Computational approaches to human spoken word recognition

人类口语单词识别的计算方法

基本信息

  • 批准号:
    1754284
  • 负责人:
  • 金额:
    $ 60.23万
  • 依托单位:
  • 依托单位国家:
    美国
  • 项目类别:
    Continuing Grant
  • 财政年份:
    2018
  • 资助国家:
    美国
  • 起止时间:
    2018-03-15 至 2023-02-28
  • 项目状态:
    已结题

项目摘要

This project addresses one of the grand challenges facing cognitive science -- how humans understand speech. People recognize words far more easily than even the best computer speech recognition systems, even though the actual sounds we hear as consonants and vowels vary greatly depending on context (what sounds come before or after), who is talking, and the setting (a quiet room versus a crowded airport). Most current models of speech recognition cannot handle the huge variability in real speech because they do not operate on the actual speech signal. Also, they do not learn, so they cannot model how people acquire language. This project addresses these challenges by comparing current models of speech recognition to each other and to human capabilities, with the goal of understanding how human speech processing is so robust and flexible. In addition, simplified "deep learning" networks will be developed and evaluated as models of human speech recognition. Deep learning networks are similar to cognitive models in that they learn abstract representations of the data, not task-specific rules or algorithms. These networks have been used to create accurate commercial speech recognition systems. By comparing them to human performance, the investigators may provide new insights into why human speech recognition is so robust. The results of this project will have technical implications (better understanding of human flexibility may aid in improving computer speech recognition) and health implications (better understanding of human speech recognition will aid in developing better interventions for language disorders). The project will also support the training of a postdoctoral researcher and a PhD student, both of whom will develop skills that can be used to contribute to research and development in academia or industry. This project focuses on the development of a "shallow deep network" model called "DeepListener" that will be compared with the behavior of human listeners. A close match in the millisecond-level behavior of the network (for example, in which words are temporarily confusable with each other) and human performance suggests that human speech processing may emerge from similar principles as those in the model. In preliminary work, DeepListener learned to recognize 93% of 2000 real words (200 words produced by 10 talkers). DeepListener will be evaluated by detailed comparison to standard neural network models of cognitive theories and to human performance. The ways in which DeepListener is similar and dissimilar to human performance and competing models will help to advance scientific theories of human speech recognition. This project will follow emerging standards for open science: experiments will be pre-registered and data and computer code will be made freely and publicly available.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
该项目解决了认知科学面临的重大挑战之一-人类如何理解语音。人们比最好的计算机语音识别系统更容易识别单词,即使我们听到的辅音和元音的实际声音根据上下文(之前或之后的声音),谁在说话以及环境(安静的房间与拥挤的机场)而变化很大。当前的大多数语音识别模型不能处理真实的语音中的巨大变化,因为它们不对实际的语音信号进行操作。此外,他们不学习,所以他们不能模拟人们如何获得语言。该项目通过将当前的语音识别模型相互比较以及与人类能力进行比较来解决这些挑战,目的是了解人类语音处理如何如此强大和灵活。 此外,简化的“深度学习”网络将作为人类语音识别的模型进行开发和评估。深度学习网络类似于认知模型,因为它们学习数据的抽象表示,而不是特定于任务的规则或算法。这些网络已被用于创建准确的商业语音识别系统。通过将它们与人类的表现进行比较,研究人员可能会对人类语音识别为什么如此强大提供新的见解。该项目的结果将具有技术意义(更好地了解人类的灵活性可能有助于改善计算机语音识别)和健康意义(更好地了解人类语音识别将有助于开发更好的语言障碍干预措施)。该项目还将支持培训一名博士后研究员和一名博士生,他们都将发展可用于促进学术界或工业界研究和开发的技能。该项目的重点是开发一种名为“Deepcourse”的“浅层深度网络”模型,该模型将与人类听众的行为进行比较。网络毫秒级行为(例如,单词暂时彼此混淆)和人类表现的密切匹配表明,人类语音处理可能源于与模型中相似的原理。在初步工作中,DeepMind学会了识别2000个真实的单词(10个说话者产生的200个单词)中的93%。深度学习将通过与认知理论的标准神经网络模型和人类表现的详细比较来评估。DeepMind与人类表现和竞争模型的相似和不同之处将有助于推进人类语音识别的科学理论。该项目将遵循开放科学的新兴标准:实验将预先注册,数据和计算机代码将免费公开。该奖项反映了NSF的法定使命,并通过使用基金会的知识价值和更广泛的影响审查标准进行评估,被认为值得支持。

项目成果

期刊论文数量(18)
专著数量(0)
科研奖励数量(0)
会议论文数量(0)
专利数量(0)
Does predictive processing imply predictive coding in models of spoken word recognition?
预测处理是否意味着口语单词识别模型中的预测编码?
LexFindR: A fast, simple, and extensible R package for finding similar words in a lexicon
LexFindR:一个快速、简单且可扩展的 R 包,用于在词典中查找相似单词
  • DOI:
    10.3758/s13428-021-01667-6
  • 发表时间:
    2021
  • 期刊:
  • 影响因子:
    5.4
  • 作者:
    Li, ZhaoBin;Crinnion, Anne Marie;Magnuson, James S.
  • 通讯作者:
    Magnuson, James S.
EARSHOT: A minimal network model of human speech recognition that operates on real speech
  • DOI:
  • 发表时间:
    2019
  • 期刊:
  • 影响因子:
    0
  • 作者:
    J. Magnuson;Heejo You;J. Rueckl;Paul D. Allopenna;Monica Li;Sahil Luthra;Rachael Steiner;Hosung Nam;M. Escabí;K. Brown;Rachel M. Theodore;Nicholas Monto
  • 通讯作者:
    J. Magnuson;Heejo You;J. Rueckl;Paul D. Allopenna;Monica Li;Sahil Luthra;Rachael Steiner;Hosung Nam;M. Escabí;K. Brown;Rachel M. Theodore;Nicholas Monto
Word length, proportion of overlap, and phonological competition in spoken word recognition
  • DOI:
  • 发表时间:
    2018
  • 期刊:
  • 影响因子:
    2.5
  • 作者:
    Elizabeth Simmons;J. Magnuson
  • 通讯作者:
    Elizabeth Simmons;J. Magnuson
Does signal reduction imply predictive coding in models of spoken word recognition?
  • DOI:
    10.3758/s13423-021-01924-x
  • 发表时间:
    2021-08
  • 期刊:
  • 影响因子:
    3.5
  • 作者:
    Luthra S;Li MYC;You H;Brodbeck C;Magnuson JS
  • 通讯作者:
    Magnuson JS
{{ item.title }}
{{ item.translation_title }}
  • DOI:
    {{ item.doi }}
  • 发表时间:
    {{ item.publish_year }}
  • 期刊:
  • 影响因子:
    {{ item.factor }}
  • 作者:
    {{ item.authors }}
  • 通讯作者:
    {{ item.author }}

数据更新时间:{{ journalArticles.updateTime }}

{{ item.title }}
  • 作者:
    {{ item.author }}

数据更新时间:{{ monograph.updateTime }}

{{ item.title }}
  • 作者:
    {{ item.author }}

数据更新时间:{{ sciAawards.updateTime }}

{{ item.title }}
  • 作者:
    {{ item.author }}

数据更新时间:{{ conferencePapers.updateTime }}

{{ item.title }}
  • 作者:
    {{ item.author }}

数据更新时间:{{ patent.updateTime }}

James Magnuson其他文献

James Magnuson的其他文献

{{ item.title }}
{{ item.translation_title }}
  • DOI:
    {{ item.doi }}
  • 发表时间:
    {{ item.publish_year }}
  • 期刊:
  • 影响因子:
    {{ item.factor }}
  • 作者:
    {{ item.authors }}
  • 通讯作者:
    {{ item.author }}

{{ truncateString('James Magnuson', 18)}}的其他基金

CRCNS US-Spain Research Proposal: Collaborative Research: Tracking and modeling the neurobiology of multilingual speech recognition
CRCNS 美国-西班牙研究提案:合作研究:跟踪和建模多语言语音识别的神经生物学
  • 批准号:
    2207770
  • 财政年份:
    2022
  • 资助金额:
    $ 60.23万
  • 项目类别:
    Continuing Grant
Collaborative Research: CompCog: Psychological, Computational, and Neural Adequacy in a Deep Learning Model of Human Speech Recognition
合作研究:CompCog:人类语音识别深度学习模型中的心理、计算和神经充分性
  • 批准号:
    2043903
  • 财政年份:
    2021
  • 资助金额:
    $ 60.23万
  • 项目类别:
    Standard Grant
NRT-UtB: Science of learning, from neurobiology to real-world application: a problem-based approach
NRT-UtB:学习科学,从神经生物学到现实世界应用:基于问题的方法
  • 批准号:
    1735225
  • 财政年份:
    2017
  • 资助金额:
    $ 60.23万
  • 项目类别:
    Standard Grant
Real-world language: Future directions in the science of communication and the communication of science
现实世界语言:传播科学和科学传播的未来方向
  • 批准号:
    1747486
  • 财政年份:
    2017
  • 资助金额:
    $ 60.23万
  • 项目类别:
    Standard Grant
IGERT: Language plasticity - Genes, Brain, Cognition and Computation
IGERT:语言可塑性 - 基因、大脑、认知和计算
  • 批准号:
    1144399
  • 财政年份:
    2012
  • 资助金额:
    $ 60.23万
  • 项目类别:
    Continuing Grant
CAREER: The Time Course of Bottom-up and Top-down Integration in Language Understanding
职业:语言理解中自下而上和自上而下整合的时间进程
  • 批准号:
    0748684
  • 财政年份:
    2008
  • 资助金额:
    $ 60.23万
  • 项目类别:
    Continuing Grant
Compensation for Coarticulation: Implications for the Basis and Architecture of Speech Perception
协同发音的补偿:对语音感知的基础和架构的影响
  • 批准号:
    0642300
  • 财政年份:
    2007
  • 资助金额:
    $ 60.23万
  • 项目类别:
    Standard Grant
Special Foreign Currency Travel Support (In Indian Currency)To Participate in the Int'l Symposium on Lectins As Tools InBiology and Medicine; Calcutta, India; January 1981
特别外币旅行支持(印度货币)参加凝集素作为生物学和医学工具的国际研讨会;
  • 批准号:
    8022021
  • 财政年份:
    1981
  • 资助金额:
    $ 60.23万
  • 项目类别:
    Standard Grant

相似国自然基金

Lagrangian origin of geometric approaches to scattering amplitudes
  • 批准号:
    24ZR1450600
  • 批准年份:
    2024
  • 资助金额:
    0.0 万元
  • 项目类别:
    省市级项目

相似海外基金

Integration of Immunologic Phenotyping with Computational Approaches to Predict Clinical Trajectory in Septic Patients
免疫表型分析与计算方法相结合来预测脓毒症患者的临床轨迹
  • 批准号:
    10708534
  • 财政年份:
    2023
  • 资助金额:
    $ 60.23万
  • 项目类别:
Bottom-up and top-down computational modeling approaches to study CMV retinitis
研究 CMV 视网膜炎的自下而上和自上而下的计算模型方法
  • 批准号:
    10748709
  • 财政年份:
    2023
  • 资助金额:
    $ 60.23万
  • 项目类别:
Integrative computational-experimental approaches to stratify monogenic disease risk
综合计算实验方法对单基因疾病风险进行分层
  • 批准号:
    10889297
  • 财政年份:
    2023
  • 资助金额:
    $ 60.23万
  • 项目类别:
Eye tracking and computational approaches to understand the roles of maturation and experience in infant looking
眼动追踪和计算方法可用于了解成熟和经验在婴儿注视中的作用
  • 批准号:
    10602809
  • 财政年份:
    2023
  • 资助金额:
    $ 60.23万
  • 项目类别:
Computational approaches to the mechanistic elucidation of the serrated pathway of human colon carcinogenesis
人类结肠癌发生锯齿状途径机制阐明的计算方法
  • 批准号:
    10590985
  • 财政年份:
    2023
  • 资助金额:
    $ 60.23万
  • 项目类别:
Computational approaches to characterize heterogeneity and improve risk stratification in complex disease phenotypes
表征复杂疾病表型异质性并改善风险分层的计算方法
  • 批准号:
    10805689
  • 财政年份:
    2023
  • 资助金额:
    $ 60.23万
  • 项目类别:
Discovering human divergent activity-regulated elements using comparative, computational, and functional approaches
使用比较、计算和功能方法发现人类不同活动调节的元素
  • 批准号:
    10779701
  • 财政年份:
    2023
  • 资助金额:
    $ 60.23万
  • 项目类别:
Computational approaches to characterize heterogeneity and improve risk stratification in complex disease phenotypes
表征复杂疾病表型异质性并改善风险分层的计算方法
  • 批准号:
    10448966
  • 财政年份:
    2022
  • 资助金额:
    $ 60.23万
  • 项目类别:
Novel computational approaches to characterize the effects of rare functional outlier variants on cis- and trans-regulatory disease processes
新的计算方法来表征罕见功能异常变异对顺式和反式调节疾病过程的影响
  • 批准号:
    10433216
  • 财政年份:
    2022
  • 资助金额:
    $ 60.23万
  • 项目类别:
Novel computational approaches to characterize the effects of rare functional outlier variants on cis- and trans-regulatory disease processes
新的计算方法来表征罕见功能异常变异对顺式和反式调节疾病过程的影响
  • 批准号:
    10679055
  • 财政年份:
    2022
  • 资助金额:
    $ 60.23万
  • 项目类别:
{{ showInfoDetail.title }}

作者:{{ showInfoDetail.author }}

知道了