Software supply chain and LLM for repair at Udem next week

[Benoit Baudry]

Hi,

Following ICSE, we host two guests next week. 

Aman Sharma will present his work on "Unreproducible Builds in Java: Causes and Mitigations", on Monday at 10h. 

André Silva will present his work on "RepairBench: Leaderboard of Frontier Models for Program Repair", on Wednesday at 11h.

Details are below

Cheers,

Benoit

Monday May 5, 10am, room 3195, André Aisenstadt building

Title : Unreproducible Builds in Java: Causes and Mitigations

Abstract : Reproducible Builds is critical for the transparency and integrity of software supply chains, yet Java builds suffer from unreproducibility due to non-deterministic metadata and bytecode inconsistencies. In this talk, we investigate the causes of unreproducibility by analyzing Maven artifacts on Reproducible Central [1], identifying key patterns that lead to inconsistent builds. We leverage existing tools such as jNorm [2], which normalizes Java bytecode by stripping non-deterministic information from class files, and oss-rebuild, a tool developed by Google to fix metadata-related unreproducibility in JAR files in order to understand to what extent these issues can be fixed.

Speaker Bio :Aman is a PhD student at KTH Royal Institute of Technology, Stockholm, Sweden and a researcher in CHAINS project funded by the Swedish Foundation for Strategic Research. He works on securing software supply chains of Java. Before that, he received his Bachelor in Technology from Indian Institute of Technology in Roorkee, India.

Wednesday May 7, 11am, room 3195, André Aisenstadt building

Title: RepairBench: Leaderboard of Frontier Models for Program Repair

Abstract: AI-driven program repair uses AI models to repair buggy software by producing patches. Rapid advancements in AI surely impact state-of-the-art performance of program repair. Yet, grasping this progress requires frequent and standardized evaluations. We propose RepairBench, a novel leaderboard for AI-driven program repair. The key characteristics of RepairBench are: 1) it is execution-based: all patches are compiled and executed against a test suite, 2) it assesses frontier models in a frequent and standardized way. RepairBench leverages two high-quality benchmarks, Defects4J and GitBug-Java, to evaluate frontier models against real-world program repair tasks. We publicly release the evaluation framework of RepairBench. We will update the leaderboard as new frontier models are released.

Bio: André Silva is a Ph.D. student at KTH Royal Institute of Technology in Stockholm,  Sweden. His research interests include the intersection of automatic program repair and machine learning. His previous work includes fine-tuning of large language models and building benchmarks for program repair. André received his M.Sc. in Computer Science from Instituto Superior Técnico, Universidade de Lisboa, Lisbon, Portugal.