Skip to main content

Claes Strannegård


Data Science and AI
Visiting address
Rännvägen 6B
41258 Göteborg
Room number
Postal address
41296 Göteborg

About Claes Strannegård

I am an associate professor in cognitive science at the Department of Philosophy, Linguistics and Theory of Science, University of Gothenburg and the Department of Applied Information Technology at Chalmers University of Technology.

My PhD thesis was about mathematical logic and belongs to the research tradition that began with Gödel’s incompleteness theorems. I spent many years in the private sector as a founder and CEO of research spinoffs in the domain of artificial intelligence. Eventually I returned to the academic world and started doing cross-disciplinary research on human reasoning and artificial general intelligence together with psychologists, neuroscientists, logicians, linguists, and mathematicians.


In our research team we have developed systems that model human reasoning in logic, arithmetic, progressive matrices, and number sequences. We have also developed versatile systems that can adapt to and operate in multiple domains. Our team won prizes from the Artificial General Intelligence Society in 2013, 2014, and 2015. A theme in our research is to exploit cognitive limitations of the human mind for constructing intelligent systems with manageable computational complexity. I work on the following research projects:

  • Automatic concept development. In this project we work with our own transparent networks that are intended to be relatively understandable by design, unlike standard (deep) neural networks. The transparency makes it relatively easy for us to define development rules that describe how the networks evolve in interaction with the environment through addition and deletion of memory structures. The model is developed in our AGI papers from 2012 (about integrating learning and reasoning), 2013 (about dynamic concept formation), and 2015 (about emotional concept development).
  • Automatic rule development. In this project we develop a program for arbitrary symbolic domains. We introduced a notion of abstract domain and developed a prototype system for learning and reasoning about arbitrary such domains. An early version of our program Alice in Wonderland was presented at CogSci 2014 and AGI 2014 and a later version at AGI 2015.
  • Deductive reasoning with bounded resources. In this project we work with proof systems with bounded cognitive resources. The systems were introduced in our 2010 JLLI paper (about propositional logic) and 2013 JLLI paper (about first-order logic). We make experiments in logical reasoning using methods of experimental psychology and build computational models by combining our proof systems with results from cognitive psychology. Our research is supported by a grant from the Swedish Research Council (VR 2012-1000) for 2013-2016.
  • Inductive reasoning with bounded resources. In this project we try to make programs that score high on IQ tests. We developed one program for progressive matrix problems and one for number sequence problems. Our strategy in both cases is to use a cognitive architecture and try to imitate human reasoning. The research was initiated in our 2013 CSR paper (about progressive matrices) and continued in our 2013 CSR paper (about number sequences).
  • Networks and Types. This project is about computational linguistics. The goal is to combine the transparent networks with the TTR-framework (type theory with records). Thus we aim for a linguistic framework, where semantics is ultimately based on perception. For instance, an association link might be formed between the phonetic sequence "cold" and a receptor for cold temperature of the network. This research is supported by a grant from the Swedish Research Council (VR 2013-4873) for 2014-2016.


I have taught courses in the fields of logic, artificial intelligence, linguistics, programming, and innovation. Currently I supervise one PhD student.


I have been working with several academic spinoff companies, including the following:

Safelogic. I founded the formal verification company Safelogic in 1999 with the idea of using resource-bounded automatic theorem proving for analyzing and verifying integrated circuit designs. Safelogic was mainly financed by venture capital and had some 20 employees when it was acquired by Jasper Design Automation in 2004. Jasper continued to develop Safelogic's tool with the main product development node still in Gothenburg (the former Safelogic team) and a sales and marketing office in Silicon Valley. In June 2014, Jasper was acquired by Cadence in a 170 MUSD deal. The development of Safelogic's original product continues and several former Safelogic employees are now employed at Cadence' recently formed node in Gothenburg.

Optisort. I founded the company Optisort in 2008 with the idea of sorting waste automatically by means of artificial neural networks. The company was restarted under the name Refind in 2012. Refind is now a leader in automatic sorting of used batteries into environmental categories. Refind machines operate in the US and in England, where it has reached a dominating market position by sorting about 2/3 of the country's used batteries. The company has predominantly been customer financed and the number of employees has been in the range 5--10. Refind appeared on the Top 25 Nordic cleantech startups list in 2015. The company recently started sorting used mobile phones in a European Horizon 2020 project.

var _gaq = _gaq || []; _gaq.push(['_setAccount', 'UA-37108064-1']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') ''; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })();