?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=Audiovisual%2C+Genre%2C+Neural+and+Topical+Textual+Embeddings+for+TV+Programme+Content+Representation&rft.creator=Nazir%2C+S&rft.creator=Cagali%2C+T&rft.creator=Sadrzadeh%2C+M&rft.creator=Newell%2C+C&rft.description=TV+programmes+have+their+contents+described+by+multiple+means%3A+textual+subtitles%2C+audiovisual+files%2C+and+metadata+such+as+genres.+In+order+to+represent+these+contents%2C+we+develop+vectorial+representations+for+their+low-level+multimodal+features%2C+group+them+with+simple+clustering+techniques%2C+and+combine+them+using+middle+and+late+fusion.+For+textual+features%2C+we+use+LSI+and+Doc2Vec+neural+embeddings%3B+for+audio%2C+MFCC's+and+Bags+of+Audio+Words%3B+for+visual%2C+SIFT%2C+and+Bags+of+Visual+Words.+We+apply+our+model+to+a+dataset+of+BBC+TV+programmes+and+use+a+standard+recommender+and+pairwise+similarity+matrices+of+content+vectors+to+estimate+viewers'+behaviours.+The+late+fusion+of+genre%2C+audio+and+video+vectors+with+both+of+the+textual+embeddings+significantly+increase+the+precision+and+diversity+of+the+results.&rft.subject=Multimedia+systems%3B+Information+filtering%3B+Recommender+systems%3B+Content-based+retrieval&rft.publisher=IEEE&rft.date=2021-01-22&rft.type=Proceedings+paper&rft.language=eng&rft.source=+++++In%3A++020+IEEE+International+Symposium+on+Multimedia+(ISM).++(pp.+pp.+197-200).++IEEE%3A+Naples%2C+Italy.+(2021)+++++&rft.format=text&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10126932%2F1%2FSabaNazir.pdf&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10126932%2F&rft.rights=open