15/07/2026
🎤 Join us for the GMUM Group Meeting!
We're excited to invite you to the GMUM Group Meeting, featuring an introductory talk on Joint-Embedding Predictive Architectures (JEPA) by Tomasz Wojnar and Marcin Przewięźlikowski, followed by a guest talk from Karan Ahuja.
Karan Ahuja is the Lisa Wissner-Slivka and Benjamin Slivka Assistant Professor of Computer Science at Northwestern University, where he directs the Sensing, Perception, Interactive Computing & Experiences (SPICE) Lab. His research develops technologies that sense, track, and understand humans to augment their daily lives. His projects have been open-sourced, deployed in real-world settings, licensed, and shipped as product features used by over 200 million users, shaping flagship products at companies including Google and Apple. Karan received his Ph.D. in Human-Computer Interaction from Carnegie Mellon University in 2023, and his work has been recognized with numerous honors, including Forbes 30 Under 30, MIT Innovators Under 35 Asia, the Siebel Fellowship, the 2024 ACM SIGCHI Outstanding Dissertation Award, and the 2025 ACM SIGCHI Special Recognition Award for pioneering input and interaction guidelines in Android XR.
➡️ who? Tomasz Wojnar and Marcin Przewięźlikowski
➡️ where? Wydział Matematyki i Informatyki UJ
➡️ when? Friday (17 July 2026) at 10.15
➡️ Talk: “Introduction to JEPA”
Abstract:
Joint-Embedding Predictive Architectures (JEPAs) offer a robust paradigm for self-supervised learning by predicting abstract representations in latent space rather than reconstructing raw pixels. During the seminar, we will present the JEPA framework, discussing its architectural principles, practical applications and mathematical foundations.
➡️ who? Karan Ahuja
➡️ where? Wydział Matematyki i Informatyki UJ
➡️ when? Friday (17 July 2026) at 12.15
➡️ Talk: “Understanding Ourselves Through Our Digital Self”
Abstract:
The rapid proliferation of consumer devices has created an unprecedented opportunity to revolutionize how we design and implement systems that understand and quantify human health and wellness in real-world settings. Today's smartphones and smartwatches offer basic health metrics like step count, pulse, and blood oxygenation, providing coarse digital representations of users' physical activities. While these simple measurements already benefit millions, there is immense untapped potential. My research develops human-centered technologies and interaction paradigms that enable consumer devices to capture rich, continuous representations of users' physical lives through advances in machine learning, sensor fusion, and edge computing. This transformation of everyday devices into intuitive and unobtrusive health monitoring systems could enable longitudinal wellness tracking, injury prevention, rehabilitation progress tracking, precise calorie counting, mental health promotion, chronic disease monitoring, and enhanced independence for aging populations. Critically, these advances must maintain both user practicality, personalization and acceptability, requiring strategic development of new sensors while maximizing the potential of existing hardware through thoughtful interface design and maintaining user privacy.
🔥 We warmly welcome everyone to join us in person and take part in these exciting discussions.
