SHL Activity Recognition Challenge 2025

Participate in the 7th SHL Activity Recognition Challenge at HASCA/Ubicomp 2025, focusing on applying foundation models to transportation mode recognition.

The seventh edition of the SHL Activity Recognition Challenge is announced, to be held at HASCA/Ubicomp 2025. This year’s challenge focuses on applying foundation models, such as large language models, to transportation mode recognition.

The task is to accurately recognize eight modes of locomotion and transportation using a user-independent approach. Participants are required to develop a solution leveraging recognized foundation models like TimesNet, Chronos, GPT, BERT, or Flamingo.

• The provided dataset includes training, validation, and testing sets.
• Submissions should implement an algorithmic pipeline using foundation models to process sensor data and output recognized activities.
• A technical paper explaining methods and development processes is required for presentation at the HASCA Workshop.

For full rules and timeline, visit http://www.shl-dataset.org/activity-recognition-challenge-2025/. Registration deadline is 30.04.2025, and submission deadline is 16.06.2025.

Contact: shldataset.challenge@gmail.com

Tags: SHL Activity Recognition Challenge, HASCA/Ubicomp 2025, Foundation Models, Transportation Mode Recognition, Machine Learning, AI Research, Activity Recognition