摘要
由一名新闻记者兼机器人与机器学习的工作人员新闻编辑每日新闻-一项关于机器学习的新研究-计算机智能现在可用。根据NewsRx记者从中华人民共和国济纳N发回的新闻报道，研究表明：“数据增强算法，如反向翻译，已经证明在各种深度学习任务中是有效的。尽管它们取得了显著的成功，但由于COD E由具有唯一性和确定性的离散标记组成，因此将数据增强算法应用于代码相关任务仍然存在障碍。”本研究经费来源于国家自然科学基金(NSFC)。我们的新闻记者从山东女子大学的研究中获得了一句话：“在这项工作中，我们提出了一种新颖而简单的数据增强方法FeaMix，它是一种基于自一致性学习的特征混合记忆批处理方法。FeaMix有两个独特之处：首先，其次，扩展了自一致性学习技术，优化了代码相关任务的语言模型，并通过大量实验验证了该方法在代码生成和代码翻译方面的有效性。
Abstract
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News – A new study on Machine Learning - Comp utational Intelligence is now available. According to news originating from Jina n, People’s Republic of China, by NewsRx correspondents, research stated, “Data augmentation algorithms, such as back translation, have shown to be effective in various deeplearning tasks. Despite their remarkable success, there has been a hurdle to applying data augmentation algorithms to code-related tasks since cod e consists of discrete tokens with uniqueness and certainty.” Financial support for this research came from National Natural Science Foundatio n of China (NSFC). Our news journalists obtained a quote from the research from Shandong Women’s Un iversity, “In this work, we propose FeaMix, a novel yet simple data augmentation approach designed for the feature mix with memory batch based on self-consisten cy learning. FeaMix has a couple of uniqueness. First, it specially selects the samples to be mixed by memory batch to guarantee that the generated features are in the same spatial distribution as the mixed features. Second, it extends the self-consistency learning technique to optimize the language model for code-rela ted tasks. With extensive experiments, we empirically validate that our method o utperforms several baseline models and traditional data augmentation methods on code generation and code translation.”
NETL
NSTL