论文概要
研究领域: ML 作者: Ajmain Inqiad Alam, Palash Roy, Chanchal K. Roy, etc. 发布时间: 2026-04-29 arXiv: 2504.21142
中文摘要
大型语言模型（LLM）在软件工程（SE）中的加速采用带来了一场静默危机：不可持续的计算成本。虽然这些模型在不同SE任务中展现出卓越能力，但它们规模难以管理、部署缓慢、内存密集且碳排放高。这一现实不仅威胁AI驱动SE的可扩展性和可及性，也威胁其长期环境可持续性。研究挑战很明确：我们必须超越准确率，将效率和环境成本作为首要设计约束。为此，我们引入碳税Transformer（CTT），一个受经济碳税原则启发的系统性多架构压缩流水线排序方案。借鉴碳定价的经济学概念，CTT将计算碳税付诸实践，惩罚架构低效并奖励部署就绪的压缩。我们在三个核心SE任务上评估CTT：代码克隆检测、代码摘要和代码生成，模型涵盖仅编码器、编码器-解码器和仅解码器架构。结果显示CTT在推理方面表现优异：(1) 内存减少高达49倍，(2) 克隆检测时间减少8-10倍，摘要生成减少3倍，代码生成减少4-7倍，(3) CO2排放减少高达81%，(4) CTT在克隆检测上保持约98%准确率，摘要约89%，代码生成高达91%（文本指标）和68%（pass@1）。
原文摘要
The accelerating adoption of Large Language Models (LLMs) in software engineering (SE) has brought with it a silent crisis: unsustainable computational cost. While these models demonstrate remarkable capabilities in different SE tasks, they are unmanageably large, slow to deploy, memory-intensive, and carbon-heavy. This reality threatens not only the scalability and accessibility of AI-powered SE, but also its long-term environmental sustainability. The research challenge is clear: we must go beyond accuracy and address efficiency and environmental cost as first-class design constraints. To meet this challenge, we introduce Carbon-Taxed Transformers (CTT), a systematic multi-architectural compression principled pipeline ordering inspired by economic carbon taxation principles. Drawing from...
--- *自动采集于 2026-04-30*
#论文 #arXiv #ML #小凯