學(xué)術(shù)報(bào)告（一）
報(bào)告題目：Neural Spectrospatial Filter: On Beamforming in the Deep Learning Era
報(bào)告時(shí)間：2023年7月10日（周一）上午9:30
報(bào)告地點(diǎn)：437bwin必贏國際官網(wǎng)B404會(huì)議室
報(bào)告人：汪德亮
報(bào)告人國籍：美國
報(bào)告人單位：美國俄亥俄州立大學(xué)
報(bào)告人簡介：DeLiang Wang received the B.S. degree and the M.S. degree from Peking (Beijing) University and the Ph.D. degree in 1991 from the University of Southern California all in computer science. Since 1991, he has been with the Department of Computer Science & Engineering and the Center for Cognitive and Brain Sciences at The Ohio State University, where he is a Professor and University Distinguished Scholar. He received the U.S. Office of Naval Research Young Investigator Award in 1996, the 2008 Helmholtz Award from the International Neural Network Society, the 2007 Outstanding Paper Award of the IEEE Computational Intelligence Society and the 2019 Best Paper Award of the IEEE Signal Processing Society. He is an IEEE Fellow and ISCA Fellow, and currently serves as Co-Editor-in-Chief of Neural Networks.
報(bào)告摘要：As the most widely-used spatial filtering approach for multi-channel signal separation, beamforming extracts the target signal arriving from a specific direction. We present an emerging approach based on multi-channel complex spectral mapping, which trains a deep neural network (DNN) to directly estimate the real and imaginary spectrograms of the target signal from those of the multi-channel noisy mixture. In this all-neural approach, the trained DNN itself becomes a nonlinear, time-varying spectrospatial filter. How does this conceptually simple approach perform relative to commonly-used beamforming techniques on different array configurations and in different acoustic environments? We examine this issue systematically on speech dereverberation, speech enhancement, and speaker separation tasks. Comprehensive evaluations show that multi-channel complex spectral mapping achieves speech separation performance comparable to or better than beamforming for different array geometries, and reduces to monaural complex spectral mapping in single-channel conditions, demonstrating the versatility of this new approach for multi-channel and single-channel speech separation. In addition, such an approach is computationally more efficient than popular mask-based beamforming. We conclude that this neural spectrospatial filter is capable of superseding traditional and mask-based beamforming.
邀請(qǐng)人：杜博、涂衛(wèi)平
學(xué)術(shù)報(bào)告（二）
報(bào)告題目：網(wǎng)絡(luò)環(huán)境下的魯棒語音處理與安全
報(bào)告時(shí)間：2023年7月10日（周一）上午10:30
報(bào)告地點(diǎn)：437bwin必贏國際官網(wǎng)B404會(huì)議室
報(bào)告人：張曉雷
報(bào)告人國籍：中國
報(bào)告人單位：西北工業(yè)大學(xué)
報(bào)告人簡介：張曉雷，西北工業(yè)大學(xué)教授，博導(dǎo)。清華大學(xué)博士、美國俄亥俄州立大學(xué)博士后。從事語音處理、機(jī)器學(xué)習(xí)、人工智能的研究工作。在Neural Networks、IEEE TPAMI、IEEE TASLP、IEEE TCYB、Computer Speech and Language等期刊、會(huì)議發(fā)表論文80余篇。出版專著1部、譯著1部。承擔(dān)國家重點(diǎn)研發(fā)計(jì)劃、國家自然科學(xué)基金重點(diǎn)項(xiàng)目等省部級(jí)以上項(xiàng)目10余項(xiàng)。入選國家與省部級(jí)青年人才計(jì)劃。獲得國際神經(jīng)網(wǎng)絡(luò)學(xué)會(huì)最佳論文、亞太信號(hào)與信息處理學(xué)會(huì)杰出講者、北京市科學(xué)技術(shù)一等獎(jiǎng)等。目前擔(dān)任Neural Networks、IEEE TASLP等國際期刊的編委、IEEE SLTC委員等。
報(bào)告摘要： 近年來，盡管大數(shù)據(jù)+深度學(xué)習(xí)在語音識(shí)別等任務(wù)上取得了顯著突破，但是語音處理在遠(yuǎn)場、強(qiáng)自然環(huán)境噪聲干擾和人造惡意攻擊下仍然表現(xiàn)出了一定的脆弱性，限制了其在智慧城市等更大范圍的應(yīng)用。如何充分利用網(wǎng)絡(luò)進(jìn)行多設(shè)備安全互聯(lián)是解決該問題的潛在方法之一。本報(bào)告將分享我們?cè)谠摲较虻囊稽c(diǎn)探索，重點(diǎn)介紹自然噪聲和遠(yuǎn)場環(huán)境下的分布式自組織陣列及智能語音應(yīng)用技術(shù)、及聲紋識(shí)別的對(duì)抗樣本攻擊與防御技術(shù)。
邀請(qǐng)人：杜博、涂衛(wèi)平
