Informal natural language that describes code functionality, such as code comments or function documentation, may contain substantial information about a programs intent. However, there is typically no guarantee that a programs implementation and natural language documentation are aligned. In the case of a conflict, leveraging information in code-adjacent natural language has the potential to enhance fault localization, debugging, and code trustworthiness. In practice, however, this information is often underutilized due to the inherent ambiguity of natural language which makes natural language intent challenging to check programmatically. The emergent abilities of Large Language Models (LLMs) have the potential to facilitate the translation of natural language intent to programmatically checkable assertions. However, it is unclear if LLMs can correctly translate informal natural language specifications into formal specifications that match programmer intent. Additionally, it is unclear if such translation could be useful in practice. In this paper, we describe nl2postcond, the problem of leveraging LLMs for transforming informal natural language to formal method postconditions, expressed as program assertions. We introduce and validate metrics to measure and compare different nl2postcond approaches, using the correctness and discriminative power of generated postconditions. We then use qualitative and quantitative methods to assess the quality of nl2postcond postconditions, finding that they are generally correct and able to discriminate incorrect code. Finally, we find that nl2postcond via LLMs has the potential to be helpful in practice; nl2postcond generated postconditions were able to catch 64 real-world historical bugs from Defects4J.<\/p>\n","protected":false},"excerpt":{"rendered":"
Informal natural language that describes code functionality, such as code comments or function documentation, may contain substantial information about a programs intent. However, there is typically no guarantee that a programs implementation and natural language documentation are aligned. In the case of a conflict, leveraging information in code-adjacent natural language has the potential to enhance […]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr-author-ordering":null,"msr_publishername":"ACM","msr_publisher_other":"","msr_booktitle":"","msr_chapter":"","msr_edition":"","msr_editors":"","msr_how_published":"","msr_isbn":"","msr_issue":"","msr_journal":"","msr_number":"","msr_organization":"","msr_pages_string":"","msr_page_range_start":"","msr_page_range_end":"","msr_series":"","msr_volume":"","msr_copyright":"","msr_conference_name":"The ACM International Conference on the Foundations of Software Engineering (FSE)","msr_doi":"","msr_arxiv_id":"","msr_s2_paper_id":"","msr_mag_id":"","msr_pubmed_id":"","msr_other_authors":"","msr_other_contributors":"","msr_speaker":"","msr_award":"","msr_affiliation":"","msr_institution":"","msr_host":"","msr_version":"","msr_duration":"","msr_original_fields_of_study":"","msr_release_tracker_id":"","msr_s2_match_type":"","msr_citation_count_updated":"","msr_published_date":"2024-7-16","msr_highlight_text":"","msr_notes":"GitHub page https:\/\/github.com\/microsoft\/nl-2-postcond\/","msr_longbiography":"","msr_publicationurl":"","msr_external_url":"","msr_secondary_video_url":"","msr_conference_url":"","msr_journal_url":"","msr_s2_pdf_url":"","msr_year":0,"msr_citation_count":0,"msr_influential_citations":0,"msr_reference_count":0,"msr_s2_match_confidence":0,"msr_microsoftintellectualproperty":true,"msr_s2_open_access":false,"msr_s2_author_ids":[],"msr_pub_ids":[],"msr_hide_image_in_river":0,"footnotes":""},"msr-research-highlight":[],"research-area":[13556,13560],"msr-publication-type":[193716],"msr-publisher":[],"msr-focus-area":[],"msr-locale":[268875],"msr-post-option":[],"msr-field-of-study":[246694,268116,251365],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-976788","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-artificial-intelligence","msr-research-area-programming-languages-software-engineering","msr-locale-en_us","msr-field-of-study-artificial-intelligence","msr-field-of-study-programming-language-88","msr-field-of-study-software-engineering"],"msr_publishername":"ACM","msr_edition":"","msr_affiliation":"","msr_published_date":"2024-7-16","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"","msr_chapter":"","msr_isbn":"","msr_journal":"","msr_volume":"","msr_number":"","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"","msr_how_published":"","msr_notes":"GitHub page https:\/\/github.com\/microsoft\/nl-2-postcond\/","msr_highlight_text":"","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":1,"msr_main_download":"","msr_publicationurl":"","msr_doi":"","msr_publication_uploader":[{"type":"doi","viewUrl":"false","id":"false","title":"https:\/\/dl.acm.org\/doi\/10.1145\/3660791","label_id":"243109","label":0},{"type":"url","viewUrl":"false","id":"false","title":"https:\/\/arxiv.org\/abs\/2310.01831","label_id":"252679","label":0}],"msr_related_uploader":[{"type":"url","viewUrl":"false","id":"false","title":"https:\/\/github.com\/microsoft\/nl-2-postcond\/","label_id":"264520","label":0},{"type":"file","viewUrl":"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2023\/10\/FSE-24-nl2postcond-final-slides.pdf","id":"1060227","title":"fse-24-nl2postcond-final-slides","label_id":"243115","label":0}],"msr_citation_count":0,"msr_citation_count_updated":"","msr_s2_paper_id":"","msr_influential_citations":0,"msr_reference_count":0,"msr_arxiv_id":"","msr_s2_author_ids":[],"msr_s2_open_access":false,"msr_s2_pdf_url":null,"msr_attachments":[{"id":1060227,"url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/07\/FSE-24-nl2postcond-final-slides.pdf"}],"msr-author-ordering":[{"type":"text","value":"Madeline Endres","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Sarah Fakhoury","user_id":42180,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Sarah Fakhoury"},{"type":"user_nicename","value":"Saikat Chakraborty","user_id":42411,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Saikat Chakraborty"},{"type":"user_nicename","value":"Shuvendu Lahiri","user_id":33640,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Shuvendu Lahiri"}],"msr_impact_theme":[],"msr_research_lab":[199565],"msr_event":[],"msr_group":[144812],"msr_project":[890049],"publication":[],"video":[],"msr-tool":[],"msr_publication_type":"inproceedings","related_content":{"projects":[{"ID":890049,"post_title":"Trusted AI-assisted Programming","post_name":"trusted-ai-assisted-programming","post_type":"msr-project","post_date":"2022-11-17 11:25:36","post_modified":"2024-10-23 15:47:00","post_status":"publish","permalink":"https:\/\/www.microsoft.com\/en-us\/research\/project\/trusted-ai-assisted-programming\/","post_excerpt":"Machine learning, in particular Large Language Models, has shown great promise at automating several aspects of programming and software development such as coding, testing, integration, static analysis, verification etc. in recent years. In this project, we leverage and extend large language models with ideas grounded in programming languages and correctness to develop trusted AI agents for all aspects of programming for reliable software development. Current work explores various directions: We study the problem of user-intent…","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/890049"}]}}]},"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/976788","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":3,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/976788\/revisions"}],"predecessor-version":[{"id":1026003,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/976788\/revisions\/1026003"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=976788"}],"wp:term":[{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=976788"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=976788"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=976788"},{"taxonomy":"msr-publisher","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publisher?post=976788"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=976788"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=976788"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=976788"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=976788"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=976788"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=976788"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=976788"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=976788"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}