© TUGraz/IGI

Welcome to the Institute of Theoretical Computer Science at TU Graz

The Institute of Theoretical Computer Science was founded in 1992 to investigate fundamental problems in information processing such as the design of computer algorithms, the complexity of computations and computational models, automated knowledge acquisition (machine learning), the complexity of learning algorithms, pattern recognition with artificial neural networks, computational geometry, and information processing in biological neural systems.

Its research integrates methods from mathematics, computer science and computational neuroscience.

In education this institute is responsible for courses and seminars that introduce students into the basic techniques and results of theoretical computer science. In addition it offers advanced courses, seminars and applied computer projects in computational geometry, computational complexity theory, machine learning, and neural networks.


NeurIPS 2022 - Tutorial and Workshop

Paper accepted for AAAI 2022

December 2022:  "Probabilistic models based on continuous latent spaces, such as variational autoencoders, can be understood as uncountable mixture models where components depend continuously on the latent code."

Alvaro H.C. Correia, Gennaro Gala, Erik Quaeghebeur, Cassio de Campos, Robert Peharz
Continuous Mixtures of Tractable Probabilistic Models

arXiv link to accepted paper

Paper accepted for NeurIPS 2022

November 2022:  "Causal models are powerful reasoning tools, but usually we don't know which model is the correct one. Traditionally, one first aims to find the correct causal model from data, which is then used for causal reasoning", states Robert Peharz.

Christian Toth, Lars Lorch, Christian Knoll, Andreas Krause, Franz Pernkopf, Robert Peharz, Julius Von Kügelgen - Active Bayesian Causal Inference

link to NeurIPS paper
more detailed tweeted Information

Kleine Zeitung Special Edition "Die Kraft der Region"

November 2022: "Machine learning will change the world like the internet and before that computers did", predicts Robert Legenstein.

learn more about this topic
Kleine Zeitung - about this Special Edition

Paper accepted at CVPR

March 2022: Stealthy bit-flip attacks are menacing deep neural network applications. Our novel defence make the attacker's life much harder. Developed in the Dependable-Systems-Lab of the SiliconAustriaLabs.

Ozan Özdenizci and Robert Legenstein
Improving robustness against stealthy weight bit-flip attacks by output code matching. In
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2022. (Link to PDF)

See the featured interview at CVPR Daily.