Public defence in Computer Science, M.Sc. Alexander Aushev

Advancing complex simulations with efficient AI techniques for real-world problem-solving. Public defence from the Aalto University School of Science, Department of Computer Science.
Doctoral hat floating above a speaker's podium with a microphone

Title of the doctoral thesis: Sample-efficient inference for agent-based cognitive models and other computationally intensive simulators

Doctoral student: Alexander Aushev
Opponent: Prof. Gilles Louppe, University of Liège, Belgium
Custos: Prof. Samuel Kaski, Aalto University School of Science, Department of Computer Science

Alex Aushev, a doctoral candidate at Aalto University, is set to present his dissertation, “Sample-efficient Inference for Agent-based Cognitive Models and Other Computationally Intensive Simulators”, on December 21 at 12:00 in lecture hall T2 of the Computer Science building (Konemiehentie 2, 02150 Espoo). His work promises substantial advancements in fields like epidemiology, cosmology, and cognitive science. 

The dissertation focuses on enhancing the efficiency of computationally intensive simulations, which are a vital tool in complex scientific and industrial domains. Specifically, Aushev's work aims to reduce computational resources without sacrificing accuracy. In areas such as agent-based computational cognitive models, where traditional methods require numerous, costly simulations, his research represents a significant breakthrough, offering considerable time and resource savings. 

At the heart of Aushev's research is 'likelihood-free inference' (LFI), an innovative approach for scenarios where traditional statistical models are insufficient due to complex simulations. His dissertation introduces novel techniques to effectively handle complex noise in simulations, accurately estimate unknown dynamics in time-series data, and swiftly and reliably select the best models from multiple candidates. These methods significantly reduce the need for numerous simulations, thereby vastly accelerating performance compared to existing approaches. 

Aushev’s research offers extensive benefits across various industries and academic fields that rely on computational simulations. By improving model selection and parameter estimation techniques, his work yields deeper insights into complex systems. This has practical applications in diverse areas such as AI, traffic systems, and financial markets. His contributions significantly advance simulation-based inference, providing researchers and professionals with sophisticated tools to more effectively tackle computational challenges. This progress lays the groundwork for quicker development and validation of computational theories in various disciplines. 

The public and media are warmly invited to attend the defense to gain insights into these cutting-edge methods that promise to transform our approach to computational simulations.

Key words: Likelihood-free inference, simulator-based inference, Bayesian optimization

Thesis available for public display 10 days prior to the defence at:

Contact details: 

Email  [email protected]
Mobile  +358 465543621

Doctoral theses in the School of Science:

  • Published:
  • Updated: