2,902 research outputs found
Recommender Systems under European AI Regulations
No full textRecommender Systems under European AI Regulations
Tommaso Di Noia, Nava Tintarev, Panagiota Fatourou, Markus Schedl
Communications of the ACM, Volume 65 Issue
Retrieving Relevant and Diverse Movie Clips Using the MFVCD-7K Multifaceted Video Clip Dataset
No full textMultimedia search is an emerging area in information retrieval (IR) and recommender systems (RS) research. However, there is a lack of standardized audiovisual datasets that include rich content descriptors, which are a necessity in content-based IR and RS. The contributions of this paper are twofold: First, we present a new multimedia dataset of movie clips, named MFVCD-7K Multifaceted Video Clip Dataset, that comes with low-level and semantic multimodal descriptions of their content (textual, audio, and visual). In addition, we showcase the use of this dataset for a novel content-based video clip retrieval and result diversification task we introduce. We investigate baseline algorithms for retrieval and diversification, and provide experimental results according to relevance and diversity measures. We believe that both dataset and baseline results constitute an important asset for the IR, RS, and multimedia communities
Frontmatter, Table of Contents, Preface, List of Authors
No full textFrontmatter, Table of Contents, Preface, List of Author
MMTF-14K: A Multifaceted Movie Trailer Dataset for Recommendation and Retrieval
No full text<p>The MMTF-14K dataset provides a stable and extensive source for devising and evaluating movie recommender systems. MMTF-14K contains <strong><a href="https://mmprj.github.io/mtrm_dataset/datasets">audio and visual descriptors</a></strong> in addition to ratings and metadata for 13,623 Hollywood-type movie trailers. The dataset therefore facilitates research on content-based recommender systems, where content refers not only to metadata, but specifically to visual and auditory characteristics of movies. The data comes also with several baselines <a href="https://mmprj.github.io/mtrm_dataset/benchmark">benchmarking results</a> for uni-modal and multi-modal recommendation systems. The dataset therefore facilitates research on movie recommendation. In addition, the rich data supports the exploration of other multimedia tasks such as popularity prediction, genre classification, or auto-tagging (aka tag prediction).</p>
<p>The MMTF-14K dataset has been created as a joint research work by <a href="http://www.ir.disco.unimib.it/yashar-deldjoo/">Yashar Deldjoo </a>(Politecnico di Milano, Italy), <a href="http://www.campus.pub.ro/lab7/gconstantin/">Mihai Gabriel Constantin </a>and <a href="http://campus.pub.ro/lab7/bionescu/">Bogdan Ionescu </a>(University Politehnica of Bucharest, Romania), <a href="http://www.cp.jku.at/people/schedl/">Markus Schedl </a>(Johannes Kepler University Linz, Austria), and <a href="https://scholar.google.it/citations?hl=en&user=dTSOPCMAAAAJ&view_op=list_works&sortby=pubdate">Paolo Cremonesi </a>(Politecnico di Milano, Italy).</p>
<p>We would like to acknowledge MovieLens here for providing a stable benchmark dataset of movies containing individual user ratings and metadata which is an enabler for doing research on movie recommendation. Please consider the <a href="http://files.grouplens.org/datasets/movielens/ml-20m-README.html">MovieLens-20M web page</a> for more details on the ratings and tags datasets.</p>
<p>For acknowledgments please use our paper:</p>
<p>@inproceedings{deldjooMMTF14K, <br>
title={MMTF-14K: A Multifaceted Movie Trailer Feature Dataset for Recommendation and Retrieval}, <br>
author={Deldjoo, Yashar and Constantin, Mihai Gabriel and Schedl, Markus and Ionescu, Bogdan and Cremonesi, Paolo}, <br>
booktitle={Proceedings of the 9th ACM Multimedia Systems Conference}, <br>
year={2018}, <br>
organization={ACM}}</p>
<p>For further inquiries you are free to contact Yashar Deldjoo through his email: <a href="mailto:[email protected]">[email protected] </a>.</p>The link to the dataset can be also found in: https://mmprj.github.io/mtrm_dataset/inde
DFU, Volume 3, Multimodal Music Processing, Complete Volume
No full textDFU, Volume 3, Multimodal Music Processing, Complete Volum
Impact of Listening Behavior on Music Recommendation.
Full text linkThe next generation of music recommendation systems will be increasingly intelligent and likely take into account user behavior for more personalized recommendations. In this work we consider user behavior when making recommendations with features extracted from a user’s history of listening events. We investigate the impact of listener’s behavior by considering features such as play counts, “mainstreaminess”, and diversity in music taste on the performance of various music recommendation approaches. The underlying dataset has been collected by crawling social media (specifically Twitter) for listening events. Each user’s listening behavior is characterized into a three dimensional feature space consisting of play count, “mainstreaminess” (i.e. the degree to which the observed user listens to currently popular artists), and diversity (i.e. the diversity of genres the observed user listens to). Drawing subsets of the 28,000 users in our dataset, according to these three dimensions, we evaluate whether these dimensions influence figures of merit of various music recommendation approaches, in particular, collaborative filtering (CF) and CF enhanced by cultural information such as users located in the same city or country
User geospatial context for music recommendation in microblogs
Full text linkMusic information retrieval and music recommendation are seeing a paradigm shift towards methods that incorporate user context aspects. However, structured experiments on a standardized music dataset to investigate the effects of do-ing so are scarce. In this paper, we compare performance of various combinations of collaborative filtering and geospatial as well as cultural user models for the task of music recom-mendation. To this end, we propose a geospatial model that uses GPS coordinates and a cultural model that uses seman-tic locations (continent, country, and state of the user). We conduct experiments on a novel standardized music collec-tion, the “Million Musical Tweets Dataset ” of listing events extracted from microblogs. Overall, we find that modeling listeners ’ location via Gaussian mixture models and comput-ing similarities from these outperforms both cultural user models and collaborative filtering. Categories and Subject Descriptors Information systems [Information retrieval]: Music rec-ommendation; Human-centered computing [Collaborative and social computing]: Social medi
Introduction to the ICWE 2022 Special Issue
No full textThe International Conference on Web Engineering (ICWE) is the premier annual conference on Web engineering and associated technologies. ICWE aims to bring together researchers and practitioners from various disciplines in academia and industry to tackle the emerging challenges in the engineering of Web applications, the problems with its associated technologies, and the impact of those technologies on society and culture. ICWE 2022 took place in Bari, Italy on 5-8 July 2022. All sessions of the conference were also offered to online participants. This special issue includes extended articles of the best papers presented at ICWE 2022.
On the Influence of User Characteristics on Music Recommendation Algorithms
Full text linkWe investigate a range of music recommendation algorithm combinations, score aggregation functions, normalization techniques, and late fusion techniques on approximately 200 million listening events collected through Last.fm. The overall goal is to identify superior combinations for the task of artist recommendation. Hypothesizing that user characteristics influence performance on these algorithmic combinations, we consider specific user groups determined by age, gender, country, and preferred genre. Overall, we find that the performance of music recommendation algorithms highly depends on user characteristics
Investigating country-specific music preferences and music recommendation algorithms with the LFM-1b dataset
Full text linkRecently, the LFM-1b dataset has been proposed to foster research and evaluation in music retrieval and music recommender systems, Schedl (Proceedings of the ACM International Conference on Multimedia Retrieval (ICMR). New York, 2016). It contains more than one billion music listening events created by more than 120,000 users of Last.fm. Each listening event is characterized by artist, album, and track name, and further includes a timestamp. Basic demographic information and a selection of more elaborate listener-specific descriptors are included as well, for anonymized users. In this article, we reveal information about LFM-1b’s acquisition and content and we compare it to existing datasets. We furthermore provide an extensive statistical analysis of the dataset, including basic properties of the item sets, demographic coverage, distribution of listening events (e.g., over artists and users), and aspects related to music preference and consumption behavior (e.g., temporal features and mainstreaminess of listeners). Exploiting country information of users and genre tags of artists, we also create taste profiles for populations and determine similar and dissimilar countries in terms of their populations’ music preferences. Finally, we illustrate the dataset’s usage in a simple artist recommendation task, whose results are intended to serve as baseline against which more elaborate techniques can be assessed.Version of recor
- …
