arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2202.03555cs.LG
收藏

data2vec:语音、视觉和语言自监督学习的通用框架

data2vec: A General Framework for Self-supervised Learning in Speech, Vision and Language

Meta AI;SambaNova(SambaNova公司)

Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli

摘要:

虽然自监督学习的一般思想在不同的模式中是相同的,但实际的算法和目标却有很大的不同,因为它们是在考虑单一模式的情况下开发的。为了让我们更接近一般的自监督学习,我们提出了data2vec,这是一个对语音、NLP或计算机视觉使用相同学习方法的框架。其核心思想是使用标准Transformer架构,基于自蒸馏设置中输入的掩蔽视图来预测完整输入数据的潜在表示。data2vec不是预测特定于模态的目标,如单词、视觉标记或人类语音单元,它们本质上是本地的,而是预测包含整个输入信息的上下文化潜在表示。在语音识别、图像分类和自然语言理解的主要基准上的实验展示了与主流方法相比的新的艺术状态或竞争性能。

英文摘要:

While the general idea of self-supervised learning is identical across modalities, the actual algorithms and objectives differ widely because they were developed with a single modality in mind. To get us closer to general self-supervised learning, we present data2vec, a framework that uses the same learning method for either speech, NLP or computer vision. The core idea is to predict latent representations of the full input data based on a masked view of the input in a self-distillation setup using a standard Transformer architecture. Instead of predicting modality-specific targets such as words, visual tokens or units of human speech which are local in nature, data2vec predicts contextualized latent representations that contain information from the entire input. Experiments on the major benchmarks of speech recognition, image classification, and natural language understanding demonstrate a new state of the art or competitive performance to predominant approaches.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑