1st International Workshop on Multimodal Media Data Analytics (MMDA 2016) Stefanos Vrochidis1 , Maite Melero2 , Leo Wanner2,3 , Jens Grivolla2 and Yannick Estève4 1 Information Technologies Institute - Centre for Research and Technology Hellas, Thessaloniki, Greece, 2 Universitat Pompeu Fabra, Barcelona, Spain, 3 Catalan Institute for Research and Advanced Studies (ICREA), Barcelona, Spain, 4 LIUM - University of Le Mans, France Workshop Title - 1st International Workshop on Multimodal Me- 3 Workshop themes dia Data Analytics (MMDA 2016) In conjunction with the 22nd European Conference on Artificial Research topics of interest for this workshop included, but were not Intelligence (ECAI) 2016. limited to: • Multimedia indexing and retrieval 1 Motivation • Multilingual content extraction • Concept extraction from multilingual text and multimedia The rapid advancements of digital technologies, as well as the pen- • Speech recognition etration of internet and social media usage have resulted in a great • Person identification in audiovisual content increase of heterogeneous multimedia data production worldwide. In • Content-based video segmentation many application fields (e.g. journalism, media monitoring) intelli- • Machine translation of media content gent systems are required for supporting the stakeholders. In recent • Sentiment analysis and opinion mining years, research has been directed towards the development of intel- • Social media analytics ligent systems and tools dealing with media content analysis. How- • Contributor and influencer analysis in social media ever, the magnitude, the diversity and the heterogeneity of the data • Context representation and extraction approaches involved require novel intelligent techniques from the broad area • Semantic content modelling and integration of artificial intelligence to deal and extract meaningful interpreta- • Multimedia topic and event detection tion and provide decision support and summarisation services. This • Semantic reasoning for media content applications need has been reflected also by relevant research projects, such as • Intelligent decision support services based on media content MULTISENSOR1 and EUMSSI2 , which have focused on providing • Multilingual summarisation of media content multimodal analytics of heterogeneous data with an aim to support • Visualisation and interactive interfaces for news journalism, media monitoring, international investments and second • Multimodal content fusion and retrieval screen applications. • Clustering and classification of multimedia In this context, MMDA presented works relevant to media man- • Web crawling and scraping agement, multimedia content analysis, concept- and event-based in- • Intelligent application for media management and retrieval dexing, semantic integration and retrieval, as well as multimodal fu- sion and content summarisation. In addition, an invited talk on au- diovisual linking was given by Roeland Ordelman, University of 4 Organisation Twente, Netherlands. Chairs • Stefanos Vrochidis (Information Technologies Institute/CERTH, 2 Aims of the workshop Greece) The goal of MMDA workshop was twofold. First, the workshop fo- • Maite Melero (UPF, Spain) cused on presenting the most recent methods for the extraction, in- • Leo Wanner (ICREA-UPF, Spain) terpretation, retrieval, fusion, aggregation, summarisation and visual- • Jens Grivolla (UPF, Spain) isation of news content. Second, it aimed at bringing together practi- • Yannick Estève (LIUM-University of Le Mans, France) tioners and researchers, both from the multimodal analytics and me- dia industry domains, to share ideas and experiences in designing Program Committee and implementing novel multimedia and multilingual analysis and • Ioannis Kompatsiaris (CERTH-ITI, Greece) retrieval techniques and tools for journalistic and media monitoring • Martha Larson (TUDelft, Netherlands) applications. • Sylvain Meignier (LIUM, France) 1 http://www.multisensorproject.eu/ • Maria Eskevich (Centre for Language Studies, Faculty of Arts, 2 http://www.eumssi.eu/ Radboud University, Netherlands) • Frederic Bechet (LIF, France) • Peter Bell (University of Edinburgh, UK) • Vladimir Alexiev (Ontotext, Bulgaria) • Ilias Giallampoukidis (CERTH-ITI, Greece) • Mariana Damova (Mozajka, Bulgaria) • Klaus Schoeffmann (Klagenfurt University, Austria) • Stephanie Elzer Schwartz (Millersville University, USA) • Dimitrios Liparas (CERTH-ITI, Greece) • Geraldine Damnati (Orange, France) • Stamatia Dasiopoulou (UPF, Spain) • Georges Linares (LIA, France) • Simon Mille (UPF, Spain) • Symeon Papadopoulos (CERTH-ITI, Greece) • Guillaume Gravier (IRISA, CNRS, France) • Ioannis Arapakis (Eurecat, Spain) • Benoit Huet (EURECOM, France) • Reinhard Busch (LT, Germany) • Gerard Casamayor (UPF, Spain) • Gregor Thurmair (LT, Germany) 5 Invited talk Title: Audiovisual Linking: Scenarios, Approaches and Evaluation Presenter: Roeland Ordelman, University of Twente, Twente, Netherlands. Abstract: The concept of (hyper)linking, well-known in the text do- main, has been inspiring researchers and practitioners in the audiovi- sual domain since many years. In this talk, an overview of a number of use scenarios that can benefit from audiovisual linking is given, some approaches that have been proposed over the last few years are presented and the evaluation of audiovisual linking is discussed. 6 List of papers • D. Liparas, S. Vrochidis, I. Kompatsiaris, G. Casamayor, L. Wan- ner, I. Arapakis, D. García Soriano, R. Busch, B. Vaisman, B. Simeonov, V. Alexiev, A. Belous, E. Jamin, N. Heise, T. Wagner, M. Jugov, M. Eckhoff, T. Forrellat and M. Puigbó, The MULTI- SENSOR project – Development of Multimedia Content Integra- tion Technologies for Journalism, Media Monitoring and Interna- tional Exporting Decision Support • J. Grivolla, Y. Estève, E. Herder, N. Le, K. Macquarrie, R. Marín, S. Meignier, M. Melero, J.M. Odobez and S. Preuß, The EUMSSI project – Event Understanding through Multimodal Social Stream Interpretation • J. Grivolla, M. Melero and T. Badia, EUMSSI: Multilayered anal- ysis of multimedia content using UIMA, MongoDB and Solr • I. Gialampoukidis, D. Liparas, S. Vrochidis and I. Kompatsiaris, Query-based Topic Detection Using Concepts and Named Entities • P.A. Broux, D. Doukhan, S. Petitrenaud, S. Meignier and J. Car- rive, An active learning method for speaker identity annotation in audio recordings • Y. Estève, S. Ghannay and N. Camelin, Recent improvements on error detection for automatic speech recognition • S. Mille, M. Ballesteros, A. Burga, G. Casamayor and L. Wan- ner, Multilingual Natural Language Generation within Abstrac- tive Summarization • J. Codina-Filbà and L. Wanner, Combining Dictionary- and Corpus-Based Concept Extraction