{"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/generalized-zero-and-few-shot-learning-via","title":"Generalized Zero- and Few-Shot Learning via Aligned Variational Autoencoders","arxiv_id":"1812.01784","date":"2018-12-05","proceeding":null,"authors":["Edgar Schönfeld","Sayna Ebrahimi","Samarth Sinha","Trevor Darrell","Zeynep Akata"],"abstract":"Many approaches in generalized zero-shot learning rely on cross-modal mapping\nbetween the image feature space and the class embedding space. As labeled\nimages are expensive, one direction is to augment the dataset by generating\neither images or image features. However, the former misses fine-grained\ndetails and the latter requires learning a mapping associated with class\nembeddings. In this work, we take feature generation one step further and\npropose a model where a shared latent space of image features and class\nembeddings is learned by modality-specific aligned variational autoencoders.\nThis leaves us with the required discriminative information about the image and\nclasses in the latent features, on which we train a softmax classifier. The key\nto our approach is that we align the distributions learned from images and from\nside-information to construct latent features that contain the essential\nmulti-modal information associated with unseen classes. We evaluate our learned\nlatent features on several benchmark datasets, i.e. CUB, SUN, AWA1 and AWA2,\nand establish a new state of the art on generalized zero-shot as well as on\nfew-shot learning. Moreover, our results on ImageNet with various zero-shot\nsplits show that our latent features generalize well in large-scale settings.","url_abs":"http://arxiv.org/abs/1812.01784v4","url_pdf":"http://arxiv.org/pdf/1812.01784v4.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":"generalized-zero-and-few-shot-learning-via","repo_url":"https://github.com/edgarschnfld/CADA-VAE-PyTorch","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"generalized-zero-and-few-shot-learning-via","repo_url":"https://github.com/sanixa/CADA-VAE-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"few-shot-learning","task_name":"Few-Shot Learning"},{"task_slug":"generalized-few-shot-learning","task_name":"Generalized Few-Shot Learning"},{"task_slug":"generalized-zero-shot-learning","task_name":"Generalized Zero-Shot Learning"},{"task_slug":"zero-shot-skeletal-action-recognition","task_name":"Zero Shot Skeletal Action Recognition"},{"task_slug":"zero-shot-learning","task_name":"Zero-Shot Learning"}],"methods":[{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/generalized-few-shot-learning-on-awa2","task":"Generalized Few-Shot Learning","dataset":"AwA2","model":"CADA-VAE","rank_in_archive_order":2,"of":6,"metrics":{"Per-Class Accuracy (1-shot)":"69.6","Per-Class Accuracy (10-shots)":"80.2","Per-Class Accuracy (2-shots)":"73.7","Per-Class Accuracy (20-shots)":"80.9","Per-Class Accuracy (5-shots)":"78.1"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-skeletal-action-recognition-on-ntu","task":"Zero Shot Skeletal Action Recognition","dataset":"NTU RGB+D","model":"CADA-VAE","rank_in_archive_order":8,"of":9,"metrics":{"Accuracy (12 unseen classes)":"28.96","Accuracy (5 unseen classes)":"76.84","Random Split Accuracy":"60.74"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-skeletal-action-recognition-on-ntu-1","task":"Zero Shot Skeletal Action Recognition","dataset":"NTU RGB+D 120","model":"CADA-VAE","rank_in_archive_order":8,"of":9,"metrics":{"Accuracy (10 unseen classes)":"59.53","Accuracy (24 unseen classes)":"35.77","Random Split Accuracy":"45.14"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-skeletal-action-recognition-on-pku","task":"Zero Shot Skeletal Action Recognition","dataset":"PKU-MMD","model":"CADA-VAE","rank_in_archive_order":6,"of":7,"metrics":{"Random Split Accuracy":"60.74"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1812.01784","atlas_url":"https://app.syntology.ai/?focus=1812.01784","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1812.01784"}},"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/edgarschnfld/CADA-VAE-PyTorch","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/sanixa/CADA-VAE-pytorch","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran_violates":2},"by_repo_kind":{"official":{"samples":1,"ran":1,"repositories":1},"listed":{"samples":1,"ran":1,"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":0,"samples":[{"code_sha256_prefix":"7ac965f971d0020b","entry":"map_label","repo":"edgarschnfld/CADA-VAE-PyTorch","repo_kind":"official","path":"model/data_loader.py","file_url":"https://github.com/edgarschnfld/CADA-VAE-PyTorch/blob/HEAD/model/data_loader.py","link_basis":"harvester_set","language":"python","status":"ran_violates","verification_level":1,"contract_check":"VIOLATES","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"7ac965f971d0020b"}},{"code_sha256_prefix":"f017532fc389cbfe","entry":"str2bool","repo":"sanixa/CADA-VAE-pytorch","repo_kind":"listed","path":"model/single_experiment.py","file_url":"https://github.com/sanixa/CADA-VAE-pytorch/blob/HEAD/model/single_experiment.py","link_basis":"plan_row","language":"python","status":"ran_violates","verification_level":1,"contract_check":"VIOLATES","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"f017532fc389cbfe"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}