魏思

发布时间:2024-06-25浏览次数:10

魏思

电话:13856056391

E-mail:siwei@iflytek.com

主要研究方向:智能语音、自然语言处理、大模型

个人简介:

魏思,博士,正高级工程师,科大讯飞股份有限公司首席科学家。长期从事智能语音、自然语言处理、通用大模型等人工智能关键核心技术研究,领导讯飞人工智能领域共性基础技术的前瞻规划和技术创新。主持国家重点研发计划“基于大数据的类人智能关键技术与系统(二期)”等多项重大项目科研工作,带领团队突破面向新课程标准的类人答题系统能力精细化评估方案、基于统一接口的多学科类人答题验证系统和自动评测技术、面向实时答题和大规模资源运算的硬件加速和云计算加速技术等关键技术,研制出具有类人考试能力和知识水平的智能答题原型验证系统和个性化学习平台等产品,在教育领域实现千万级别学生用户的典型示范应用。现任科技创新2030—“新一代人工智能”重大项目“大模型系统能力涌现的机理机制研究”的项目负责人,探索大模型系统能力涌现的机理机制理论,基于图文音等“多模态”数据,从宏观微观等“多尺度”角度出发,研究语言理解生成、知识问答、逻辑推理以及多模态等“多层次”涌现能力的原理,并应用于数学解答、自由对话、情感分析等“多场景”任务的效果提升,形成“理论研究与实践验证”的闭环体系。已发表高水平论文70余篇,授权发明专利100余项,获得国家科技进步奖二等奖1项、安徽省科技进步奖一等奖2项,被授予中国青年五四奖章集体、安徽省技术领军人才、MIT TR35 China“远见者”等荣誉。

获奖情况:

2018年,安徽省科学技术进步奖一等奖

2011年,国家科学技术进步奖二等奖

2020年,中国青年五四奖章集体

2009年,安徽省自然科学优秀学术论文奖

2008年,安徽省科学技术进步奖一等奖

2007年,合肥市科学技术进步奖一等奖

2006年,中国科学技术大学学术论坛一等奖

2017年,《麻省理工科技评论》TR35 China“远见者”

2017年,安徽省技术领军人才

2016年,安徽省五一劳动奖章

2010年,安徽省杰出青年科技创新奖

主要论著:

[1]    魏思,巩捷甫,宋巍,宋子尧,王士进. 利用深层语言分析改进中文作文自动评分方法, 中文信息学报, 36(4), 2022.

[2]    魏思,沈双宏,黄振亚,刘淇,陈恩红,苏喻,王士进. 融合通用题目表征学习的神经知识追踪方法研究, 中文信息学报, 36(4), 2022.

[3]    魏思,吴奎,竺博,王士进. 语音评测技术助力英语口语教学与评价, 人工智能, 2019.

[4]    魏思,凌震华,杜俊,刘聪,胡国平,胡郁. 语音信号与信息处理中的深度学习方法, 中国计算机学会通讯, 11(8), 2015.

[5]    Si Wei, Guoping Hu, Yu Hu, Renhua Wang. A new method for mispronunciation detection using Support Vector Machine based on Pronunciation Space Models, Speech Communication, 51(01), 2009.

[6]    Si Wei, Haikun Wang, Qingsheng Liu, Renhua Wang. CDF-Matching for Automatic Tone Error Detection in Mandarin Call System, ICASSP 2007, 4, 2007.

[7]    魏思,刘庆升,胡郁,王仁华. 普通话水平测试电子化系统, 中文信息学报, 20(6), 2006.

[8]    魏思,刘庆升,胡郁,王仁华. 带方言口音普通话自动水平测试, NSMMSC 2005, 24, 2005.

[9]    王士进,巩捷甫,汪意发,宋巍,陈志刚,魏思. 基于要点匹配的文科主观题通用评分, 中文信息学报, 37(6), 2023.

[10]王士进,汪成成,张丹,魏思,王渊. 基于多模态语义分析的试题推荐方法, 中文信息学报, 37(6), 2023.

[11]Boyuan Zheng, Xiaoyu Yang, Yuping Ruan, Zhenhua Ling, Quan Liu, Si Wei, Xiaodan Zhu. SemEval-2021 Task 4: Reading Comprehension of Abstract Meaning, ACL-IJCNLP 2021, 2021.

[12]Jiachen Gu, Tianda Li, Quan Liu, Zhenhua Ling, Zhiming Su, Si Wei, Xiaodan Zhu, Speaker-Aware BERT for Multi-Turn Response Selection in Retrieval-Based Chatbots. CIKM 2020, 2020.

[13]Hui Liu, Zhan Shi, Jiachen Gu, Quan Liu, Si Wei, Xiaodan Zhu. End-to-End Transition-Based Online Dialogue Disentanglement, IJCAI 2020, 2020.

[14]Jianshu Zhang, Jun Du, Yongxin Yang, Yizhe Song, Si Wei, Lirong Dai. A Tree-Structured Decoder for Image-to-Markup Generation, ICML 2020, 119, 2020.

[15]Kailin Chen, Cheng Peng, Yi Li, Dalong Cheng, Si Wei. Accurate Automated Keypoint Detections for Spinal Curvature Estimation, CSI 2019, 11963, 2020.

[16]陈志刚,华磊,刘权,尹坤,魏思,胡国平. 深度学习建模下的自动句子填空技术, 中文信息学报, 33(8), 2019.

[17]Qian Chen, Xiaodan Zhu, Zhenhua Ling, Diana Inkpen, Si Wei. Neural Natural Language Inference Models Enhanced with External Knowledge, ACL 2018, 2018.

[18]Yu Su, Qingwen Liu, Qi Liu, Zhenya Huang, Yu Yin, Enhong Chen, Chris H. Q, Ding, Si Wei, Guoping Hu. Exercise-Enhanced Sequential Modeling for Student Performance Prediction, AAAI 2018, 2018.

[19]Junhua Liu, Zhenhua Ling, Si Wei, Guoping Hu, Lirong Dai. Improving the Decoding Efficiency of Deep Neural Network Acoustic Models by Cluster-Based Senone Selection, Journal of Signal Processing Systems, 90(1), 2018.

[20]Qian Chen, Xiaodan Zhu, Zhenhua Ling, Diana Inkpen, Si Wei. Natural Language Inference with External Knowledge, ICLR 2017, 2017.

[21]Jianshu Zhang, Jun Du, Shiliang Zhang, Dan Liu, Yulong Hu, Jinshui Hu, Si Wei, Lirong Dai. Watch, attend and parse: An end-to-end neural network based approach to handwritten mathematical expression recognition, Pattern Recognition, 71, 2017.

[22]Qian Chen, Xiaodan Zhu, Zhenhua Ling, Si Wei, Hui Jiang, Diana Inkpen. Recurrent Neural Network-Based Sentence Encoder with Gated Attention for Natural Language Inference, EMNLP 2017, 2017.

[23]Quan Liu, Hui Jiang, Andrew Evdokimov, Zhenhua Ling, Xiaodan Zhu, Si Wei, Yu Hu. Cause-Effect Knowledge Acquisition and Neural Association Model for Solving A Set of Winograd Schema Problems, IJCAI 2017, 2017.

[24]Lijuan Liu, Chuang Ding, Yajun Hu, Zhenhua Ling, Yuan Jiang, Ming Zhou, Si Wei. The iFLYTEK System for Blizzard Machine Learning Challenge 2017-ES1, ASRU 2017, 2017.

[25]Qian Chen, Xiaodan Zhu, Zhenhua Ling, Si Wei, Hui Jiang, Diana Inkpen. Enhanced LSTM for Natural Language Inference, ACL 2017, 2017.

[26]Yiming Cui, Zhipeng Chen, Si Wei, Shijin Wang, Ting Liu, Guoping Hu. Attention-over-Attention Neural Networks for Reading Comprehension, ACL 2017, 2017.

[27]许苏魁,戴礼荣,魏思,刘庆峰,高前勇. 自由表述口语语音评测后验概率估计改进方法, 中文信息学报, 31 (2), 2017.

[28]Shiliang Zhang, Cong Liu, Hui Jiang, Si Wei, Lirong Dai, Yu Hu. Nonrecurrent Neural Structure for Long-Term Dependence, TASLP, 25(04), 2017.

[29]Quan Liu, Hui Jiang, Zhenhua Ling, Xiaodan  Zhu, Si Wei, Yu Hu. Commonsense Knowledge Enhanced Embeddings for Solving Pronoun Disambiguation Problems in Winograd Schema Challenge, AAAI 2017, 2017.

[30]Quan Liu, Hui Jiang, Zhenhua Ling, Xiaodan Zhu, Si Wei, Yu Hu. Combing Context and Commonsense Knowledge Through Neural Networks for Solving Winograd Schema Problems, AAAI 2017, 2017.

[31]Zhenya Huang, Qi Liu, Enhong Chen, Hongke Zhao, Mingyong Gao, Si Wei, Yu Su, Guoping Hu. Question Difficulty Prediction for READING Problems in Standard Tests, AAAI 2017, 2017.

[32]Dan Liu, Wei Lin, Shiliang Zhang, Si Wei, Hui Jiang. The USTC NELSLIP Systems for Trilingual Entity Detection and Linking Tasks at TAC KBP 2016, TAC 2016, 2016.

[33]Junhua Liu, Zhenhua Ling, Si Wei, Guoping Hu, Lirong Dai. Cluster-based senone selection for the efficient calculation of deep neural network acoustic models, ISCSLP 2016, 2016.

[34]Jian Tang, Shiliang Zhang, Si Wei, Lirong Dai. Future Context Attention for Unidirectional LSTM Based Acoustic Model, INTERSPEECH 2016, 2016.

[35]Shiliang Zhang, Hui Jiang, Shifu Xiong, Si Wei, Lirong Dai. Compact Feedforward Sequential Memory Networks for Large Vocabulary Continuous Speech Recognition, INTERSPEECH 2016, 2016.

[36]Qian Chen, Xiaodan Zhu, Zhenhua Ling, Si Wei, Hui Jiang. Distraction-Based Neural Networks for Modeling Document, IJCAI 2016, 2016.

[37]Shiliang Zhang, Hui Jiang, Si Wei, Lirong Dai. Rectified linear neural networks with tied-scalar regularization for LVCSR, INTERSPEECH 2015, 2015.

[38]Jun Du, Jianfang Zhai, Jinshui Hu, Bo Zhu, Si Wei, Lirong Dai. Writer adaptive feature extraction based on convolutional neural networks for online handwritten Chinese character recognition, ICDAR 2015, 2015.

[39]Zhigang Chen, Wei Lin, Qian Chen, Xiaoping Chen, Si Wei, Hui Jiang, Xiaodan Zhu. Revisiting Word Embedding for Contrasting Meaning, ACL 2015, 2015.

[40]Quan Liu, Hui Jiang, Si Wei, Zhenhua Ling, Yu Hu. Learning Semantic Word Embeddings based on Ordinal Knowledge Constraints, ACL 2015, 2015.

[41]Jingjie Li, Ian Vince McLoughlin, Cong Liu, Shaofei Xue, Si Wei. Multi-task deep neural network acoustic models with model adaptation using discriminative speaker identity for whisper recognition, ICASSP 2015, 2015.

[42]Sukui Xu, Si Wei, Zhenhua Ling, Qianyong Gao, Lirong Dai, Qingfeng Liu. A Statistical Modeling Approach to Automatic Evaluation of Mandarin Pronunciation, Journal of the Phonetic Society of Japan, 19(1), 2015.

[43]Bing Jiang, Yan Song, Si Wei, Ian Vince McLoughlin, Lirong Dai. Task-aware deep bottleneck features for spoken language identification, INTERSPEECH 2014, 2014.

[44]Bing Jiang, Yan Song, Si Wei, Mengge Wang, Ian Vince McLoughlin, Lirong Dai. Performance evaluation of deep bottleneck features for spoken language identification, ISCSLP 2014, 2014.

[45]Jun Du, Jinshui Hu, Bo Zhu, Si Wei, Lirong Dai. Writer Adaptation Using Bottleneck Features and Discriminative Linear Regression for Online Handwritten Chinese Character Recognition, ICFHR 2014, 2014.

[46]Jun Du, Jinshui Hu, Bo Zhu, Si Wei, Lirong Dai. A Study of Designing Compact Classifiers Using Deep Neural Networks for Online Handwritten Chinese Character Recognition, ICPR 2014, 2014.

[47]Diyuan Liu, Si Wei, Wu Guo, Yebo Bao, Shifu Xiong, Lirong Dai. Lattice based optimization of bottleneck feature extractor with linear transformation, ICASSP 2014, 2014.

[48]Yan Song, Bing Jiang, Yebo Bao, Si Wei, Lirong Dai. i-vector representation based on bottleneck features for language identification, Electronics Letters, 49(24), 2013.

[49]严可,魏思,戴礼荣. 针对发音质量评测的声学模型优化算法, 中文信息学报, 2013.

[50]潘逸倩,魏思,戴礼荣,刘庆峰. 基于得分域多维特征分类器的声纹密码系统, 模式识别与人工智能, 5(7), 2012.

[51]Hussein Hussein, Hue San Do, H. Mixdorff, Hongwei Ding, Qianyong Gao, Guoping Hue, Si Wei, Zhao Chao. Mandarin tone perception and production by German learners, SLaTE 2011, 2011.

[52]龚澍,魏思,胡国平,刘庆峰. 普通话发音检错系统中TANDEM方法的研究与实现, 计算机应用与软件, 2011.

[53]Si Wei, Qianyong Gao, Guoping Hu, Yu Hu. Robust pronunciation evaluation in adverse environments, ISCSLP 2010, 2010.

[54]Si Wei, Qianyong Gao, Guoping Hu, Yu Hu. Robust Speech Evaluation System in adverse environments, ISCSLP 2010, 2010.

[55]Hussein Hussein, Hansjorg Mixdorff, Hue San Do, Si Wei, Shu Gong, Hongwei Ding, Qianyong Gao,Guoping Hu. Towards a computer-aided pronunciation training system for German learners of Mandarin, SLaTE 2009, 2010.

[56]Heng Lu, Zhen-Hua Ling, Si Wei, Lirong Dai, Ren-Hua Wang. Automatic error detection for unit selection speech synthesis using log likelihood ratio based SVM classifier, INTERSPEECH 2010, 2010.

[57]严可,胡国平,魏思,李萌涛,杨晓果,冯国栋. 计算机用于英语背诵题的自动评分技术初探, 计算机应用与软件, 27(7), 2010.

[58]Hussein Hussein, H. Mixdorff, Hue San Do, Si Wei, Qianyong Gao, Shu Gong, Hongwei Ding, Guoping Hu. Development of a Computer-Aided Pronunciation Training System for Teaching Mandarin for German Learners - Pronunciation Errors, ESSV 2010, 2010.

[59]严可,胡国平,魏思,戴礼荣. 面向大规模英语口语机考的复述题自动评分技术研究, NCMMSC 2009, 2009.

[60]严可,胡国平,魏思,戴礼荣,李萌涛,杨晓果,冯国栋. 面向大规模英语口语机考的复述题自动评分技术, 清华大学学报(自然科学版), 49(S1), 2009.

[61]刘庆升,胡郁,魏思,王仁华. 基于KLD差的统计错误模式生成算法, 数据采集与处理, 24(1), 2009.

[62]Yiqian Pan, Si Wei, Renhua Wang. Tone Evaluation of Chinese Continuous Speech Based on Prosodic Words, ISCSLP 2008, 2008.

[63]Si Wei, Yiqian Pan, Guoping Hu, Yu Hu, Renhua Wang. Pronunciation Space Models for Pronunciation Evaluation, ISCSLP 2008, 2008.

[64]Heng Lu, Zhenhua Ling, Si Wei, Yuhu, Lirong Dai, Renhua Wang. Heteronym Verification for Mandarin Speech Synthesis, ISCSLP 2008, 2008.

[65]潘逸倩,魏思,王仁华. 基于韵律信息的连续语流调型评测研究, 中文信息学报, 22(4), 2008.

[66]严可,魏思,王仁华. VTS降噪算法在语音评测中的应用, 第八届中国语音学学术会议, 2008.

[67]刘庆升,魏思,胡郁,郭武,王仁华. 基于语言学知识的发音质量评价算法改进, 中文信息学报, 21(4), 2007.

[68]Qingsheng Liu, Si Wei, Yu Hu, Wu Guo, Renhua Wang. The Application of Phone Weight in Putonghua Pronunciation Quality Assessment, ISCSLP 2006, 11, 2006.

[69]Renhua Wang, Qingfeng Liu, Si Wei. Putonghua Proficiency Test and Evaluation, Advances in Chinese Spoken Language Processing, 2006.

[70]Si Wei, Qingsheng Liu, Yu Hu, Renhua Wang. Automatic Mandarin pronunciation scoring for native learners with dialect accent, INTERSPEECH 2006, 2006.