v1.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":"neural-network-encapsulation","repo_url":"https://github.com/hli2020/nn_capsulation","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"neural-network-encapsulation","repo_url":"https://github.com/hli2020/nn_encapsulation","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1808.03749","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1808.03749"}},"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":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/hli2020/nn_encapsulation","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/hli2020/nn_capsulation","reach":{"status":"ok"}}],"summary":{"ran_draft_wrong":2},"by_repo_kind":{"official":{"samples":2,"ran":2,"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":2,"samples":[{"code_sha256_prefix":"c4647dd5ef46a3fd","entry":"normalize_cap","repo":"hli2020/nn_capsulation","repo_kind":"official","path":"layers/cap_layer.py","file_url":"https://github.com/hli2020/nn_capsulation/blob/HEAD/layers/cap_layer.py","link_basis":"plan_row","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"c4647dd5ef46a3fd"}},{"code_sha256_prefix":"ac7650b0121cd6c9","entry":"squash","repo":"hli2020/nn_capsulation","repo_kind":"official","path":"layers/cap_layer.py","file_url":"https://github.com/hli2020/nn_capsulation/blob/HEAD/layers/cap_layer.py","link_basis":"plan_row","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"invariant","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"ac7650b0121cd6c9"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}