{"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/glyce-glyph-vectors-for-chinese-character","title":"Glyce: Glyph-vectors for Chinese Character Representations","arxiv_id":"1901.10125","date":"2019-01-29","proceeding":"NeurIPS 2019 12","authors":["Yuxian Meng","Wei Wu","Fei Wang","Xiaoya Li","Ping Nie","Fan Yin","Muyu Li","Qinghong Han","Xiaofei Sun","Jiwei Li"],"abstract":"It is intuitive that NLP tasks for logographic languages like Chinese should benefit from the use of the glyph information in those languages. However, due to the lack of rich pictographic evidence in glyphs and the weak generalization ability of standard computer vision models on character data, an effective way to utilize the glyph information remains to be found. In this paper, we address this gap by presenting Glyce, the glyph-vectors for Chinese character representations. We make three major innovations: (1) We use historical Chinese scripts (e.g., bronzeware script, seal script, traditional Chinese, etc) to enrich the pictographic evidence in characters; (2) We design CNN structures (called tianzege-CNN) tailored to Chinese character image processing; and (3) We use image-classification as an auxiliary task in a multi-task learning setup to increase the model's ability to generalize. We show that glyph-based models are able to consistently outperform word/char ID-based models in a wide range of Chinese NLP tasks. We are able to set new state-of-the-art results for a variety of Chinese NLP tasks, including tagging (NER, CWS, POS), sentence pair classification, single sentence classification tasks, dependency parsing, and semantic role labeling. For example, the proposed model achieves an F1 score of 80.6 on the OntoNotes dataset of NER, +1.5 over BERT; it achieves an almost perfect accuracy of 99.8\\% on the Fudan corpus for text classification. Code found at https://github.com/ShannonAI/glyce.","url_abs":"https://arxiv.org/abs/1901.10125v5","url_pdf":"https://arxiv.org/pdf/1901.10125v5.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":"glyce-glyph-vectors-for-chinese-character","repo_url":"https://github.com/ShannonAI/glyce","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"glyce-glyph-vectors-for-chinese-character","repo_url":"https://github.com/zhangyuwangumass/Glyph-based-Chinese-Character-Embedding","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":null,"task_name":"Chinese Dependency Parsing"},{"task_slug":"chinese-named-entity-recognition","task_name":"Chinese Named Entity Recognition"},{"task_slug":null,"task_name":"Chinese Part-of-Speech Tagging"},{"task_slug":"chinese-semantic-role-labeling","task_name":"Chinese Semantic Role Labeling"},{"task_slug":"chinese-sentence-pair-classification","task_name":"Chinese Sentence Pair Classification"},{"task_slug":"chinese-word-segmentation","task_name":"Chinese Word Segmentation"},{"task_slug":"classification-1","task_name":"Classification"},{"task_slug":"dependency-parsing","task_name":"Dependency Parsing"},{"task_slug":"document-classification","task_name":"Document Classification"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"multi-task-learning","task_name":"Multi-Task Learning"},{"task_slug":"cg","task_name":"NER"},{"task_slug":"pos","task_name":"POS"},{"task_slug":"part-of-speech-tagging","task_name":"Part-Of-Speech Tagging"},{"task_slug":"semantic-role-labeling","task_name":"Semantic Role Labeling"},{"task_slug":"semantic-textual-similarity","task_name":"Semantic Textual Similarity"},{"task_slug":"sentence","task_name":"Sentence"},{"task_slug":"sentence-classification","task_name":"Sentence Classification"},{"task_slug":"sentence-pair-classification","task_name":"Sentence-Pair Classification"},{"task_slug":"sentiment-analysis","task_name":"Sentiment Analysis"},{"task_slug":"text-classification","task_name":"Text Classification"},{"task_slug":"image-classification","task_name":"image-classification"},{"task_slug":"text-classification-1","task_name":"text-classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/chinese-named-entity-recognition-on-msra","task":"Chinese Named Entity Recognition","dataset":"MSRA","model":"Glyce + BERT","rank_in_archive_order":8,"of":21,"metrics":{"F1":"95.54","Precision":"95.57","Recall":"95.51"},"uses_additional_data":false},{"leaderboard":"/sota/chinese-named-entity-recognition-on-ontonotes","task":"Chinese Named Entity Recognition","dataset":"OntoNotes 4","model":"Glyce + BERT","rank_in_archive_order":8,"of":15,"metrics":{"F1":"80.62","Precision":"81.87","Recall":"81.4"},"uses_additional_data":false},{"leaderboard":"/sota/chinese-named-entity-recognition-on-resume","task":"Chinese Named Entity Recognition","dataset":"Resume NER","model":"Glyce + BERT","rank_in_archive_order":5,"of":13,"metrics":{"F1":"96.54","Precision":"96.62","Recall":"96.48"},"uses_additional_data":false},{"leaderboard":"/sota/chinese-named-entity-recognition-on-weibo-ner","task":"Chinese Named Entity Recognition","dataset":"Weibo NER","model":"Glyce + BERT","rank_in_archive_order":9,"of":18,"metrics":{"F1":"67.6","Precision":"67.68","Recall":"67.71"},"uses_additional_data":false},{"leaderboard":"/sota/chinese-word-segmentation-on-as","task":"Chinese Word Segmentation","dataset":"AS","model":"Glyce + BERT","rank_in_archive_order":1,"of":2,"metrics":{"F1":"96.7","Precision":"96.6","Recall":"96.8"},"uses_additional_data":false},{"leaderboard":"/sota/chinese-word-segmentation-on-cityu","task":"Chinese Word Segmentation","dataset":"CITYU","model":"Glyce + BERT","rank_in_archive_order":2,"of":2,"metrics":{"F1":"97.9","Precision":"97.9","Recall":"98"},"uses_additional_data":false},{"leaderboard":"/sota/chinese-word-segmentation-on-msr","task":"Chinese Word Segmentation","dataset":"MSR","model":"Glyce + BERT","rank_in_archive_order":5,"of":6,"metrics":{"F1":"98.3","Precision":"98.2","Recall":"98.3"},"uses_additional_data":false},{"leaderboard":"/sota/chinese-word-segmentation-on-pku","task":"Chinese Word Segmentation","dataset":"PKU","model":"Glyce + BERT","rank_in_archive_order":2,"of":4,"metrics":{"F1":"96.7","Precision":"97.1","Recall":"96.4"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1901.10125","atlas_url":"https://app.syntology.ai/?focus=1901.10125","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1901.10125"}},"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/ShannonAI/glyce","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/zhangyuwangumass/Glyph-based-Chinese-Character-Embedding","reach":{"status":"ok"}}],"summary":{"ran_draft_wrong":1,"ran_honours":2,"unverified":5},"by_repo_kind":{"official":{"samples":8,"ran":3,"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":"0f786c407fb1ee4c","entry":"swish","repo":"ShannonAI/glyce","repo_kind":"official","path":"glyce/layers/bert_basic_model.py","file_url":"https://github.com/ShannonAI/glyce/blob/HEAD/glyce/layers/bert_basic_model.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":2,"contract_check":"MISDECLARED","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"0f786c407fb1ee4c"}},{"code_sha256_prefix":"56a9ab06b860b180","entry":"gelu","repo":"ShannonAI/glyce","repo_kind":"official","path":"glyce/layers/bert_basic_model.py","file_url":"https://github.com/ShannonAI/glyce/blob/HEAD/glyce/layers/bert_basic_model.py","link_basis":"harvester_set","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"56a9ab06b860b180"}},{"code_sha256_prefix":"50e1ffed03f484ec","entry":"gelu","repo":"ShannonAI/glyce","repo_kind":"official","path":"glyce/glyph_cnn_models/glyph_group_cnn.py","file_url":"https://github.com/ShannonAI/glyce/blob/HEAD/glyce/glyph_cnn_models/glyph_group_cnn.py","link_basis":"harvester_set","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"50e1ffed03f484ec"}},{"code_sha256_prefix":"68147246e5e691bf","entry":"channel_shuffle","repo":"ShannonAI/glyce","repo_kind":"official","path":"glyce/glyph_cnn_models/downsample.py","file_url":"https://github.com/ShannonAI/glyce/blob/HEAD/glyce/glyph_cnn_models/downsample.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"68147246e5e691bf"}},{"code_sha256_prefix":"0438badc6adb3fb8","entry":"channel_shuffle","repo":"ShannonAI/glyce","repo_kind":"official","path":"glyce/glyph_cnn_models/utils.py","file_url":"https://github.com/ShannonAI/glyce/blob/HEAD/glyce/glyph_cnn_models/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"0438badc6adb3fb8"}},{"code_sha256_prefix":"483cee36456abbed","entry":"count_params","repo":"ShannonAI/glyce","repo_kind":"official","path":"glyce/glyph_cnn_models/utils.py","file_url":"https://github.com/ShannonAI/glyce/blob/HEAD/glyce/glyph_cnn_models/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"483cee36456abbed"}},{"code_sha256_prefix":"14797096cd0b979b","entry":"masked_softmax","repo":"ShannonAI/glyce","repo_kind":"official","path":"glyce/glyph_cnn_models/self_attention.py","file_url":"https://github.com/ShannonAI/glyce/blob/HEAD/glyce/glyph_cnn_models/self_attention.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"14797096cd0b979b"}},{"code_sha256_prefix":"2f417d486bcad523","entry":"weighted_sum","repo":"ShannonAI/glyce","repo_kind":"official","path":"glyce/glyph_cnn_models/self_attention.py","file_url":"https://github.com/ShannonAI/glyce/blob/HEAD/glyce/glyph_cnn_models/self_attention.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"2f417d486bcad523"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}