{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2024,8,7]],"date-time":"2024-08-07T07:35:32Z","timestamp":1723016132996},"publisher-location":"California","reference-count":0,"publisher":"International Joint Conferences on Artificial Intelligence Organization","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2018,7]]},"abstract":"<jats:p>Making sense of the surrounding context and ongoing\n\nevents through not only the visual inputs but\n\nalso acoustic cues is critical for various AI applications.\n\nThis paper presents an attempt to learn\n\na neural network model that recognizes more than\n\n500 different sound events from the audio part of\n\nuser generated videos (UGV). Aside from the large\n\nnumber of categories and the diverse recording conditions\n\nfound in UGV, the task is challenging because\n\na sound event may occur only for a short\n\nperiod of time in a video clip. Our model specifically\n\ntackles this issue by combining a main subnet\n\nthat aggregates information from the entire clip\n\nto make clip-level predictions, and a supplementary\n\nsubnet that examines each short segment of the clip\n\nfor segment-level predictions. As the labeled data\n\navailable for model training are typically on the clip\n\nlevel, the latter subnet learns to pay attention to segments\n\nselectively to facilitate attentional segment-level\n\nsupervision. We call our model the M&amp;mnet,\n\nfor it leverages both \u201cM\u201dacro (clip-level) supervision\n\nand \u201cm\u201dicro (segment-level) supervision derived\n\nfrom the macro one. Our experiments show\n\nthat M&amp;mnet works remarkably well for recognizing\n\nsound events, establishing a new state-of-theart\n\nfor DCASE17 and AudioSet data sets. Qualitative\n\nanalysis suggests that our model exhibits strong\n\ngains for short events. In addition, we show that the\n\nmicro subnet is computationally light and we can\n\nuse multiple micro subnets to better exploit information\n\nin different temporal scales.<\/jats:p>","DOI":"10.24963\/ijcai.2018\/463","type":"proceedings-article","created":{"date-parts":[[2018,7,5]],"date-time":"2018-07-05T01:49:10Z","timestamp":1530755350000},"page":"3336-3342","source":"Crossref","is-referenced-by-count":14,"title":["Learning to Recognize Transient Sound Events using Attentional Supervision"],"prefix":"10.24963","author":[{"given":"Szu-Yu","family":"Chou","sequence":"first","affiliation":[{"name":"Graduate Institute of Networking and Multimedia, National Taiwan University"},{"name":"Research Center for Information Technology Innovation, Academia Sinica"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jyh-Shing","family":"Jang","sequence":"additional","affiliation":[{"name":"Graduate Institute of Networking and Multimedia, National Taiwan University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yi-Hsuan","family":"Yang","sequence":"additional","affiliation":[{"name":"Research Center for Information Technology Innovation, Academia Sinica"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"10584","event":{"number":"27","sponsor":["International Joint Conferences on Artificial Intelligence Organization (IJCAI)"],"acronym":"IJCAI-2018","name":"Twenty-Seventh International Joint Conference on Artificial Intelligence {IJCAI-18}","start":{"date-parts":[[2018,7,13]]},"theme":"Artificial Intelligence","location":"Stockholm, Sweden","end":{"date-parts":[[2018,7,19]]}},"container-title":["Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence"],"original-title":[],"deposited":{"date-parts":[[2018,7,5]],"date-time":"2018-07-05T01:53:04Z","timestamp":1530755584000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ijcai.org\/proceedings\/2018\/463"}},"subtitle":[],"proceedings-subject":"Artificial Intelligence Research Articles","short-title":[],"issued":{"date-parts":[[2018,7]]},"references-count":0,"URL":"https:\/\/doi.org\/10.24963\/ijcai.2018\/463","relation":{},"subject":[],"published":{"date-parts":[[2018,7]]}}}