SWE-Future: Forecast-Conditioned Data Synthesis for Future-Oriented Software Engineering Agents
The paper proposes SWE-Future, a method for forecasting future coding tasks in software repositories and uses the forecasts to synthesize a dataset for coding-agents.
The paper introduces a novel approach to forecasting future coding tasks and using the forecasts to synthesize a dataset for coding-agents, which is different from prior work that relies on direct historical replay or synthetic-data generation.
Before reading this…
Applications
- →Creating realistic benchmarks for coding-agents
- →Guiding future-oriented coding-task synthesis
To understand this paper, make sure you know these concepts first:
- Software engineeringfind papers →
- Data synthesisfind papers →
- Forecastingfind papers →
Abstract
More Like ThisRealistic coding-agent benchmarks often replay public GitHub issues and pull requests, making them vulnerable to overlap with model pretraining, fine-tuning, synthetic-data generation, or benchmark-driven model selection. Fully synthetic tasks avoid direct historical replay, but can drift away from real repository needs. We propose SWE-Future, a forecast-conditioned data synthesis method for future-oriented coding tasks. Given a forecast snapshot at time $T_0$, the method uses only pre-$T_0$ repository evidence to forecast future feature implementation/enhancement, bugfix, and refactor task families. We first validate this forecasting step retrospectively: after forecasts are fixed, later pull requests are used only to measure whether the predicted task families match future repository work. In an 80-repository study, the forecaster achieves 58.1\% future-work relevance under the main semantic matching metric. We then use validated forecast families as conditioning signals to synthesize a 200-task coding-agent dataset across 61 repositories from a task-generation snapshot, rather than replaying the later pull requests used for validation. SWE-Future shows that repository-evolution forecasts can guide realistic, future-oriented coding-task synthesis while reducing direct dependence on historical pull-request replay.