ZurichAI: Jürgen Schmidhuber
[about]
NOTE: This event is fully booked. Registration will be checked for entry!
We're honoured to host an original legend in the field of machine learning: Prof. Jürgen Schmidhuber. Jürgen has pioneered and advanced many techniques that are foundational to modern AI:
Jürgen led groups at TUM and IDSIA which developed the foundations of today’s most-cited neural nets, e.g., LSTM (1997, most-cited AI of the 20th century), Highway Network (2015 - its later variant ResNet became the most-cited AI of the 21st), DanNet (2011 - the first deep GPU CNN to win image-recognition competitions, before AlexNet/VGG!); Generative Adversarial Network (1990, for Artificial Curiosity), unnormalized linear Transformer (1991). The core ideas behind GANs, Transformers, pre-training, and neural distillation can all be traced to his lab's work.
Note: We're on the 19th floor this time for extra seating capacity, not the usual 14th.
[speakers]
Jürgen Schmidhuber
[details]
- time
- 21 oct 2025 18:00
- location
- ETH AI Center
- address
- Andreasstrasse 5, OAT, 19th floor, 8050, Zürich
- format
- talk
- status
- finished
- tags
- access
- OAT building.