{"id":112,"date":"2025-11-12T21:12:07","date_gmt":"2025-11-12T20:12:07","guid":{"rendered":"https:\/\/project.inria.fr\/llm4code\/?page_id=112"},"modified":"2026-09-17T10:03:14","modified_gmt":"2026-09-17T08:03:14","slug":"publications","status":"publish","type":"page","link":"https:\/\/project.inria.fr\/llm4code\/publications\/","title":{"rendered":"Publications"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\"><strong>2026<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Baudart G., Mahboubi A.,  A Case Study in Mechanizing Computer-Aided Proofs: Small Gaps Between Primes, AITP <a href=\"https:\/\/github.com\/LLM4Rocq\/small-prime-gap\">https:\/\/github.com\/LLM4Rocq\/small-prime-gap<\/a><\/li>\n\n\n\n<li>Colas, C., Perez, J., Nisioti, E., Mocherla, A., Oudeyer, P. Y., Moulin-Frier, C., &amp; Derex, M. (2026). Discovering Adaptive Transmission Programs for Collective Innovation. <a href=\"https:\/\/arxiv.org\/pdf\/2608.24545\">https:\/\/arxiv.org\/pdf\/2608.24545<\/a><\/li>\n\n\n\n<li>Baudart G. , Lelarge M., Library Before Proof: Making LLM-Generated Rocq Usable by Mathematicians, <em>AI4Research@ICML<\/em><\/li>\n\n\n\n<li>Mandel L., Baudart G., Vaziri M., Hirzel M. PPDL: LLM-Based Flows as Probabilistic Programs,&nbsp;<em>ICML<\/em><\/li>\n\n\n\n<li>Baudart G., Viennot J., St\u00e9rin T., Lelarge M., Putnam 2025 Problems in Rocq using Opus 4.6 and Rocq-MCP, Experiment report <a href=\"https:\/\/arxiv.org\/abs\/2603.20405\">https:\/\/arxiv.org\/abs\/2603.20405<\/a><\/li>\n\n\n\n<li>Viennot, J., Baudart, G., Gallego Arias, E.J., Lelarge, Stoskopf T., Tacq: Context Aware Tactic Recommendation for Rocq, JFLA <a href=\"https:\/\/hal.science\/hal-05428141v1\">https:\/\/hal.science\/hal-05428141<\/a><\/li>\n\n\n\n<li>Kier, K., Giagnorio, A., AbedelKader, O., Zaitsev, O., Peharz, R., Robbes, R., Bavota, G., Ducasse, S. Teaching LLMs a Low-Resource Language: Enhancing Code Completion in Pharo. ICSME 2026. https:\/\/arxiv.org\/abs\/2607.04939 <\/li>\n\n\n\n<li>Allain, A., Blot, A., Khelladi, D.E., Acher, M. Code Transformation Rule Synthesis using LLMs: Potential and Limits. ASE 2026. <a href=\"https:\/\/hal.science\/hal-05607247\/\">https:\/\/arxiv.org\/abs\/2609.03592<\/a><\/li>\n\n\n\n<li>Acher, M., J\u00e9z\u00e9quel, J.-M. Do Programming Languages Still Matter to Your AI Coding Agent Teammate? Evidence at Scale from Chess Engines. arXiv preprint arXiv:2606.13763, 2026. https:\/\/arxiv.org\/abs\/2606.13763  <\/li>\n\n\n\n<li>Acher, M., Kebaili, Z., Espinasse, J.-B., Khelladi, D.E. Coding Agents Can Develop COBOL Systems: Evidence from a Human-Guided Multi-Case Study. HAL preprint hal-05743207, 2026. https:\/\/hal.science\/hal-05743207 <\/li>\n\n\n\n<li>Espinasse, J.-B., Khelladi, D.E., Acher, M. Spec2COBOLRot: An Agentic-AI Degradation Loop for Realistic COBOL Corpus Generation. AISM@ASE 2026 (Workshop on AI for Software Modernization).<\/li>\n\n\n\n<li>Reux, C., Acher, M., Khelladi, D.E., Quinton, C., Barais, O. Imperfect Visual Verification for Code Edition: A Case Study on TikZ. arXiv preprint arXiv:2606.15693, 2026. https:\/\/arxiv.org\/abs\/2606.15693 <\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>2025<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Boulet, T., Hinaut, X., Moulin-Frier, C. Software Engineering Agents for Embodied Controller Generation: A Study in Minigrid Environments. EfficientReasoning@NeurIPS 2025. <a href=\"https:\/\/inria.hal.science\/hal-05330526\">https:\/\/inria.hal.science\/hal-05330526<\/a><\/li>\n\n\n\n<li>Coignion, T., Quinton, C., Rouvoy, R. When Faster Isn&#8217;t Greener: The Hidden Costs of LLM-Based Code Optimization. ASE 2025. <a href=\"https:\/\/hal.archives-ouvertes.fr\/hal-05227453\">https:\/\/hal.archives-ouvertes.fr\/hal-05227453<\/a><\/li>\n\n\n\n<li>D\u00f6derlein, J.B., Kouadio, N.H., Acher, M., Khelladi, D.E., Combemale, B. Piloting Copilot, Codex, and StarCoder2: Hot temperature, cold prompts, or black magic? Journal of Systems and Software 2025. <a href=\"https:\/\/arxiv.org\/abs\/2210.14699\">https:\/\/arxiv.org\/abs\/2210.14699<\/a><\/li>\n\n\n\n<li>Pourcel, J., Colas, C., Oudeyer, P.Y. Self-Improving Language Models for Evolutionary Program Synthesis: A Case Study on ARC-AGI. ICML 2025. <a href=\"https:\/\/arxiv.org\/abs\/2507.14172\">https:\/\/arxiv.org\/abs\/2507.14172<\/a><\/li>\n\n\n\n<li>Reux, C., Acher, M., Khelladi, D.E., Barais, O., Quinton, C. LLM Code Customization with Visual Results: A Benchmark on TikZ. EASE 2025. <a href=\"https:\/\/hal.archives-ouvertes.fr\/hal-05049250\">https:\/\/hal.archives-ouvertes.fr\/hal-05049250<\/a><\/li>\n\n\n\n<li>Spieker, H., Matricon, T., Belmecheri, N., Betten, J.E., Le Bartz Lyan, G., Borges, H., Mazouni, Q., Gross, D., Gotlieb, A., Acher, M. Prompting for Performance: Exploring LLMs for Configuring Software. ICTAI 2025. <a href=\"https:\/\/arxiv.org\/abs\/2507.09790\">https:\/\/arxiv.org\/abs\/2507.09790<\/a><\/li>\n\n\n\n<li>Stoskopf, T., Cohen, C., Tabareau, N., Babel-formal: Translation of Proofs between Lean and Rocq. MathAI@NeurIPS 2025. <a href=\"https:\/\/hal.science\/hal-05342510\/\">https:\/\/hal.science\/hal-05342510\/<\/a><\/li>\n\n\n\n<li>Stoskopf, T., Viennot, J., Cohen, C., LLM4Docq: Bootstrapping Documentation for MathComp with LLMs and Expert Feedback, Rocqshop@ITP 2025. <a href=\"https:\/\/coq-workshop.gitlab.io\/2025\/files\/EA9.pdf\">https:\/\/coq-workshop.gitlab.io\/2025\/files\/EA9.pdf<\/a><\/li>\n\n\n\n<li>Viennot, J., Baudart, G., Gallego Arias, E.J., Lelarge, M. MiniF2F in Rocq: Automatic Translation Between Proof Assistants\u2013A Case Study. MathAI@NeurIPS 2025. <a href=\"https:\/\/arxiv.org\/abs\/2503.04763\">https:\/\/arxiv.org\/abs\/2503.04763<\/a><\/li>\n\n\n\n<li>Zine, N., Quinton, C., Rouvoy, R. LLM-based Co-Evolution of Configurable Software Systems. SPLC 2025. <a href=\"https:\/\/hal.archives-ouvertes.fr\/hal-05090995\">https:\/\/hal.archives-ouvertes.fr\/hal-05090995<\/a><\/li>\n\n\n\n<li>Acher, M. Une meilleure IA pour d\u00e9velopper du code, Inria Communication 2025, <a href=\"https:\/\/www.inria.fr\/fr\/meilleure-ia-developper-code-diverse\">https:\/\/www.inria.fr\/fr\/meilleure-ia-developper-code-diverse<\/a><\/li>\n\n\n\n<li>Baudart, G. Faire communiquer un mod\u00e8le de langage et un assistant de preuve, Blog Binaire, 2025 <a href=\"https:\/\/www.lemonde.fr\/blog\/binaire\/2025\/05\/23\/faire-communiquer-un-modele-de-langage-et-un-assistant-de-preuve\/\">https:\/\/www.lemonde.fr\/blog\/binaire\/2025\/05\/23\/faire-communiquer-un-modele-de-langage-et-un-assistant-de-preuve\/<\/a><\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>2024<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Pourcel, G., Carta, T., Kova\u010d, G., Oudeyer, P.Y. Autotelic LLM-based exploration for goal-conditioned RL. IMOL@NeurIPS 2024. <a href=\"https:\/\/inria.hal.science\/hal-04861896\">https:\/\/inria.hal.science\/hal-04861896<\/a><\/li>\n\n\n\n<li>Pourcel, J., Colas, C., Molinaro, G., Oudeyer, P.Y., Teodorescu, L. Aces: generating diverse programming puzzles with autotelic language models and semantic descriptors. NeurIPS 2024. <a href=\"https:\/\/arxiv.org\/abs\/2310.10692\">https:\/\/arxiv.org\/abs\/2310.10692<\/a><\/li>\n\n\n\n<li>Teodorescu, L., Baudart, G., Gallego Arias, E.J., Lelarge, M. NLIR: Natural Language Intermediate Representation for Mechanized Theorem Proving. MathAI@NeurIPS 2024. <a href=\"https:\/\/hal.science\/hal-04886208\">https:\/\/hal.science\/hal-04886208<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>2026 2025 2024<\/p>\n<p> <a class=\"continue-reading-link\" href=\"https:\/\/project.inria.fr\/llm4code\/publications\/\"><span>Continue reading<\/span><i class=\"crycon-right-dir\"><\/i><\/a> <\/p>\n","protected":false},"author":2510,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":"","_members_access_role":[],"_members_access_error":""},"class_list":["post-112","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/project.inria.fr\/llm4code\/wp-json\/wp\/v2\/pages\/112","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/project.inria.fr\/llm4code\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/project.inria.fr\/llm4code\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/project.inria.fr\/llm4code\/wp-json\/wp\/v2\/users\/2510"}],"replies":[{"embeddable":true,"href":"https:\/\/project.inria.fr\/llm4code\/wp-json\/wp\/v2\/comments?post=112"}],"version-history":[{"count":13,"href":"https:\/\/project.inria.fr\/llm4code\/wp-json\/wp\/v2\/pages\/112\/revisions"}],"predecessor-version":[{"id":160,"href":"https:\/\/project.inria.fr\/llm4code\/wp-json\/wp\/v2\/pages\/112\/revisions\/160"}],"wp:attachment":[{"href":"https:\/\/project.inria.fr\/llm4code\/wp-json\/wp\/v2\/media?parent=112"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}