Faruk Polat

E-mail

polatf@metu.edu.tr

Department

Department of Computer Engineering

ORCID

0000-0003-0509-9153

Scopus Author ID

7003321824

Web of Science Researcher ID

ABA-3585-2020

Population-based exploration in reinforcement learning through repulsive reward shaping using eligibility traces Bal, Melis Ilayda; İyigün, Cem; Polat, Faruk; Aydın, Hüseyin (2024-01-01) Efficient exploration plays a key role in accelerating the learning performance and sample efficiency of reinforcement learning tasks. In this paper we propose a framework that serves as a population-based repulsive reward...
Solving an industry-inspired generalization of lifelong MAPF problem including multiple delivery locations Polat, Faruk (2023-08-01)
Landmark based guidance for reinforcement learning agents under partial observability Demir, Alper; Çilden, Erkin; Polat, Faruk (2022-01-01) Under partial observability, a reinforcement learning agent needs to estimate its true state by solely using its observation semantics. However, this interpretation has a drawback, which is called perceptual aliasing, avoi...
LIMP: Incremental Multi-agent Path Planning with LPA* Yorganci, Mucahit Alkan; Semiz, Fatih; Polat, Faruk (2022-01-01) The multi-agent pathfinding (MAPF) problem is defined as finding conflict-free paths for more than one agent. There exist optimal and suboptimal solvers for MAPF, and most of the solvers focus on the MAPF problem in static...
Multiagent Pickup and Delivery for Capacitated Agents Çilden, Evren; Polat, Faruk (2022-01-01) In Multi-Agent Pickup and Delivery (MAPD), multiple robots continuously receive tasks to pick up packages and deliver them to predefined destinations in an automated warehouse. If the capacity of agents is increased, agent...
Incremental multi-agent path finding Semiz, Fatih; Polat, Faruk (Elsevier BV, 2021-03-01) Existing multi-agent path finding (MAPF) algorithms are offline methods that aim at finding conflict-ree paths for more than one agent. In many real-life applications it is possible that a multi-agent plan cannot be fully ...
Compact Frequency Memory for Reinforcement Learning with Hidden States. Polat, Faruk; Cilden, Erkin (2019-10-28) Memory-based reinforcement learning approaches keep track of past experiences of the agent in environments with hidden states. This may require extensive use of memory that limits the practice of these methods in a real-li...
Effective feature reduction for link prediction in location-based social networks Bayrak, Ahmet Engin; Polat, Faruk (SAGE Publications, 2019-10-01) In this study, we investigated feature-based approaches for improving the link prediction performance for location-based social networks (LBSNs) and analysed their performances. We developed new features based on time, com...
Automatic landmark discovery for learning agents under partial observability DEMİR, ALPER; Cilden, Erkin; Polat, Faruk (Cambridge University Press (CUP), 2019-08-02) In the reinforcement learning context, a landmark is a compact information which uniquely couples a state, for problems with hidden states. Landmarks are shown to support finding good memoryless policies for Partially Obse...
GENERATING EFFECTIVE INITIATION SETS FOR SUBGOAL-DRIVEN OPTIONS DEMİR, ALPER; Cilden, Erkin; Polat, Faruk (World Scientific Pub Co Pte Lt, 2019-03-01) Options framework is one of the prominent models serving as a basis to improve learning speed by means of temporal abstractions. An option is mainly composed of three elements: initiation set, option's local policy and ter...

Citation Formats