{"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/discosg-towards-discourse-level-text-scene","title":"DiscoSG: Towards Discourse-Level Text Scene Graph Parsing through Iterative Graph Refinement","arxiv_id":"2506.15583","date":"2025-06-18","proceeding":null,"authors":["Shaoqing Lin","Chong Teng","Fei Li","Donghong Ji","Lizhen Qu","Zhuang Li"],"abstract":"Vision-Language Models (VLMs) now generate discourse-level, multi-sentence visual descriptions, challenging text scene graph parsers originally designed for single-sentence caption-to-graph mapping. Current approaches typically merge sentence-level parsing outputs for discourse input, often missing phenomena like cross-sentence coreference, resulting in fragmented graphs and degraded downstream VLM task performance. To address this, we introduce a new task, Discourse-level text Scene Graph parsing (DiscoSG), supported by our dataset DiscoSG-DS, which comprises 400 expert-annotated and 8,430 synthesised multi-sentence caption-graph pairs for images. Each caption averages 9 sentences, and each graph contains at least 3 times more triples than those in existing datasets. While fine-tuning large PLMs (i.e., GPT-4) on DiscoSG-DS improves SPICE by approximately 48% over the best sentence-merging baseline, high inference cost and restrictive licensing hinder its open-source use, and smaller fine-tuned PLMs struggle with complex graphs. We propose DiscoSG-Refiner, which drafts a base graph using one small PLM, then employs a second PLM to iteratively propose graph edits, reducing full-graph generation overhead. Using two Flan-T5-Base models, DiscoSG-Refiner still improves SPICE by approximately 30% over the best baseline while achieving 86 times faster inference than GPT-4. It also consistently improves downstream VLM tasks like discourse-level caption evaluation and hallucination detection. Code and data are available at: https://github.com/ShaoqLin/DiscoSG","url_abs":"https://arxiv.org/abs/2506.15583v1","url_pdf":"https://arxiv.org/pdf/2506.15583v1.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":"discosg-towards-discourse-level-text-scene","repo_url":"https://github.com/shaoqlin/discosg","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"discosg-towards-discourse-level-text-scene","repo_url":"https://github.com/zhuang-li/factual","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"graph-generation","task_name":"Graph Generation"},{"task_slug":"hallucination","task_name":"Hallucination"},{"task_slug":"sentence","task_name":"Sentence"}],"methods":[{"method_slug":"absolute-position-encodings","method_name":"Absolute Position Encodings"},{"method_slug":"base","method_name":"BASE"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"gpt-4","method_name":"GPT-4"},{"method_slug":"label-smoothing","method_name":"Label Smoothing"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"transformer","method_name":"Transformer"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=2506.15583","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2506.15583"}},"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":"deterministic:regex_extraction","url":"https://github.com/ShaoqLin/DiscoSG","reach":{"status":"ok"}}],"summary":{"ran_violates":1,"unverified":18},"by_repo_kind":{"official":{"samples":19,"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":19,"samples":[{"code_sha256_prefix":"a834eb2f904a8668","entry":"are_tuples_match","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"capture/capture.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/capture/capture.py","link_basis":"harvester_set","language":"python","status":"ran_violates","verification_level":1,"contract_check":"VIOLATES","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"a834eb2f904a8668"}},{"code_sha256_prefix":"d6e2c6fe36767d2c","entry":"cal_agreement","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"caparena_metrics_all_score.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/caparena_metrics_all_score.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"d6e2c6fe36767d2c"}},{"code_sha256_prefix":"e6e6f6dc760babc9","entry":"cal_model_level_agreement","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"caparena_metrics_all_score.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/caparena_metrics_all_score.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"e6e6f6dc760babc9"}},{"code_sha256_prefix":"65be5c4b4da73e42","entry":"clean_text_caparena","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"convert_to_winner_style.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/convert_to_winner_style.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"65be5c4b4da73e42"}},{"code_sha256_prefix":"e5b0cb48bb329644","entry":"collect_unique_captions","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"src/factual_scene_graph/dataset_utils.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/src/factual_scene_graph/dataset_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"e5b0cb48bb329644"}},{"code_sha256_prefix":"0a2e445f0debd3d8","entry":"collect_unique_captions_flickr8k","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"src/factual_scene_graph/dataset_utils.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/src/factual_scene_graph/dataset_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"0a2e445f0debd3d8"}},{"code_sha256_prefix":"8b28307b99f57703","entry":"compute_correlation","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"src/factual_scene_graph/eval_utils.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/src/factual_scene_graph/eval_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"8b28307b99f57703"}},{"code_sha256_prefix":"66b7066c73cbe754","entry":"determine_winner","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"convert_to_winner_style.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/convert_to_winner_style.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"66b7066c73cbe754"}},{"code_sha256_prefix":"ca3043f1941bb849","entry":"extract_triples","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"src/factual_scene_graph/triple_utils.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/src/factual_scene_graph/triple_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"ca3043f1941bb849"}},{"code_sha256_prefix":"891f7322773a89aa","entry":"find_matching_metrics","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"convert_to_winner_style.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/convert_to_winner_style.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"891f7322773a89aa"}},{"code_sha256_prefix":"6962fa94d60714a9","entry":"get_synonyms","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"capture/capture.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/capture/capture.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"6962fa94d60714a9"}},{"code_sha256_prefix":"fbcbdb56cb865fc0","entry":"load_parsed_captions","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"src/factual_scene_graph/parse_utils.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/src/factual_scene_graph/parse_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"fbcbdb56cb865fc0"}},{"code_sha256_prefix":"0cc82cec90a528c2","entry":"merge_sentence_results","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"capture/capture.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/capture/capture.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"0cc82cec90a528c2"}},{"code_sha256_prefix":"5275fcf2194b983c","entry":"normalize_triple","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"src/factual_scene_graph/triple_utils.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/src/factual_scene_graph/triple_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"5275fcf2194b983c"}},{"code_sha256_prefix":"f0be599ef5e5abbf","entry":"parse_captions","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"src/factual_scene_graph/parse_utils.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/src/factual_scene_graph/parse_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"f0be599ef5e5abbf"}},{"code_sha256_prefix":"4c4b9f1c9d3001d8","entry":"parse_captions_fix","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"src/factual_scene_graph/parse_utils.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/src/factual_scene_graph/parse_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"4c4b9f1c9d3001d8"}},{"code_sha256_prefix":"a1d8dfe0ba8f9a0c","entry":"prepare_data_for_elo","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"caparena_metrics_all_score.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/caparena_metrics_all_score.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"a1d8dfe0ba8f9a0c"}},{"code_sha256_prefix":"f1cc0f7668994de6","entry":"remove_space","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"src/factual_scene_graph/triple_utils.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/src/factual_scene_graph/triple_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"f1cc0f7668994de6"}},{"code_sha256_prefix":"a5b0a0b31855c4e0","entry":"save_results_to_json","repo":"ShaoqLin/DiscoSG","repo_kind":"official","path":"src/factual_scene_graph/log_utils.py","file_url":"https://github.com/ShaoqLin/DiscoSG/blob/HEAD/src/factual_scene_graph/log_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"a5b0a0b31855c4e0"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}