Interdisciplinary Workshop on Analysis of TV content

March 14th - National Institute of Informatics, Tokyo

9:30 - 18:00

In the context of ERC Demoserie european project, the ERC Demoserie team and the Japanese French Laboratory of Informatics organize an interdisciplinary workshop on TV content analysis bringing together researchers in social science and researchers in computer science.

Workshop is open to the public upon prior registration

Detailed program

9:30 : Opening  -  Pr. Philippe Codognet, Japanese French Laboratory for Informatics, National Institute of Informatics

Session 1. TV series analysis (chair: Philippe Codognet)

The ERC Demoserie research project aims to study TV series' aesthetic potential for visualising political and ethical issues and examine how 'security TV series' are conceived both by their creators and audiences. It also has the objective to grasp the influence of ‘security TV series’ on their viewers, in particular how they create shared and shareable moral values in the EU and beyond, elucidate TV series’ crucial role in developing a collective understanding of democracy in its relation to fiction, politics and security and explore how they raise awareness for the protection of individuals and societies in current and future crises.

We have been engaged in research on understanding and generating stories and supporting creative writing, with an emphasis on the perspectives of professional creators. This talk introduces a series of studies and shows how we envisioned and implemented interdisciplinary research. Moreover, mainly focusing on our proposed system, "COMPASS," the importance of incorporating the perspectives of experts in AI research and development will be confirmed. Our COMPASS is a creative writing support system that suggests completing the missing information that storytellers unintentionally omitted. We conducted a user study of four professional creators who use Japanese and confirmed the system's usefulness. We hope this effort will be a basis for further research on collaborative creation between creators and AI.

Keynote

Status on user's behaviour while choosing TV series 

After a brief description of the TV series new usages and the new user's paradigm introduced by the OTT propositions, we will present Betaseries analysis on the user's behaviours while choosing a series to watch (based on pools), his satisfaction level towards recommandation tools today and his expectations on this subject.

Lunch break 

Session 2. Emotions on TV screen (chair: Camille Guinaudeau)

A fundamental understanding of human perception, sentiment, and impressions evoked from multimedia contents can benefit various applications starting from film-making over marketing up to video recommendations. In my research, I alleviate knowledge from human-generated data such as Social Media or the Web, to analyze the impression of multimedia contents. The first half of the talk introduces our method of estimating evoked emotions of Social media videos based on knowledge from their respective user-generated comments. The second half introduces research on estimating the perceived impression of a word or sound (i.e., the mental image it creates), based on analyzing Vision and Language data. Through these researches, we can get a step closer to fully understanding human impressions of multimedia contents.

Challenges of an AI-Powered Content-Based Search Engine for TV Series and Films

 A few months ago, the DEMOSERIES team discussed with some fellow AI researchers the possibility of creating an AI-powered content-based search engine for TV series (which would also apply to films, reason why I include them here). My field of research is not AI, but having developed since several years a specialisation in ethics of AI, I used my general knowledge of it to imagine how this search engine could work and the challenges that might arise during its elaboration. Some of my reflections might appear obvious or naive to AI specialists, or maybe simply wrong, but I am curious about their reactions, and maybe some insights will be interesting for them.

Session 3. Gender,  Media and Politics (chair: Sandra Laugier)

The Gender Equality Monitor project is a multidisciplinary research project that aims to describe automatically representation and treatment differences existing between women and men in the French-language media such as TV, radio, newspapers and song lyrics collections. In this context, automatic tools have been developed for gender classification from audio or video features, and have been applied for analysis of french media. This presentation will describe the tools and their performance as well as the preliminary conclusions of the studies carried out on the representations of genres in media and its impact on models trained from multimedia data.

This talk will consider potential advantages and limitations of AI-based video and audio analysis of gender representation in TV series. Drawing on our recent study of female representation in three contemporary Russian series Two Hills (Start, 2022-), An Ordinary Woman (TV-3/Premiere, 2018-2021), and The Sleepers (Yuri Bykov, Pervyi Kanal, 2017), we will discuss our methodology and potential applications of AI in this kind of investigation. 

Analyzing the appearances of political figures in large-scale news archives is increasingly important with the growing availability of large-scale news archives and developments in computer vision. We have developed a deep learning-based method combining face detection, tracking, and classification, without requiring any re-training when targeting new individuals, only needing few images of target individuals. We validate our method by testing it against two news archives spanning 10 to 20 years, one containing three US cable news (CNN, Fox News, MSNBC) and the other including two major Japanese news programs (NHK news7, HODO station). With this, we allow researchers to explore huge TV archives on information based on TV screen time for specific individuals of interest.

Closing discussion

Access

The workshop will take place in Room 1208, 12th floor, at the National Institute of Informatics, Tokyo, Japan

Organizing committee