István Gellért Knáb

Mr. István Gellért Knáb


Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
25 May 2024 in Electronics
Verified
Journal Article
Open access
PDF file
BKBálint Kővári
Bálint Kővári
Bálint PelenczeiBálint Pelenczei
Bálint Pelenczei
István Gellért KnábIstván Gellért Knáb
István Gellért Knáb
Tamás BécsiTamás Bécsi
Tamás Bécsi
In recent years, Reinforcement Learning (RL) has excelled in the realm of autonomous vehicle control, which is distinguished by the absence of limitations, such as specific training data or the necessity for explicit mathematical model identification. Particularly in the context of lane keeping, a diverse set of rewarding strategies yields a spectrum of realizable policies. Nevertheless, the challenge lies in discerning the optimal behavior that maximizes performance. Traditional approaches entail exhaustive training through a trial-and-error strategy across conceivable reward functions, which is a process notorious for its time-consuming nature and substantial financial implications. Contrary to conventional methodologies, the Monte Carlo Tree Search (MCTS) enables the prediction of reward function quality through Monte Carlo simulations, thereby eliminating the need for exhaustive training on all available reward functions. The findings obtained from MCTS simulations can be effectively leveraged to selectively train only the most suitable RL models. This approach helps alleviate the resource-heavy nature of traditional RL processes through altering the training pipeline. This paper validates the theoretical framework concerning the unique property of the Monte Carlo Tree Search algorithm by emphasizing its generality through highlighting crossalgorithmic and crossenvironmental capabilities while also showcasing its potential to reduce training costs.

29 July 2024 in IEEE Access
Verified
Journal Article
Open access
BKBálint Kővári
Bálint Kővári
István Gellért KnábIstván Gellért Knáb
István Gellért Knáb
Domokos Esztergár-KissDomokos Esztergár-Kiss
Domokos Esztergár-Kiss
Szilard AradiSzilard Aradi
Szilard Aradi
Tamás BécsiTamás Bécsi
Tamás Bécsi
With increasing realised traffic on transport networks, greenhouse gas emissions show a similar trend. Reducing them is a modern aspiration, creating a better place to live and moving towards sustainability. Expanding the infrastructure is often not an appropriate solution, as the system would only be fully utilised at peak times, while at less frequent times it would not even approach capacity and would require huge investment costs. An alternative to further construction work is the implementation of intelligent traffic systems, where smoother flows can achieve higher capacity by reducing the variability in the system. In a motorway environment, a common approach is Variable Speed Limit Control, where the road is divided into zones and individual speed limits are used to increase or decrease the load on the cells. This paper proposes a solution in which individual cells make decisions cooperatively, in contrast to classical state machine-based methods. Thanks to the jointly formulated goal of the agents, a predictive control method is created that leads to a reduction in emissions due to avoided shock waves and reduced waiting times. This paper presents a solution that provides a universal solution across multiple application lengths, illustrating the power of deep learning. https://github.com/istvan-knab/Variable-speed-limit-control.

16 October 2025 in IEEE Access
Verified
Journal Article
Open access
István Gellért KnábIstván Gellért Knáb
István Gellért Knáb
Tamás BécsiTamás Bécsi
Tamás Bécsi
BKBálint Kővári
Bálint Kővári
A primary limiting factor in modern deep learning is the availability of computational resources, a constraint that becomes particularly pronounced in the context of complex reinforcement learning tasks. Hardware performance limitations often result in prolonged training procedures, which are frequently correlated with increased operational costs. In many cases, these limitations place such solutions at a disadvantage, which is why cost reduction remains one of the main priorities. In the field of deep reinforcement learning, a commonly employed method for this purpose is the guided selection of training samples, which emphasizes the most informative state transitions, thereby allowing the agent to maximize its learning from the most critical experiences. However, similarly to the decision-making process, it is also crucial to ensure that the agent has information about which samples are important. Therefore, exploration must be incorporated in this context as well, enabling the agent to acquire the most diverse possible knowledge regarding both the samples themselves and their informational value. The paper introduces a sampling method for continuous-output reinforcement learning agents, which balances, in a problem-independent manner, the necessary exploration and the selection of the most informative samples that enhance learning speed. Due to the presented Upper Confidence Bound-based sampling, both exploration and exploitation are ensured simultaneously, and by tuning their balance appropriately, training can be accelerated. This approach improves the efficiency of the training process, which often constitutes the primary bottleneck.