2026/04/06 by Samira Hajizadeh, Suman Jana · 1 voice
Computer Science · Materials Science · #Code (set theory) #Execution time #Machine Learning in Materials Science #Pairwise comparison #Profiling (computer programming) #Security token #Software Engineering Research #Speedup #Task (project management) #Topic Modeling #cs.AI #cs.CL #cs.LG #cs.PL #cs.SE
paper · pdf · doi:10.48550/arxiv.2604.05137
openalex publication_date 2026/04/06 · arxiv published 2026/04/06 · arxiv updated 2026/04/06 · openalex created_date 2026/04/09 · openalex updated_date 2026/07/28
Large language models (LLMs) often generate code that is functionally correct but inefficient in runtime and memory. Prior approaches to improving code efficiency typically rely on absolute execution feedback, such as profiling a single program's runtime or memory usage, which is costly and provides weak guidance for refinement. We propose Relative Contrastive Feedback (RCF), an inference-time feedback mechanism that requires no model fine-tuning or parameter updates. RCF compares two structurally similar programs for the same task and highlights the differences associated with better efficiency. Building on this idea, we introduce EffiPair, an inference-time iterative refinement framework that operates entirely at test time by generating multiple candidate solutions, identifying informative program pairs with large efficiency gaps, summarizing their execution differences into lightweight feedback, and using this signal to produce more efficient solutions. By replacing isolated scalar feedback with pairwise contrastive comparisons, EffiPair provides more direct guidance while reducing profiling and prompting overhead. Experiments on code-efficiency benchmarks show that EffiPair consistently improves efficiency while preserving correctness. For instance, with DeepSeek-Chat V3.2, EffiPair achieves up to 1.5x speedup over generation without performance feedback, while reducing token usage by more than 90% compared to prior work.