Expressnews
February, 2025: Amrita Vishwa Vidyapeetham, the only partner from India in a prestigious international collaboration, has played a significant role in the creation of DiversityOne, a groundbreaking dataset that sets new benchmarks in modelling everyday human behavior using smartphone sensor data. Funded under the European Commission’s H2020 “WeNet: The Internet of Us” project, DiversityOne is one of the largest and most diverse datasets of its kind, designed to fuel cutting-edge research in artificial intelligence, ubiquitous computing, and behavioral sciences.
As the only Indian partner in this European project, Amrita has contributed to the dataset that captures raw, multimodal sensor data from 782 participants across eight countries: India, China, Mongolia, Denmark, Italy, the UK, Paraguay, and Mexico. Spanning four weeks of data collection, the dataset includes 26 types of smartphone sensor data, such as GPS, accelerometer, app usage, and screen time. It also incorporates over 350,000 self-reports paired with psycho-social surveys and demographic information, offering context-rich behavioral insights and enabling detailed cross-cultural and cross-country studies.
The dataset reflects global cultural and lifestyle diversity while serving as a robust resource for researchers working on responsible AI, time series modeling, domain adaptation, and cross-cultural studies. Its size, scale, and inclusivity make it a first-of-its-kind resource for the research community.
The research paper, “DiversityOne: A Multi-Country Smartphone Sensor Dataset for Everyday Life BehaviorModeling,” accepted by the prestigious Proceedings of the ACM on Interactive, Mobile, Wearable, and Ubiquitous Technologies (IMWUT) journal, highlights the scope and technical significance of the dataset. The data offers unique opportunities for AI researchers to address challenges in behavioral prediction, domain generalization, and ubiquitous computing while advancing human-centric AI and machine learning applications.
Amrita collaborated with 15 leading global institutions, including Università di Trento (Italy), EPFL (Switzerland), Aalborg University (Denmark), Jilin University (China), The London School of Economics and Political Science (UK), National University of Mongolia, Universidad Católica “Nuestra Señora de la Asunción” (Paraguay), Idiap Research Institute (Switzerland), University of Tubingen (Germany), Open University of Cyprus (Latsia, Cyprus), Ben Gurion University (Israel) and the Spanish National Research Council (CSIC, Barcelona, Spain) and the Instituto Potosino (IPIYCT, Mexico).
By integrating diverse data from both the Global North and Global South, DiversityOne enables researchers to explore cross-cultural differences and similarities in human behavior. It provides a testbed for innovations in AI-driven behavioralmodeling, responsible AI, and time series adaptation across cultures, with applications in health monitoring, urban planning, and digital wellbeing.
The Amrita lead of this project, Dr. Shyam Diwakar, Director of the Mind Brain Center at Amrita Vishwa Vidyapeethamremarked, “DiversityOne is a testament to the power of global collaboration. Amrita University is proud to contribute to a resource that not only sets new standards in behavioralmodeling but also opens doors to impactful AI solutions that embrace cultural diversity. We are looking forward to see how researchers across the globe will use this dataset that is a testbed for artificial intelligence.”
DiversityOne is openly available to researchers worldwide, providing a comprehensive resource to drive technological advancements in AI and behavioral sciences.
Access the dataset and additional details here:
Research Paper: arxiv.org/pdf/2502.03347
Dataset Information: DiversityOne Website
Dataset Catalogue: DiversityOne Dataset
Amrita Vishwa Vidyapeetham is a globally recognized leader in interdisciplinary research and education, committed to addressing the world’s most pressing challenges through innovation, compassion, and collaboration.
