arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2405.08237cs.CLcs.SDeess.AS
收藏

一种预测学习模型可模拟连续语音神经表征中发现的时间动态与语境效应

A predictive learning model can simulate temporal dynamics and context effects found in neural representations of continuous speech

University of Edinburgh(爱丁堡大学);University of Maryland(马里兰大学)

Oli Danyi Liu, Hao Tang, Naomi Feldman, Sharon Goldwater

AI总结:

本研究用基于无标注语音训练的预测学习计算模型模拟语音神经表征分析,发现其与大脑存在相似时间动态及音位跨语境泛化特性,同时指出仅靠该分析不足以证明语境不变编码的存在。

摘要:

语音感知涉及对序列呈现项目的存储与整合。认知神经科学领域的最新研究已发现人类语音神经编码中存在可能助力这种时间加工的时间与语境特征。本研究中,我们用从一个计算模型中提取的表征开展了类似分析的模拟,该模型以预测后续声学信号为学习目标,在无标注语音上训练而成。我们的模拟揭示出与脑信号中相似的时间动态,表明这些特性无需语言知识即可产生。大脑与该模型共有的另一特性是,音位的编码模式支持一定程度的跨语境泛化。但我们发现证据表明,这些泛化的有效性取决于具体语境,这意味着仅靠该分析不足以支持语境不变编码的存在。

英文摘要:

Speech perception involves storing and integrating sequentially presented items. Recent work in cognitive neuroscience has identified temporal and contextual characteristics in humans' neural encoding of speech that may facilitate this temporal processing. In this study, we simulated similar analyses with representations extracted from a computational model that was trained on unlabelled speech with the learning objective of predicting upcoming acoustics. Our simulations revealed temporal dynamics similar to those in brain signals, implying that these properties can arise without linguistic knowledge. Another property shared between brains and the model is that the encoding patterns of phonemes support some degree of cross-context generalization. However, we found evidence that the effectiveness of these generalizations depends on the specific contexts, which suggests that this analysis alone is insufficient to support the presence of context-invariant encoding.

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。

↑