MindForge: Teaching Small Language Models Whole-Life-Cycle Software Engineering via Source-Free Program Synthesis
Coding agents have made substantial progress on software engineering tasks that modify existing codebases, including bug fixing and feature implementation. However, constructing a complete program from scratch remains a major challenge: even the frontier models evaluated on ProgramBench fully resolve fewer than 1% of tasks. One obstacle is the lack of scalable training environments for this from-scratch setting, spanning the whole software engineering life cycle, as existing environment-construction frameworks focus only on a single phase in software development. To address this gap, we introduce MindForge, an automated pipeline that converts open-source command-line programs into source-free environments that expose only a compiled reference executable and its documentation. Using MindForge, we construct training environments from repositories disjoint from those in ProgramBench, and curate a high-quality data recipe consisting of program synthesis trajectories using GLM-5.2 as the teacher agent. Fine-tuning Qwen3.6-27B on these trajectories increases its ProgramBench average test pass rate from 37.98% to 49.51%, achieving performance comparable to substantially larger frontier models. Moreover, the fine-tuned model consistently improves over the base model across all seven unseen software engineering benchmarks, spanning long-horizon repository generation and translation, bug fixing, feature implementation, and cross-language issue resolution, with absolute gains of 31.00 points on RepoZero-C2Rust, 14.16 on DeepSWE, 10.70/4.56 on NL2Repo-Bench (with/without tests), 5.04 on SWE-bench Verified, 5.93 on SWE-bench Pro, 5.22 on SWE-bench Multilingual, and 4.94 on FeatBench.
Code (0)
등록된 구현이 없습니다.
Tasks
Program SynthesisSimilar Papers 제목 키워드 기반
MindForge: Empowering Embodied Agents with Theory of Mind for Lifelong Collaborative Learning
Contemporary embodied agents, such as Voyager in Minecraft, have demonstrated promising capabilities in open-ended individual learning. However, when powered with open large language models (LLMs), these agents often str…
MinecraftTransfer LearningA Whole-Person Function Dictionary for the Mobility, Self-Care and Domestic Life Domains: a Seedset Expansion Approach
Whole-person functional limitations in the areas of mobility, self-care and domestic life affect a majority of individuals with disabilities. Detecting, recording and monitoring such limitations would benefit those indiv…
Mitigating belief projection in explainable artificial intelligence via Bayesian Teaching
State-of-the-art deep-learning systems use decision rules that are challenging for humans to model. Explainable AI (XAI) attempts to improve human understanding but rarely accounts for how people typically reason about u…
Explainable artificial intelligenceExplainable Artificial Intelligence (XAI)image-classificationImage ClassificationMeta-Teaching: Curriculum Generation for Lifelong Learning
Meta-learning will be crucial to creating lifelong, generalizable AI. In practice, however, it is hard to define the meta-training task distribution that is used to train meta-learners. If made too small, tasks are too s…
Lifelong learningMeta-LearningSpatialAfford: Teaching Compact VLMs Where to Look and Where to Ground for Affordance
Affordance grounding aims to localize the functional region for interaction, such as the handle to grasp or the button to press, rather than the whole object. This makes it more challenging than generic visual grounding …
Reinforcement LearningSpatial ReasoningVisual Grounding