Affective Film Dataset from India (AFDI): Creation and Validation with an Indian Sample

We describe the creation of an affective film dataset for researchers interested in studying a broad spectrum of emotional experiences. Two hundred twenty-two 60-seconds long video clips were selected based on multimedia content analysis and screened in the lab with 407 participants. The participants' ratings mapped to 31 emotion categories in the first stage. Based on the selection criteria, 69 audio-visual clips were selected. These selected affective clips were then presented to 271 participants. Participants rated these clips on rating scales and categorized them into emotion categories. The affective clips were able to induce 19 basic and complex emotion categories reliably. Since the presented dataset is comprised of film clips based on both Indian and western content, the dataset can effectively be used for cross-cultural emotion research. From the dataset, researchers can select emotional movie clips based on the ratings and quantitative measures, including the reliability measures presented in this work. We also show a continuity of emotional experiences using an advanced visualisation technique to complement the existing knowledge based on V-A space with the information on how the transitions among emotion categories are taking place.

Dataset on Emotions using Naturalistic Stimuli (DENS)

Emotion is a constructed phenomenon that emerges from the dynamic interaction of multiple components neurologically, physiologically and behaviorally. Such dynamics can not be captured by static and controlled experiments. Hence, the study of emotion with a naturalistic paradigm is needed. In this dataset, multimedia naturalistic stimuli are used to acquire the emotional dynamics using EEG, ECG, EMG and behavioural scales. The stimuli are multimedia videos collected from youtube for 372 affective words, analyzed with multimedia content analysis to filter out non-emotional stimuli and then validated with university students. The validated stimuli had the least variance in subjective ratings on self-assessment scales. The stimuli are then used to acquire neurological dynamics along with peripheral channels and subjective ratings-valence, arousal, dominance, liking, familiarity, relevance and emotion category. Both the raw data and pre-processed data is provided along with the pre-processing pipeline. This data can be utilized to study dynamic activation and connectivity in the whole brain source localization study, understand the mutual interaction between the central and autonomic nervous system, understand temporal hierarchy using multiresolution tools, and perform machine learning-based classification and complex networks analysis. The data is accessible at \url{10.18112/openneuro.ds003751.v1.0.0}

Report on SEBD 2020

This paper reports on the 28th Italian Symposium on Advanced Database Systems (SEBD 2020), held online as a virtual conference from the 21st to the 24th of June 2020. The topics that were addressed in this edition of the conference were organized in the sessions: ontologies and data integration, anomaly detection and dependencies, text analysis and search, deep learning, noSQL data, trajectories and diffusion, health and medicine, context and ranking, social and knowledge graphs, multimedia content analysis, security issues, and data mining.

Towards Neural-Symbolic AI for Media Understanding

Methods based on Machine Learning have become state-of-the-art in various segments of computing, especially in the fields of computer vision, speech recognition, and natural language processing. Such methods, however, generally work best when applied to specific tasks in specific domains where large training datasets are available. This paper presents an overview of the state-of-the-art in the area of Deep Learning for Multimedia Content Analysis (image, audio, and video), and describe recent works that propose The integration of deep learning with symbolic AI reasoning. We draw a picture of the future by discussing envisaged use cases that address media understanding gaps which can be solved by the integration of machine learning and symbolic AI, the so-called Neuro-Symbolic integration.

Pleasing the ‘Bubble:’ Abe Shinzō’s Strategic Self-Exhibition on Facebook

Not only is Abe Shinzō on the way to becoming Japan’s longest-serving Prime Minister in the country’s history. With more than 1 million followers on Twitter and slightly less than 600 hundred thousand fans on Facebook, he is by far the most successful Japanese political leader on social media. Commentators have described Abe’s turn to social networking services (SNS) as a “revenge” against “traditional” media against the background of a growing use of SNSs by other major Japanese political actors. At any rate, particularly through Facebook, combining text and pictures of himself on and off duty, Abe has successfully established his own mode to communicate with and “exhibit” himself to voters, citizens and the global community of netizens. This paper aims to address the following research question: on which themes and key concepts is this “presentation of the self” based? In other words, how is the Prime Minister communication staff constructing Abe’s “social” image and to which audience is this aimed? Based on Goffman’s theorization and later application of his work on the study of online social interactions, this paper illustrates the strive to ensure the consistency of Abe’s use of the SNS with previously expressed concepts and ideas (e.g., in the 2006 book “A Beautiful Country”), with the aim of pleasing the “bubble” of like-minded individuals constituting Abe’s (online) support base, and avoid issues that might possibly harm the Prime Minister’s reputation. Abe’s Facebook activity (a combination of text and pictures) during a critical time in his second tenure (2017), in which he faced cronyism allegations while coping with gaffes and scandals involving cabinet members, provided a case in point for multimedia content analysis.