Part of the TPC Seminar Series
Speaker: Jonas Hübotter, Doctoral Researcher, Learning and Adaptive Systems Group at ETH Zurich
Date: Wednesday, March 5, 2025
Time: 9:00 A.M. to 10:00 A.M. (Central Time)
Location: Virtual
Abstract:
The standard paradigm of machine learning separates training and testing. Training aims to learn a model by inductively extracting general rules from data, and testing applies this model to new, unseen data. We investigate an alternative transductive paradigm where the model is trained specifically for the given task. We show that this paradigm can improve state-of-the-art LLMs, and explore how to select data for test-time training.
Biography:
Jonas Hübotter is a doctoral researcher in the Learning and Adaptive Systems Group at ETH Zurich working with Andreas Krause. Prior to this, he obtained a Master’s degree in Theoretical Computer Science and Machine Learning from ETH Zurich and a Bachelor’s degree in Computer Science and Mathematics from the Technical University of Munich. He has worked in industry at Citadel Securities and Uncountable. His research focuses on local learning, transductive learning, active learning, and reasoning. Fundamentally, Jonas is motivated by the goal of AI systems that learn and improve over time, adapt to their environments, and exhibit strong reasoning capabilities.
Additional Zoom Information:
Meeting ID: 160 220 8288
Passcode: 257460
+16692545252,,1602208288#,,,,*257460# US (San Jose)
+16468287666,,1602208288#,,,,*257460# US (New York)
Dial by your location
- +1 669 254 5252 US (San Jose)
- +1 646 828 7666 US (New York)
- +1 646 964 1167 US (US Spanish Line)
- +1 669 216 1590 US (San Jose)
- +1 415 449 4000 US (US Spanish Line)
- +1 551 285 1373 US (New Jersey)