arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2108.08877cs.CL
收藏

Sentence-T5:基于预训练文本到文本模型的可扩展句子编码器

Sentence-T5: Scalable Sentence Encoders from Pre-trained Text-to-Text Models

Google Research(谷歌研究院)

Jianmo Ni, Gustavo Hernández Ábrego, Noah Constant, Ji Ma, Keith B. Hall, Daniel Cer, Yinfei Yang

摘要:

我们提供了文本到文本Transformers(T5)的句子嵌入的第一次探索。句子嵌入在语言处理任务中有着广泛的应用。虽然T5在序列到序列映射问题的语言任务上取得了令人印象深刻的性能,但目前还不清楚如何从编码器-解码器模型中生成句子嵌入。我们研究了三种提取T5句子嵌入的方法:两种只利用T5编码器,一种使用完整的T5编码器-解码器模型。为了支持我们的调查,我们建立了一个新的句子表示转移基准,SentGLUE,它扩展了SentEval工具包从GLUE基准的九个任务。我们的仅编码器模型在SentEval和SentGLUE传输任务上的性能优于Sentence-BERT和SimCSE句子嵌入,包括语义文本相似性(STS)。发现将T5从数百万个参数放大到数十亿个参数可以产生一致的进一步改进。最后,我们的编码器-解码器方法实现了一个新的国家的最先进的STS时,使用句子嵌入。我们的模型发布在https://tfhub.dev/google/collections/sentence-t5/1。

英文摘要:

We provide the first exploration of sentence embeddings from text-to-text transformers (T5). Sentence embeddings are broadly useful for language processing tasks. While T5 achieves impressive performance on language tasks cast as sequence-to-sequence mapping problems, it is unclear how to produce sentence embeddings from encoder-decoder models. We investigate three methods for extracting T5 sentence embeddings: two utilize only the T5 encoder and one uses the full T5 encoder-decoder model. To support our investigation, we establish a new sentence representation transfer benchmark, SentGLUE, which extends the SentEval toolkit to nine tasks from the GLUE benchmark. Our encoder-only models outperforms Sentence-BERT and SimCSE sentence embeddings on both SentEval and SentGLUE transfer tasks, including semantic textual similarity (STS). Scaling up T5 from millions to billions of parameters is found to produce consistent further improvements. Finally, our encoder-decoder method achieves a new state-of-the-art on STS when using sentence embeddings. Our models are released at https://tfhub.dev/google/collections/sentence-t5/1.

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。

↑