Sepp Hochreiter
Sepp Hochreiter helped develop Long Short-Term Memory with Jürgen Schmidhuber. Their 1997 paper addressed a difficulty in recurrent neural networks: learning useful relationships across long stretches of a sequence. LSTM provides a controlled memory path through which information can remain available. Hochreiter later became a professor at Johannes Kepler University Linz, where his research linked machine learning with biological data. He also led the LIT AI Lab. In 2024, he worked with Maximilian Beck and colleagues on xLSTM, a research effort to extend recurrent networks for larger language models. He is affiliated with JKU’s machine learning institute and AI Lab, with research publications dated 2026.
- Now at
- Johannes Kepler University of Linz
- Works
- 316
- Citations
- 126,064
- h-index
- 55
Publication numbers from OpenAlex, as of 14 October 2026.
Timeline
-
1991
Finding a problem in network memory
Analyzed the loss of training signals across time in recurrent networks.
-
1997
Publishing LSTM with Schmidhuber
Published LSTM with Jürgen Schmidhuber to support learning across long delays.
-
2006
Joining Johannes Kepler University
Became a professor and bioinformatics institute leader at JKU Linz.
-
2017
Building an AI research lab
Led the new LIT AI Lab at Linz.
-
2024
Returning to LSTM at larger scale
Joined the team that published xLSTM, an extension of recurrent network memory.
Sources and credits
Photo: C.Stadler/Bwag, via Wikimedia Commons. License: CC BY-SA 4.0. Source: commons.wikimedia.org. Cropped to a square and resized. Ask us to remove or replace this photo.
