{"release_id":19,"data":{"id":90729,"release_id":19,"source_id":"sgkb.science-research.nus-explainable-ai-scientific-discovery-framework-2026","kind":"article","fingerprint":"276d2bb675d06b44f773ab33e789ebf6a9fff4a7a62d65d38ab2eee164ff3140","title_en":"NUS Explainable-AI Framework for Scientific Discovery 2026","title_zh":"NUS \u9762\u5411\u79d1\u5b66\u53d1\u73b0\u7684\u53ef\u89e3\u91ca AI \u6846\u67b6 2026","domain":"science-research","status":"verified","assurance":null,"subtype":null,"lat":null,"lon":null,"date":"2026-09-02","payload":{"bundle_sha256":"45bf7e4a6e04877a9385caa9d5613e31f964959fced55c1117ce10170e4f73ce","checks":{"citations":"passed","translation":"passed"},"data":{"aliases":["NUS explainable AI framework","XAI scientific discovery Singapore","explainable AI high-stakes applications NUS"],"aliases_i18n":[],"classification_confidence":"low","classification_source":"domain-default","document_id":"sgkb.science-research.nus-explainable-ai-scientific-discovery-framework-2026","document_sha256":"1714379523916c3b095879b030072597934c2969a6a2d58f721820da0d21ccb9","document_type":"article","domain":"science-research","entities":["sg.university.nus","sg.research.nus-xai-scientific-discovery-framework"],"facets":["business","infrastructure","institution","law","policy"],"geographic_scope":["national"],"jurisdiction":"SG","language":"en","last_verified":"2026-09-02","path":"science-research\/nus-explainable-ai-scientific-discovery-framework-2026.md","related":["ai-for-science","national-ai-strategy-and-governance","university-research-ecosystem"],"sources":[{"accessed":"2026-09-02","published":"2026-09-02","publisher":"National University of Singapore, College of Design and Engineering","tier":1,"title":"NUS CDE \u2014 An explainable AI framework for scientific discovery and high-stakes applications","url":"https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/"}],"status":"verified","summary":"An NUS College of Design and Engineering Perspective proposes a workflow for turning AI-discovered patterns into testable scientific hypotheses and assessing explanations before high-stakes use.","tags":["tech","university","regulation","policy"],"temporal_scope":"current","title":"NUS Explainable-AI Framework for Scientific Discovery 2026","topics":["sgtopic.science-research.nus-explainable-ai-scientific-discovery-framework-2026"],"volatility":"slow"},"date":"2026-09-02","domain":"science-research","evidence_paths":["source\/science-research\/nus-explainable-ai-scientific-discovery-framework-2026.md","artefacts\/retrieval-chunks.jsonl","artefacts\/document-classification.jsonl"],"expected_version":12,"input_hash":"6b4473cb6eb284d4e2f96fe3b50751fc4323c5b090012ba8ecff0593c89a26e9","kind":"article","lineage":[{"chunk_id":"sgkb.science-research.nus-explainable-ai-scientific-discovery-framework-2026#overview","citation_urls":["https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/"],"content_sha256":"91eaec0810538e6480195919d8d2faf544672eecce6012cdd6b283c31fda21b3","document_id":"sgkb.science-research.nus-explainable-ai-scientific-discovery-framework-2026","document_sha256":"1714379523916c3b095879b030072597934c2969a6a2d58f721820da0d21ccb9","section_sources":[{"accessed":"2026-09-02","published":"2026-09-02","publisher":"National University of Singapore, College of Design and Engineering","tier":1,"title":"NUS CDE \u2014 An explainable AI framework for scientific discovery and high-stakes applications","url":"https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/"}]},{"chunk_id":"sgkb.science-research.nus-explainable-ai-scientific-discovery-framework-2026#from-prediction-to-testable-hypothesis","citation_urls":["https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/"],"content_sha256":"a760b73e4656a9590ed54a4cad1ff4de713d9fd99a539ceba67e3f5555b0eb5d","document_id":"sgkb.science-research.nus-explainable-ai-scientific-discovery-framework-2026","document_sha256":"1714379523916c3b095879b030072597934c2969a6a2d58f721820da0d21ccb9","section_sources":[{"accessed":"2026-09-02","published":"2026-09-02","publisher":"National University of Singapore, College of Design and Engineering","tier":1,"title":"NUS CDE \u2014 An explainable AI framework for scientific discovery and high-stakes applications","url":"https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/"}]},{"chunk_id":"sgkb.science-research.nus-explainable-ai-scientific-discovery-framework-2026#discovery-optimisation-and-certification","citation_urls":["https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/"],"content_sha256":"bb017c0895b6c4cfe38d8f9490ba2fc7b984f09c12cb5f148fdab690a1b5d8dc","document_id":"sgkb.science-research.nus-explainable-ai-scientific-discovery-framework-2026","document_sha256":"1714379523916c3b095879b030072597934c2969a6a2d58f721820da0d21ccb9","section_sources":[{"accessed":"2026-09-02","published":"2026-09-02","publisher":"National University of Singapore, College of Design and Engineering","tier":1,"title":"NUS CDE \u2014 An explainable AI framework for scientific discovery and high-stakes applications","url":"https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/"}]},{"chunk_id":"sgkb.science-research.nus-explainable-ai-scientific-discovery-framework-2026#high-stakes-boundary","citation_urls":["https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/"],"content_sha256":"c85d487e8950aae1484589ea627d6e1ddf43b4c57c402ac36f8bf14766e62ec6","document_id":"sgkb.science-research.nus-explainable-ai-scientific-discovery-framework-2026","document_sha256":"1714379523916c3b095879b030072597934c2969a6a2d58f721820da0d21ccb9","section_sources":[{"accessed":"2026-09-02","published":"2026-09-02","publisher":"National University of Singapore, College of Design and Engineering","tier":1,"title":"NUS CDE \u2014 An explainable AI framework for scientific discovery and high-stakes applications","url":"https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/"}]}],"links":[{"relation":"related","target_id":"sgkb.science-research.ai-for-science"},{"relation":"related","target_id":"sgkb.business-industry.national-ai-strategy-and-governance"},{"relation":"related","target_id":"sgkb.science-research.university-research-ecosystem"},{"relation":"topics","target_id":"sgtopic.science-research.nus-explainable-ai-scientific-discovery-framework-2026"},{"relation":"entities","target_id":"sg.university.nus"},{"relation":"entities","target_id":"sg.research.nus-xai-scientific-discovery-framework"}],"locales":{"en":{"markdown":"# NUS Explainable-AI Framework for Scientific Discovery 2026\n\nThe National University of Singapore\u2019s College of Design and Engineering (NUS CDE) reported a **Perspective** proposing a framework for using explainable artificial intelligence (**XAI**) in scientific discovery and high-stakes applications. The Perspective was co-authored by NUS Assistant Professor Gianmarco Mengaldo and published in *Nature Communications* on **6 August 2026**; NUS published its Singapore research feature on **2 September 2026** ([NUS CDE, 2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)).\n\n## From prediction to testable hypothesis\n\nThe proposed workflow treats an AI explanation as a lead for investigation rather than as a scientific explanation that can be accepted on appearance alone. Researchers would identify which inputs or patterns influenced a model\u2019s prediction, translate that clue into a hypothesis, and then test it through experiments, simulations or established scientific principles. NUS illustrates the idea with AI analysis of turbulent airflow: an explanation can identify a feature worth investigating, while further testing determines whether it reflects a genuine physical process or merely a recurring pattern in the training data ([NUS CDE, 2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)).\n\n## Discovery, optimisation and certification\n\nNUS describes three linked uses for the approach. Explainability can support **discovery** by surfacing patterns that suggest how a system functions; tested insights can support **optimisation** by improving a design or control strategy; and records of what influenced a model can contribute to **certification** and reliability assessment. In a reported turbulence example, XAI identified influential airflow patterns and helped researchers train a second AI system to reduce drag more effectively than a system trained only on the final drag outcome ([NUS CDE, 2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)).\n\n## High-stakes boundary\n\nFor healthcare, aviation, energy, infrastructure and autonomous systems, the feature says evaluators would need to examine both model accuracy and whether an explanation faithfully reflects what influenced the model, including under unfamiliar conditions. Explanations may strengthen safety and regulatory assessments, but NUS explicitly presents them as supplementary evidence rather than replacements for existing testing, certification or regulatory processes. The article describes a proposed framework and ongoing exploration, not a deployed regulatory standard or a claim that the method has already validated every high-stakes AI system ([NUS CDE, 2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)).","summary":"An NUS College of Design and Engineering Perspective proposes a workflow for turning AI-discovered patterns into testable scientific hypotheses and assessing explanations before high-stakes use.","title":"NUS Explainable-AI Framework for Scientific Discovery 2026"},"zh":{"markdown":"# NUS \u9762\u5411\u79d1\u5b66\u53d1\u73b0\u7684\u53ef\u89e3\u91ca AI \u6846\u67b6 2026\n\n\u65b0\u52a0\u5761\u56fd\u7acb\u5927\u5b66\u8bbe\u8ba1\u4e0e\u5de5\u7a0b\u5b66\u9662\uff08NUS CDE\uff09\u62a5\u544a\u4e86\u4e00\u7bc7 **\u89c2\u70b9\u6587\u7ae0**\uff0c\u63d0\u51fa\u5728\u79d1\u5b66\u53d1\u73b0\u548c\u9ad8\u98ce\u9669\u5e94\u7528\u4e2d\u4f7f\u7528\u53ef\u89e3\u91ca\u4eba\u5de5\u667a\u80fd\uff08**XAI**\uff09\u7684\u6846\u67b6\u3002\u8be5\u89c2\u70b9\u6587\u7ae0\u7531 NUS \u52a9\u7406\u6559\u6388 Gianmarco Mengaldo \u5408\u8457\uff0c\u4e8e **6 August 2026** \u53d1\u8868\u5728 *Nature Communications*\uff1bNUS \u4e8e **2 September 2026** \u53d1\u5e03\u4e86\u65b0\u52a0\u5761\u7814\u7a76\u4e13\u9898\u62a5\u9053\uff08[NUS CDE\uff0c2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)\uff09\u3002\n\n## \u4ece\u9884\u6d4b\u5230\u53ef\u68c0\u9a8c\u7684\u5047\u8bbe\n\n\u8fd9\u5957\u62df\u8bae\u5de5\u4f5c\u6d41\u7a0b\u5c06 AI \u89e3\u91ca\u89c6\u4e3a\u8c03\u67e5\u7ebf\u7d22\uff0c\u800c\u4e0d\u662f\u4ec5\u51ed\u8868\u9762\u73b0\u8c61\u5c31\u80fd\u63a5\u53d7\u7684\u79d1\u5b66\u89e3\u91ca\u3002\u7814\u7a76\u4eba\u5458\u4f1a\u627e\u51fa\u54ea\u4e9b\u8f93\u5165\u6216\u6a21\u5f0f\u5f71\u54cd\u4e86\u6a21\u578b\u7684\u9884\u6d4b\uff0c\u5c06\u8fd9\u4e00\u7ebf\u7d22\u8f6c\u5316\u4e3a\u5047\u8bbe\uff0c\u7136\u540e\u901a\u8fc7\u5b9e\u9a8c\u3001\u6a21\u62df\u6216\u65e2\u6709\u79d1\u5b66\u539f\u7406\u8fdb\u884c\u68c0\u9a8c\u3002NUS \u4ee5 AI \u5206\u6790\u6e4d\u6d41\u6c14\u6d41\u4e3a\u4f8b\u8bf4\u660e\u8fd9\u4e00\u601d\u8def\uff1a\u89e3\u91ca\u53ef\u4ee5\u6307\u51fa\u503c\u5f97\u7814\u7a76\u7684\u7279\u5f81\uff0c\u800c\u8fdb\u4e00\u6b65\u68c0\u9a8c\u5219\u53ef\u786e\u5b9a\u5b83\u53cd\u6620\u7684\u662f\u771f\u5b9e\u7269\u7406\u8fc7\u7a0b\uff0c\u8fd8\u662f\u8bad\u7ec3\u6570\u636e\u4e2d\u53cd\u590d\u51fa\u73b0\u7684\u6a21\u5f0f\uff08[NUS CDE\uff0c2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)\uff09\u3002\n\n## \u53d1\u73b0\u3001\u4f18\u5316\u4e0e\u8ba4\u8bc1\n\nNUS \u63cf\u8ff0\u4e86\u8fd9\u79cd\u65b9\u6cd5\u7684\u4e09\u79cd\u5173\u8054\u7528\u9014\u3002\u53ef\u89e3\u91ca\u6027\u53ef\u4ee5\u901a\u8fc7\u63ed\u793a\u6697\u793a\u7cfb\u7edf\u8fd0\u884c\u65b9\u5f0f\u7684\u6a21\u5f0f\u6765\u652f\u6301 **\u53d1\u73b0**\uff1b\u7ecf\u8fc7\u68c0\u9a8c\u7684\u89c1\u89e3\u53ef\u4ee5\u901a\u8fc7\u6539\u8fdb\u8bbe\u8ba1\u6216\u63a7\u5236\u7b56\u7565\u6765\u652f\u6301 **\u4f18\u5316**\uff1b\u8bb0\u5f55\u54ea\u4e9b\u56e0\u7d20\u5f71\u54cd\u4e86\u6a21\u578b\uff0c\u5219\u6709\u52a9\u4e8e **\u8ba4\u8bc1** \u548c\u53ef\u9760\u6027\u8bc4\u4f30\u3002\u5728\u4e00\u9879\u5df2\u62a5\u544a\u7684\u6e4d\u6d41\u6848\u4f8b\u4e2d\uff0cXAI \u627e\u5230\u4e86\u6709\u5f71\u54cd\u529b\u7684\u6c14\u6d41\u6a21\u5f0f\uff0c\u5e76\u5e2e\u52a9\u7814\u7a76\u4eba\u5458\u8bad\u7ec3\u7b2c\u4e8c\u4e2a AI \u7cfb\u7edf\uff0c\u4f7f\u5176\u6bd4\u4ec5\u6839\u636e\u6700\u7ec8\u963b\u529b\u7ed3\u679c\u8bad\u7ec3\u7684\u7cfb\u7edf\u66f4\u6709\u6548\u5730\u964d\u4f4e\u963b\u529b\uff08[NUS CDE\uff0c2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)\uff09\u3002\n\n## \u9ad8\u98ce\u9669\u5e94\u7528\u7684\u754c\u9650\n\n\u5bf9\u4e8e\u533b\u7597\u4fdd\u5065\u3001\u822a\u7a7a\u3001\u80fd\u6e90\u3001\u57fa\u7840\u8bbe\u65bd\u548c\u81ea\u4e3b\u7cfb\u7edf\uff0c\u8be5\u4e13\u9898\u62a5\u9053\u6307\u51fa\uff0c\u8bc4\u4f30\u8005\u9700\u8981\u8003\u5bdf\u6a21\u578b\u51c6\u786e\u6027\uff0c\u4ee5\u53ca\u89e3\u91ca\u662f\u5426\u5fe0\u5b9e\u53cd\u6620\u4e86\u5f71\u54cd\u6a21\u578b\u7684\u56e0\u7d20\uff0c\u5305\u62ec\u5728\u964c\u751f\u6761\u4ef6\u4e0b\u7684\u8868\u73b0\u3002\u89e3\u91ca\u53ef\u4ee5\u52a0\u5f3a\u5b89\u5168\u548c\u76d1\u7ba1\u8bc4\u4f30\uff0c\u4f46 NUS \u660e\u786e\u5c06\u5176\u4f5c\u4e3a\u8865\u5145\u8bc1\u636e\uff0c\u800c\u975e\u73b0\u6709\u6d4b\u8bd5\u3001\u8ba4\u8bc1\u6216\u76d1\u7ba1\u6d41\u7a0b\u7684\u66ff\u4ee3\u54c1\u3002\u6587\u7ae0\u4ecb\u7ecd\u7684\u662f\u4e00\u5957\u62df\u8bae\u6846\u67b6\u548c\u6b63\u5728\u8fdb\u884c\u7684\u63a2\u7d22\uff0c\u5e76\u975e\u5df2\u90e8\u7f72\u7684\u76d1\u7ba1\u6807\u51c6\uff0c\u4e5f\u6ca1\u6709\u58f0\u79f0\u8be5\u65b9\u6cd5\u5df2\u7ecf\u9a8c\u8bc1\u4e86\u6bcf\u4e2a\u9ad8\u98ce\u9669 AI \u7cfb\u7edf\uff08[NUS CDE\uff0c2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)\uff09\u3002","summary":"\u65b0\u52a0\u5761\u56fd\u7acb\u5927\u5b66\u8bbe\u8ba1\u4e0e\u5de5\u7a0b\u5b66\u9662\u7684\u4e00\u7bc7\u89c2\u70b9\u6587\u7ae0\u63d0\u51fa\u4e86\u4e00\u5957\u5de5\u4f5c\u6d41\u7a0b\uff0c\u7528\u4e8e\u5c06 AI \u53d1\u73b0\u7684\u6a21\u5f0f\u8f6c\u5316\u4e3a\u53ef\u68c0\u9a8c\u7684\u79d1\u5b66\u5047\u8bbe\uff0c\u5e76\u5728\u9ad8\u98ce\u9669\u5e94\u7528\u524d\u8bc4\u4f30\u89e3\u91ca\u3002","title":"NUS \u9762\u5411\u79d1\u5b66\u53d1\u73b0\u7684\u53ef\u89e3\u91ca AI \u6846\u67b6 2026"}},"source_id":"sgkb.science-research.nus-explainable-ai-scientific-discovery-framework-2026","status":"verified","summary_en":"An NUS College of Design and Engineering Perspective proposes a workflow for turning AI-discovered patterns into testable scientific hypotheses and assessing explanations before high-stakes use.","summary_zh":"\u65b0\u52a0\u5761\u56fd\u7acb\u5927\u5b66\u8bbe\u8ba1\u4e0e\u5de5\u7a0b\u5b66\u9662\u7684\u4e00\u7bc7\u89c2\u70b9\u6587\u7ae0\u63d0\u51fa\u4e86\u4e00\u5957\u5de5\u4f5c\u6d41\u7a0b\uff0c\u7528\u4e8e\u5c06 AI \u53d1\u73b0\u7684\u6a21\u5f0f\u8f6c\u5316\u4e3a\u53ef\u68c0\u9a8c\u7684\u79d1\u5b66\u5047\u8bbe\uff0c\u5e76\u5728\u9ad8\u98ce\u9669\u5e94\u7528\u524d\u8bc4\u4f30\u89e3\u91ca\u3002","title_en":"NUS Explainable-AI Framework for Scientific Discovery 2026","title_zh":"NUS \u9762\u5411\u79d1\u5b66\u53d1\u73b0\u7684\u53ef\u89e3\u91ca AI \u6846\u67b6 2026","wiki":{"domain":"science-research","last_verified":"2026-09-02","locales":{"en":{"markdown":"# NUS Explainable-AI Framework for Scientific Discovery 2026\n\nThe National University of Singapore\u2019s College of Design and Engineering (NUS CDE) reported a **Perspective** proposing a framework for using explainable artificial intelligence (**XAI**) in scientific discovery and high-stakes applications. The Perspective was co-authored by NUS Assistant Professor Gianmarco Mengaldo and published in *Nature Communications* on **6 August 2026**; NUS published its Singapore research feature on **2 September 2026** ([NUS CDE, 2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)).\n\n## From prediction to testable hypothesis\n\nThe proposed workflow treats an AI explanation as a lead for investigation rather than as a scientific explanation that can be accepted on appearance alone. Researchers would identify which inputs or patterns influenced a model\u2019s prediction, translate that clue into a hypothesis, and then test it through experiments, simulations or established scientific principles. NUS illustrates the idea with AI analysis of turbulent airflow: an explanation can identify a feature worth investigating, while further testing determines whether it reflects a genuine physical process or merely a recurring pattern in the training data ([NUS CDE, 2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)).\n\n## Discovery, optimisation and certification\n\nNUS describes three linked uses for the approach. Explainability can support **discovery** by surfacing patterns that suggest how a system functions; tested insights can support **optimisation** by improving a design or control strategy; and records of what influenced a model can contribute to **certification** and reliability assessment. In a reported turbulence example, XAI identified influential airflow patterns and helped researchers train a second AI system to reduce drag more effectively than a system trained only on the final drag outcome ([NUS CDE, 2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)).\n\n## High-stakes boundary\n\nFor healthcare, aviation, energy, infrastructure and autonomous systems, the feature says evaluators would need to examine both model accuracy and whether an explanation faithfully reflects what influenced the model, including under unfamiliar conditions. Explanations may strengthen safety and regulatory assessments, but NUS explicitly presents them as supplementary evidence rather than replacements for existing testing, certification or regulatory processes. The article describes a proposed framework and ongoing exploration, not a deployed regulatory standard or a claim that the method has already validated every high-stakes AI system ([NUS CDE, 2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)).","summary":"An NUS College of Design and Engineering Perspective proposes a workflow for turning AI-discovered patterns into testable scientific hypotheses and assessing explanations before high-stakes use.","title":"NUS Explainable-AI Framework for Scientific Discovery 2026"},"zh":{"markdown":"# NUS \u9762\u5411\u79d1\u5b66\u53d1\u73b0\u7684\u53ef\u89e3\u91ca AI \u6846\u67b6 2026\n\n\u65b0\u52a0\u5761\u56fd\u7acb\u5927\u5b66\u8bbe\u8ba1\u4e0e\u5de5\u7a0b\u5b66\u9662\uff08NUS CDE\uff09\u62a5\u544a\u4e86\u4e00\u7bc7 **\u89c2\u70b9\u6587\u7ae0**\uff0c\u63d0\u51fa\u5728\u79d1\u5b66\u53d1\u73b0\u548c\u9ad8\u98ce\u9669\u5e94\u7528\u4e2d\u4f7f\u7528\u53ef\u89e3\u91ca\u4eba\u5de5\u667a\u80fd\uff08**XAI**\uff09\u7684\u6846\u67b6\u3002\u8be5\u89c2\u70b9\u6587\u7ae0\u7531 NUS \u52a9\u7406\u6559\u6388 Gianmarco Mengaldo \u5408\u8457\uff0c\u4e8e **6 August 2026** \u53d1\u8868\u5728 *Nature Communications*\uff1bNUS \u4e8e **2 September 2026** \u53d1\u5e03\u4e86\u65b0\u52a0\u5761\u7814\u7a76\u4e13\u9898\u62a5\u9053\uff08[NUS CDE\uff0c2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)\uff09\u3002\n\n## \u4ece\u9884\u6d4b\u5230\u53ef\u68c0\u9a8c\u7684\u5047\u8bbe\n\n\u8fd9\u5957\u62df\u8bae\u5de5\u4f5c\u6d41\u7a0b\u5c06 AI \u89e3\u91ca\u89c6\u4e3a\u8c03\u67e5\u7ebf\u7d22\uff0c\u800c\u4e0d\u662f\u4ec5\u51ed\u8868\u9762\u73b0\u8c61\u5c31\u80fd\u63a5\u53d7\u7684\u79d1\u5b66\u89e3\u91ca\u3002\u7814\u7a76\u4eba\u5458\u4f1a\u627e\u51fa\u54ea\u4e9b\u8f93\u5165\u6216\u6a21\u5f0f\u5f71\u54cd\u4e86\u6a21\u578b\u7684\u9884\u6d4b\uff0c\u5c06\u8fd9\u4e00\u7ebf\u7d22\u8f6c\u5316\u4e3a\u5047\u8bbe\uff0c\u7136\u540e\u901a\u8fc7\u5b9e\u9a8c\u3001\u6a21\u62df\u6216\u65e2\u6709\u79d1\u5b66\u539f\u7406\u8fdb\u884c\u68c0\u9a8c\u3002NUS \u4ee5 AI \u5206\u6790\u6e4d\u6d41\u6c14\u6d41\u4e3a\u4f8b\u8bf4\u660e\u8fd9\u4e00\u601d\u8def\uff1a\u89e3\u91ca\u53ef\u4ee5\u6307\u51fa\u503c\u5f97\u7814\u7a76\u7684\u7279\u5f81\uff0c\u800c\u8fdb\u4e00\u6b65\u68c0\u9a8c\u5219\u53ef\u786e\u5b9a\u5b83\u53cd\u6620\u7684\u662f\u771f\u5b9e\u7269\u7406\u8fc7\u7a0b\uff0c\u8fd8\u662f\u8bad\u7ec3\u6570\u636e\u4e2d\u53cd\u590d\u51fa\u73b0\u7684\u6a21\u5f0f\uff08[NUS CDE\uff0c2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)\uff09\u3002\n\n## \u53d1\u73b0\u3001\u4f18\u5316\u4e0e\u8ba4\u8bc1\n\nNUS \u63cf\u8ff0\u4e86\u8fd9\u79cd\u65b9\u6cd5\u7684\u4e09\u79cd\u5173\u8054\u7528\u9014\u3002\u53ef\u89e3\u91ca\u6027\u53ef\u4ee5\u901a\u8fc7\u63ed\u793a\u6697\u793a\u7cfb\u7edf\u8fd0\u884c\u65b9\u5f0f\u7684\u6a21\u5f0f\u6765\u652f\u6301 **\u53d1\u73b0**\uff1b\u7ecf\u8fc7\u68c0\u9a8c\u7684\u89c1\u89e3\u53ef\u4ee5\u901a\u8fc7\u6539\u8fdb\u8bbe\u8ba1\u6216\u63a7\u5236\u7b56\u7565\u6765\u652f\u6301 **\u4f18\u5316**\uff1b\u8bb0\u5f55\u54ea\u4e9b\u56e0\u7d20\u5f71\u54cd\u4e86\u6a21\u578b\uff0c\u5219\u6709\u52a9\u4e8e **\u8ba4\u8bc1** \u548c\u53ef\u9760\u6027\u8bc4\u4f30\u3002\u5728\u4e00\u9879\u5df2\u62a5\u544a\u7684\u6e4d\u6d41\u6848\u4f8b\u4e2d\uff0cXAI \u627e\u5230\u4e86\u6709\u5f71\u54cd\u529b\u7684\u6c14\u6d41\u6a21\u5f0f\uff0c\u5e76\u5e2e\u52a9\u7814\u7a76\u4eba\u5458\u8bad\u7ec3\u7b2c\u4e8c\u4e2a AI \u7cfb\u7edf\uff0c\u4f7f\u5176\u6bd4\u4ec5\u6839\u636e\u6700\u7ec8\u963b\u529b\u7ed3\u679c\u8bad\u7ec3\u7684\u7cfb\u7edf\u66f4\u6709\u6548\u5730\u964d\u4f4e\u963b\u529b\uff08[NUS CDE\uff0c2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)\uff09\u3002\n\n## \u9ad8\u98ce\u9669\u5e94\u7528\u7684\u754c\u9650\n\n\u5bf9\u4e8e\u533b\u7597\u4fdd\u5065\u3001\u822a\u7a7a\u3001\u80fd\u6e90\u3001\u57fa\u7840\u8bbe\u65bd\u548c\u81ea\u4e3b\u7cfb\u7edf\uff0c\u8be5\u4e13\u9898\u62a5\u9053\u6307\u51fa\uff0c\u8bc4\u4f30\u8005\u9700\u8981\u8003\u5bdf\u6a21\u578b\u51c6\u786e\u6027\uff0c\u4ee5\u53ca\u89e3\u91ca\u662f\u5426\u5fe0\u5b9e\u53cd\u6620\u4e86\u5f71\u54cd\u6a21\u578b\u7684\u56e0\u7d20\uff0c\u5305\u62ec\u5728\u964c\u751f\u6761\u4ef6\u4e0b\u7684\u8868\u73b0\u3002\u89e3\u91ca\u53ef\u4ee5\u52a0\u5f3a\u5b89\u5168\u548c\u76d1\u7ba1\u8bc4\u4f30\uff0c\u4f46 NUS \u660e\u786e\u5c06\u5176\u4f5c\u4e3a\u8865\u5145\u8bc1\u636e\uff0c\u800c\u975e\u73b0\u6709\u6d4b\u8bd5\u3001\u8ba4\u8bc1\u6216\u76d1\u7ba1\u6d41\u7a0b\u7684\u66ff\u4ee3\u54c1\u3002\u6587\u7ae0\u4ecb\u7ecd\u7684\u662f\u4e00\u5957\u62df\u8bae\u6846\u67b6\u548c\u6b63\u5728\u8fdb\u884c\u7684\u63a2\u7d22\uff0c\u5e76\u975e\u5df2\u90e8\u7f72\u7684\u76d1\u7ba1\u6807\u51c6\uff0c\u4e5f\u6ca1\u6709\u58f0\u79f0\u8be5\u65b9\u6cd5\u5df2\u7ecf\u9a8c\u8bc1\u4e86\u6bcf\u4e2a\u9ad8\u98ce\u9669 AI \u7cfb\u7edf\uff08[NUS CDE\uff0c2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)\uff09\u3002","summary":"\u65b0\u52a0\u5761\u56fd\u7acb\u5927\u5b66\u8bbe\u8ba1\u4e0e\u5de5\u7a0b\u5b66\u9662\u7684\u4e00\u7bc7\u89c2\u70b9\u6587\u7ae0\u63d0\u51fa\u4e86\u4e00\u5957\u5de5\u4f5c\u6d41\u7a0b\uff0c\u7528\u4e8e\u5c06 AI \u53d1\u73b0\u7684\u6a21\u5f0f\u8f6c\u5316\u4e3a\u53ef\u68c0\u9a8c\u7684\u79d1\u5b66\u5047\u8bbe\uff0c\u5e76\u5728\u9ad8\u98ce\u9669\u5e94\u7528\u524d\u8bc4\u4f30\u89e3\u91ca\u3002","title":"NUS \u9762\u5411\u79d1\u5b66\u53d1\u73b0\u7684\u53ef\u89e3\u91ca AI \u6846\u67b6 2026"}},"path":"science-research\/nus-explainable-ai-scientific-discovery-framework-2026.md","related_source_ids":["sgkb.science-research.ai-for-science","sgkb.business-industry.national-ai-strategy-and-governance","sgkb.science-research.university-research-ecosystem"],"source_hash":"6b4473cb6eb284d4e2f96fe3b50751fc4323c5b090012ba8ecff0593c89a26e9","source_revision":"094b7bf3a1ab75da7b1952ee33089c1733835fe6","sources":[{"title":"NUS CDE \u2014 An explainable AI framework for scientific discovery and high-stakes applications","url":"https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/"}]}},"search_text":"NUS Explainable-AI Framework for Scientific Discovery 2026\nNUS \u9762\u5411\u79d1\u5b66\u53d1\u73b0\u7684\u53ef\u89e3\u91ca AI \u6846\u67b6 2026\nAn NUS College of Design and Engineering Perspective proposes a workflow for turning AI-discovered patterns into testable scientific hypotheses and assessing explanations before high-stakes use.\n\u65b0\u52a0\u5761\u56fd\u7acb\u5927\u5b66\u8bbe\u8ba1\u4e0e\u5de5\u7a0b\u5b66\u9662\u7684\u4e00\u7bc7\u89c2\u70b9\u6587\u7ae0\u63d0\u51fa\u4e86\u4e00\u5957\u5de5\u4f5c\u6d41\u7a0b\uff0c\u7528\u4e8e\u5c06 AI \u53d1\u73b0\u7684\u6a21\u5f0f\u8f6c\u5316\u4e3a\u53ef\u68c0\u9a8c\u7684\u79d1\u5b66\u5047\u8bbe\uff0c\u5e76\u5728\u9ad8\u98ce\u9669\u5e94\u7528\u524d\u8bc4\u4f30\u89e3\u91ca\u3002\n{\"aliases\":[\"NUS explainable AI framework\",\"XAI scientific discovery Singapore\",\"explainable AI high-stakes applications NUS\"],\"aliases_i18n\":[],\"classification_confidence\":\"low\",\"classification_source\":\"domain-default\",\"document_id\":\"sgkb.science-research.nus-explainable-ai-scientific-discovery-framework-2026\",\"document_sha256\":\"1714379523916c3b095879b030072597934c2969a6a2d58f721820da0d21ccb9\",\"document_type\":\"article\",\"domain\":\"science-research\",\"entities\":[\"sg.university.nus\",\"sg.research.nus-xai-scientific-discovery-framework\"],\"facets\":[\"business\",\"infrastructure\",\"institution\",\"law\",\"policy\"],\"geographic_scope\":[\"national\"],\"jurisdiction\":\"SG\",\"language\":\"en\",\"last_verified\":\"2026-09-02\",\"path\":\"science-research\\\/nus-explainable-ai-scientific-discovery-framework-2026.md\",\"related\":[\"ai-for-science\",\"national-ai-strategy-and-governance\",\"university-research-ecosystem\"],\"sources\":[{\"accessed\":\"2026-09-02\",\"published\":\"2026-09-02\",\"publisher\":\"National University of Singapore, College of Design and Engineering\",\"tier\":1,\"title\":\"NUS CDE \u2014 An explainable AI framework for scientific discovery and high-stakes applications\",\"url\":\"https:\\\/\\\/cde.nus.edu.sg\\\/news\\\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\\\/\"}],\"status\":\"verified\",\"summary\":\"An NUS College of Design and Engineering Perspective proposes a workflow for turning AI-discovered patterns into testable scientific hypotheses and assessing explanations before high-stakes use.\",\"tags\":[\"tech\",\"university\",\"regulation\",\"policy\"],\"temporal_scope\":\"current\",\"title\":\"NUS Explainable-AI Framework for Scientific Discovery 2026\",\"topics\":[\"sgtopic.science-research.nus-explainable-ai-scientific-discovery-framework-2026\"],\"volatility\":\"slow\"}\n# NUS Explainable-AI Framework for Scientific Discovery 2026\n\nThe National University of Singapore\u2019s College of Design and Engineering (NUS CDE) reported a **Perspective** proposing a framework for using explainable artificial intelligence (**XAI**) in scientific discovery and high-stakes applications. The Perspective was co-authored by NUS Assistant Professor Gianmarco Mengaldo and published in *Nature Communications* on **6 August 2026**; NUS published its Singapore research feature on **2 September 2026** ([NUS CDE, 2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)).\n\n## From prediction to testable hypothesis\n\nThe proposed workflow treats an AI explanation as a lead for investigation rather than as a scientific explanation that can be accepted on appearance alone. Researchers would identify which inputs or patterns influenced a model\u2019s prediction, translate that clue into a hypothesis, and then test it through experiments, simulations or established scientific principles. NUS illustrates the idea with AI analysis of turbulent airflow: an explanation can identify a feature worth investigating, while further testing determines whether it reflects a genuine physical process or merely a recurring pattern in the training data ([NUS CDE, 2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)).\n\n## Discovery, optimisation and certification\n\nNUS describes three linked uses for the approach. Explainability can support **discovery** by surfacing patterns that suggest how a system functions; tested insights can support **optimisation** by improving a design or control strategy; and records of what influenced a model can contribute to **certification** and reliability assessment. In a reported turbulence example, XAI identified influential airflow patterns and helped researchers train a second AI system to reduce drag more effectively than a system trained only on the final drag outcome ([NUS CDE, 2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)).\n\n## High-stakes boundary\n\nFor healthcare, aviation, energy, infrastructure and autonomous systems, the feature says evaluators would need to examine both model accuracy and whether an explanation faithfully reflects what influenced the model, including under unfamiliar conditions. Explanations may strengthen safety and regulatory assessments, but NUS explicitly presents them as supplementary evidence rather than replacements for existing testing, certification or regulatory processes. The article describes a proposed framework and ongoing exploration, not a deployed regulatory standard or a claim that the method has already validated every high-stakes AI system ([NUS CDE, 2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)).\n# NUS \u9762\u5411\u79d1\u5b66\u53d1\u73b0\u7684\u53ef\u89e3\u91ca AI \u6846\u67b6 2026\n\n\u65b0\u52a0\u5761\u56fd\u7acb\u5927\u5b66\u8bbe\u8ba1\u4e0e\u5de5\u7a0b\u5b66\u9662\uff08NUS CDE\uff09\u62a5\u544a\u4e86\u4e00\u7bc7 **\u89c2\u70b9\u6587\u7ae0**\uff0c\u63d0\u51fa\u5728\u79d1\u5b66\u53d1\u73b0\u548c\u9ad8\u98ce\u9669\u5e94\u7528\u4e2d\u4f7f\u7528\u53ef\u89e3\u91ca\u4eba\u5de5\u667a\u80fd\uff08**XAI**\uff09\u7684\u6846\u67b6\u3002\u8be5\u89c2\u70b9\u6587\u7ae0\u7531 NUS \u52a9\u7406\u6559\u6388 Gianmarco Mengaldo \u5408\u8457\uff0c\u4e8e **6 August 2026** \u53d1\u8868\u5728 *Nature Communications*\uff1bNUS \u4e8e **2 September 2026** \u53d1\u5e03\u4e86\u65b0\u52a0\u5761\u7814\u7a76\u4e13\u9898\u62a5\u9053\uff08[NUS CDE\uff0c2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)\uff09\u3002\n\n## \u4ece\u9884\u6d4b\u5230\u53ef\u68c0\u9a8c\u7684\u5047\u8bbe\n\n\u8fd9\u5957\u62df\u8bae\u5de5\u4f5c\u6d41\u7a0b\u5c06 AI \u89e3\u91ca\u89c6\u4e3a\u8c03\u67e5\u7ebf\u7d22\uff0c\u800c\u4e0d\u662f\u4ec5\u51ed\u8868\u9762\u73b0\u8c61\u5c31\u80fd\u63a5\u53d7\u7684\u79d1\u5b66\u89e3\u91ca\u3002\u7814\u7a76\u4eba\u5458\u4f1a\u627e\u51fa\u54ea\u4e9b\u8f93\u5165\u6216\u6a21\u5f0f\u5f71\u54cd\u4e86\u6a21\u578b\u7684\u9884\u6d4b\uff0c\u5c06\u8fd9\u4e00\u7ebf\u7d22\u8f6c\u5316\u4e3a\u5047\u8bbe\uff0c\u7136\u540e\u901a\u8fc7\u5b9e\u9a8c\u3001\u6a21\u62df\u6216\u65e2\u6709\u79d1\u5b66\u539f\u7406\u8fdb\u884c\u68c0\u9a8c\u3002NUS \u4ee5 AI \u5206\u6790\u6e4d\u6d41\u6c14\u6d41\u4e3a\u4f8b\u8bf4\u660e\u8fd9\u4e00\u601d\u8def\uff1a\u89e3\u91ca\u53ef\u4ee5\u6307\u51fa\u503c\u5f97\u7814\u7a76\u7684\u7279\u5f81\uff0c\u800c\u8fdb\u4e00\u6b65\u68c0\u9a8c\u5219\u53ef\u786e\u5b9a\u5b83\u53cd\u6620\u7684\u662f\u771f\u5b9e\u7269\u7406\u8fc7\u7a0b\uff0c\u8fd8\u662f\u8bad\u7ec3\u6570\u636e\u4e2d\u53cd\u590d\u51fa\u73b0\u7684\u6a21\u5f0f\uff08[NUS CDE\uff0c2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)\uff09\u3002\n\n## \u53d1\u73b0\u3001\u4f18\u5316\u4e0e\u8ba4\u8bc1\n\nNUS \u63cf\u8ff0\u4e86\u8fd9\u79cd\u65b9\u6cd5\u7684\u4e09\u79cd\u5173\u8054\u7528\u9014\u3002\u53ef\u89e3\u91ca\u6027\u53ef\u4ee5\u901a\u8fc7\u63ed\u793a\u6697\u793a\u7cfb\u7edf\u8fd0\u884c\u65b9\u5f0f\u7684\u6a21\u5f0f\u6765\u652f\u6301 **\u53d1\u73b0**\uff1b\u7ecf\u8fc7\u68c0\u9a8c\u7684\u89c1\u89e3\u53ef\u4ee5\u901a\u8fc7\u6539\u8fdb\u8bbe\u8ba1\u6216\u63a7\u5236\u7b56\u7565\u6765\u652f\u6301 **\u4f18\u5316**\uff1b\u8bb0\u5f55\u54ea\u4e9b\u56e0\u7d20\u5f71\u54cd\u4e86\u6a21\u578b\uff0c\u5219\u6709\u52a9\u4e8e **\u8ba4\u8bc1** \u548c\u53ef\u9760\u6027\u8bc4\u4f30\u3002\u5728\u4e00\u9879\u5df2\u62a5\u544a\u7684\u6e4d\u6d41\u6848\u4f8b\u4e2d\uff0cXAI \u627e\u5230\u4e86\u6709\u5f71\u54cd\u529b\u7684\u6c14\u6d41\u6a21\u5f0f\uff0c\u5e76\u5e2e\u52a9\u7814\u7a76\u4eba\u5458\u8bad\u7ec3\u7b2c\u4e8c\u4e2a AI \u7cfb\u7edf\uff0c\u4f7f\u5176\u6bd4\u4ec5\u6839\u636e\u6700\u7ec8\u963b\u529b\u7ed3\u679c\u8bad\u7ec3\u7684\u7cfb\u7edf\u66f4\u6709\u6548\u5730\u964d\u4f4e\u963b\u529b\uff08[NUS CDE\uff0c2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)\uff09\u3002\n\n## \u9ad8\u98ce\u9669\u5e94\u7528\u7684\u754c\u9650\n\n\u5bf9\u4e8e\u533b\u7597\u4fdd\u5065\u3001\u822a\u7a7a\u3001\u80fd\u6e90\u3001\u57fa\u7840\u8bbe\u65bd\u548c\u81ea\u4e3b\u7cfb\u7edf\uff0c\u8be5\u4e13\u9898\u62a5\u9053\u6307\u51fa\uff0c\u8bc4\u4f30\u8005\u9700\u8981\u8003\u5bdf\u6a21\u578b\u51c6\u786e\u6027\uff0c\u4ee5\u53ca\u89e3\u91ca\u662f\u5426\u5fe0\u5b9e\u53cd\u6620\u4e86\u5f71\u54cd\u6a21\u578b\u7684\u56e0\u7d20\uff0c\u5305\u62ec\u5728\u964c\u751f\u6761\u4ef6\u4e0b\u7684\u8868\u73b0\u3002\u89e3\u91ca\u53ef\u4ee5\u52a0\u5f3a\u5b89\u5168\u548c\u76d1\u7ba1\u8bc4\u4f30\uff0c\u4f46 NUS \u660e\u786e\u5c06\u5176\u4f5c\u4e3a\u8865\u5145\u8bc1\u636e\uff0c\u800c\u975e\u73b0\u6709\u6d4b\u8bd5\u3001\u8ba4\u8bc1\u6216\u76d1\u7ba1\u6d41\u7a0b\u7684\u66ff\u4ee3\u54c1\u3002\u6587\u7ae0\u4ecb\u7ecd\u7684\u662f\u4e00\u5957\u62df\u8bae\u6846\u67b6\u548c\u6b63\u5728\u8fdb\u884c\u7684\u63a2\u7d22\uff0c\u5e76\u975e\u5df2\u90e8\u7f72\u7684\u76d1\u7ba1\u6807\u51c6\uff0c\u4e5f\u6ca1\u6709\u58f0\u79f0\u8be5\u65b9\u6cd5\u5df2\u7ecf\u9a8c\u8bc1\u4e86\u6bcf\u4e2a\u9ad8\u98ce\u9669 AI \u7cfb\u7edf\uff08[NUS CDE\uff0c2 September 2026](https:\/\/cde.nus.edu.sg\/news\/an-explainable-ai-framework-for-scientific-discovery-and-high-stakes-applications\/)\uff09\u3002"},"related":[{"id":90727,"source_id":"sgkb.science-research.nus-business-analytics-centre-ai-fund-2026","kind":"article","title_en":"NUS Business Analytics Centre AI Scholarship and Fund 2026","title_zh":"NUS \u5546\u4e1a\u5206\u6790\u4e2d\u5fc3\u4eba\u5de5\u667a\u80fd\u5956\u5b66\u91d1\u4e0e\u57fa\u91d1 2026","assurance":null,"subtype":null},{"id":90732,"source_id":"sgkb.science-research.nus-openai-ai-collaboration-2026","kind":"article","title_en":"NUS\u2013OpenAI Strategic AI Collaboration 2026","title_zh":"NUS\u2013OpenAI 2026 \u5e74\u6218\u7565\u4eba\u5de5\u667a\u80fd\u5408\u4f5c","assurance":null,"subtype":null},{"id":90738,"source_id":"sgkb.science-research.returning-singaporean-scientists-scheme-nus-2026","kind":"article","title_en":"Returning Singaporean Scientists Scheme: NUS Appointments 2026","title_zh":"\u65b0\u52a0\u5761\u79d1\u5b66\u5bb6\u56de\u6d41\u8ba1\u5212\uff1aNUS \u4efb\u547d\uff082026\uff09","assurance":null,"subtype":null},{"id":90713,"source_id":"sgkb.science-research.ai-for-science","kind":"article","title_en":"Singapore's AI-for-Science initiative","title_zh":"\u65b0\u52a0\u5761\u201c\u4eba\u5de5\u667a\u80fd\u4fc3\u8fdb\u79d1\u5b66\u201d\u5021\u8bae","assurance":null,"subtype":null},{"id":89740,"source_id":"sgkb.business-industry.national-ai-strategy-and-governance","kind":"article","title_en":"Singapore's National AI Strategy and Governance","title_zh":"\u65b0\u52a0\u5761\u56fd\u5bb6\u4eba\u5de5\u667a\u80fd\u6218\u7565\u4e0e\u6cbb\u7406","assurance":null,"subtype":null},{"id":90748,"source_id":"sgkb.science-research.university-research-ecosystem","kind":"article","title_en":"University Research Ecosystem","title_zh":"\u5927\u5b66\u7814\u7a76\u751f\u6001\u7cfb\u7edf","assurance":null,"subtype":null},{"id":83286,"source_id":"sg.university.nus","kind":"entity","title_en":"National University of Singapore","title_zh":"\u65b0\u52a0\u5761\u56fd\u7acb\u5927\u5b66","assurance":null,"subtype":"autonomous-university"},{"id":82552,"source_id":"sg.research.nus-xai-scientific-discovery-framework","kind":"entity","title_en":"NUS explainable-AI framework for scientific discovery and high-stakes applications","title_zh":"NUS \u7528\u4e8e\u79d1\u5b66\u53d1\u73b0\u548c\u9ad8\u98ce\u9669\u5e94\u7528\u7684\u53ef\u89e3\u91ca\u4eba\u5de5\u667a\u80fd\u6846\u67b6","assurance":null,"subtype":"research-methods-framework"},{"id":92917,"source_id":"sgtopic.science-research.nus-explainable-ai-scientific-discovery-framework-2026","kind":"topic","title_en":"NUS Explainable-AI Framework for Scientific Discovery 2026","title_zh":"NUS \u79d1\u5b66\u53d1\u73b0\u53ef\u89e3\u91ca\u4eba\u5de5\u667a\u80fd\u6846\u67b6 2026","assurance":null,"subtype":null}]}