報(bào)告題目: Real-World Large-Scale Visual Recognition and Search

報(bào)告專家：阿里巴巴研究員/資深總監(jiān) 華先勝

報(bào)告時(shí)間：2015年11月14日 周六 上午9:30- 11:30

報(bào)告地點(diǎn)：文科樓模擬法庭B-204

報(bào)告內(nèi)容：Image recognition and search have been studied for decades and became very hot again in recently years mainly due to the rapid development of distributed computing, high-dimensional indexing, deep learning and big data analytical techniques. Many visual search and recognition research prototypes and preliminary products are available to the public. However, have we solved all the big technical and non-technical challenges? Has ImageNet solved the recognition problem? What are the key factors of realizing a real-world visual recognition/search system? Are semantic gaps still there, and do they really matter? What is the right business model for those technologies? Which direction that visual search/recognition is going towards? What are still missing? In this talk, we will discuss all these questions based on a real-world visual search and recognition engine that has a considerable number of users.

華先勝，2015年加入阿里巴巴搜索事業(yè)部，任研究員/資深總監(jiān)，負(fù)責(zé)大數(shù)據(jù)多媒體內(nèi)容分析及搜索算法團(tuán)隊(duì)。在此之前先后在微軟中國研究院、微軟總部Bing搜索引擎以及微軟美國研究院任主管研究員、首席研發(fā)主管以及資深研究員。北京大學(xué)博士畢業(yè)，一直從事圖像、視頻內(nèi)容分析和搜索方面的研究和開發(fā)工作。在國際會議和期刊上發(fā)表論文250余篇，擁有50余項(xiàng)授權(quán)專利。2008年因視頻搜索方面的突出貢獻(xiàn)獲麻省理工學(xué)院《技術(shù)評論》 TR35“全球35個(gè)35歲以下青年創(chuàng)新者”稱號，2012年任多媒體領(lǐng)域頂級國際會議 ACM Multimeida 程序委員會主席，2012和2013年連續(xù)兩年任IEEE多媒體會議ICME程序委員會主席，還是多個(gè)多媒體、AI及搜索領(lǐng)域國際會議的領(lǐng)域主席（包括KDD，SIGIR，IJCAI，MM，ICME等），曾任或現(xiàn)任IEEE Trans. On Multimedia及ACM Trans. On Intelligent Systems and Technology雜志副編，曾獲包括ACM Multimedia和IEEE Trans. On CSVT等在內(nèi)的多項(xiàng)最佳論文獎(jiǎng)。