Ing. Marko Ruman

Ph.D student

Marko Ruman
Research interests: Reinforcement learning, transfer learning, aplications of machine learning.

Biography

Publication list

Education 

Present Jun 2018 Ph.D. - Mathematical Engineering State Doctoral Exam with honors Faculty of Nuclear Sciences and Physical Engineering, Czech Technical University, Prague, Czech Republic 

Jun 2018 Sep 2016 MSc. - Mathematical Engineering GPA: 3.97 Magna cum laude Faculty of Nuclear Sciences and Physical Engineering, Czech Technical University, Prague, Czech Republic Aug 

2016 Sep 2013 BSc. - Mathematical Engineering GPA: 3.55 Magna cum laude Faculty of Nuclear Sciences and Physical Engineering, Czech Technical University, Prague, Czech Republic 

Work Experience 

Present Jul 2015 Research Assistant - Institute of Information Theory and Automation, Czech Academy of Sciences 

ˆ Co-authored 7 papers on reinforcement learning, learning of dynamic stochastic environments and modelling human decision making as Markov decision processes 

ˆ Presented own research at conferences NIPS, ICANN and ICCAIRO 

ˆ Developed a Python library implementing the researched universally approximating model of dynamic environments called Mixture Ratios 

ˆ Recently developed a novel method for knowledge transfer in deep reinforcement learning 

ˆ Participated as a team member of two international and four Czech research projects 

ˆ Served as a member of the Organising Committee of COST 2019 GAMENET Conference Research interests 

ˆ Knowledge transfer in deep reinforcement learning ˆ Bayesian learning of dynamic stochastic environments 

ˆ Markov decision processes 

Jan 2020 Apr 2018 Co-Founder - Optimifica 

ˆ Developed a custom ML model for predicting customers’ interest in specific products used by retail companies 

ˆ Achieved 40% success rate when recommending 5 products for a new basket for retail companies with low-frequency customer base (2-3 visits per year on average) 

ˆ Increased margin of retail companies by 5% by utilising micro-targeted product recommendations produced by the custom ML model Apr 2018 Oct 2004 ML and web development - Freelancer 

ˆ Coded my first website when I was 12, worked on many machine learning and web development projects since then

Journal articles

  1. Ruman Marko, Guy Tatiana Valentine: Knowledge Transfer in Deep Reinforcement Learning via an RL-Specific GAN-Based Correspondence Function, IEEE Access 12 1 (2024), p. 177204-177218 Download Download DOI: 10.1109/ACCESS.2024.3497589 [2024]
  2. Kárný Miroslav, Ruman Marko: Mixture ratio modeling of dynamic systems, International Journal of Adaptive Control and Signal Processing 35 5 (2021), p. 660-675 Download Download DOI: 10.1002/acs.3219 [2021]

Other publications

  1. Ruman Marko, Kárný Miroslav: Dynamic Mixture Ratio Model, Proceedings of the 2019 International Conference on Control, Artificial Intelligence, Robotics & Optimization (ICCAIRO), p. 92-99, The Institute of Electrical and Electronics Engineers, Inc. (2020) Download DOI: 10.1109/ICCAIRO47923.2019.00023 [2020]
  2. Guy Tatiana Valentine, Ruman Marko, Hůla František, Kárný Miroslav: Experimental Performance of Deliberation-Aware Responder in Multi-Proposer Ultimatum Game, Proceedings of the NIPS 2016 Workshop on Imperfect Decision Makers, p. 51-60, Eds: Guy Tatiana Valentine, Kárný Miroslav, Rios-Insua D., Wolpert D. H. Download [2017]
  3. Ruman Marko, Hůla František, Kárný Miroslav, Guy Tatiana Valentine: Deliberation-aware Responder in Multi-Proposer Ultimatum Game, Artificial Neural Networks and Machine Learning – ICANN 2016, p. 230-237 Download DOI: 10.1007/978-3-319-44778-0_27 [2016]
  4. Hůla František, Ruman Marko, Kárný Miroslav: Adaptive Proposer for Ultimatum Game, Artificial Neural Networks and Machine Learning – ICANN 2016, p. 330-338 Download DOI: 10.1007/978-3-319-44778-0_39 [2016]