A State-of-the-Art Review of Deep Reinforcement Learning Techniques for Real-Time Strategy Games

Ashraf, Nesma M.; Mostafa, Reham R.; Sakr, Rasha H.; Rashad, M. Z.

doi:10.1007/978-3-030-72080-3_17

Cited by 4 publications

(3 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The DQN agent combines deep neural networks (DNNs) and Q-learning, i.e. the action-value function Q ( s, a ) is represented and approximated by a DNN (Ashraf et al, 2021). The update scheme of Q realizes one iteration of the Bellman equation, i.e.…”

Section: Methodsmentioning

confidence: 99%

“…Modern machine learning techniques, are now able to solve behavioral tasks that are comparable in complexity to real-world behavior and to reach human performance in some specific domains (Ashraf et al, 2021). These techniques have been used to model hippocampus functionality on an abstract level (Diekmann and Cheng, 2023) and to explain the emergence of complex task-relevant spatial codes (Vijayabaskaran and Cheng, 2022).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Global remapping emerges as the mechanism for renewal of context-dependent behavior in a reinforcement learning model

Kappel,

Cheng

2023

Preprint

View full text Add to dashboard Cite

The hippocampal formation exhibits complex and context-dependent activity patterns and dynamics, e.g., place cell activity during spatial navigation in rodents or remapping of place fields when the animal switches between contexts. Furthermore, rodents show context-dependent renewal of extinguished behavior. However, the link between context-dependent neural codes and context-dependent renewal is not fully understood. We use a reinforcement learning agent based on deep neural networks to study the learning dynamics that occur during spatial learning and context switching in a simulated ABA extinction and renewal paradigm in a 3D virtual environment. Despite its simplicity, the network exhibits a number of features typically found in the CA1 and CA3 regions of the hippocampus. A significant proportion of neurons in deeper layers of the network are tuned to a specific spatial position of the agent in the environment - similar to place cells in the hippocampus. These spatial representations exhibit global remapping when the agent is exposed to a new context. The spatial maps are restored when the agent returns to the previous context, accompanied by renewal of the conditioned behavior. Remapping is facilitated by memory replay of experiences during training. These complex spatial representations and dynamics occur spontaneously in the hidden layer of a deep network during learning. Our results show that integrated codes that jointly represent spatial and task-relevant contextual variables are the mechanism underlying renewal in a simulated DQN agent.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Global remapping emerges as the mechanism for renewal of context-dependent behavior in a reinforcement learning model

Kappel,

Cheng

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Classifier. CNN is one of the most common DNN models used in image and video recognition [58], [59], recommender systems [60], image classification [61], natural language processing [62], and particularly SSVEP-based BCI [41], [49]- [51]. In the current study, CNN is applied as a robust classifier to the input features (explained in the feature extraction section) for recognizing the targets.…”

Section: B the Proposed Architecture For Single-flicker Detectionmentioning

confidence: 99%

A Novel Fast ICA-FBCCA Algorithm and Convolutional Neural Network for Single-Flicker SSVEP-Based BCIs

Aghili,

Kilani,

Rouhani

et al. 2024

IEEE Access

View full text Add to dashboard Cite

Brain-computer interface (BCI) systems have been developed to assist individuals with neuromuscular disorders to communicate with their surroundings using their brain signals. One attractive branch of BCI is steady-state visual evoked potential (SSVEP), which has acceptable speed and accuracy and is non-invasive. However, SSVEP-based EEG signals suffer from eye-fatigue problems, resulting in artifacts that affect the accuracy of the system. Thus, researchers are still working to improve SSVEP-based BCI systems. This paper proposes robust machine-learning algorithm for single-flicker SSVEP detection. A novel approach based on fast independent component analysis and filter-bank canonical correlation analysis (fast ICA-FBCCA) is developed to extract features from the single-flicker SSVEP signal. The clean features learned by fast ICA-FBCCA are then applied to a discrete wavelet transform (DWT) technique and fed to a convolutional neural network (CNN) with only one convolutional layer and a smaller number of parameters. The effectiveness of the proposed technique is evaluated using two datasets. The results were evaluated using two datasets. The findings clearly demonstrate that the proposed method outperforms traditional methods, with average target recognition accuracy and standard deviation values of 97 ± 3.1% among 6 subjects for dataset 1 and 82.12 ± 10.7% among 12 subjects for dataset 2. Overall, these findings suggest that the proposed method is a promising approach for improving the accuracy and reliability of the single-flicker SSVEP-based BCI systems. INDEX TERMSbrain-computer interface (BCI); single-flicker steady-state visual evoked potential; fast independent component analysis (fast ICA); filter-bank canonical correlation analysis (FBCCA); convolutional neural network (CNN) I. INTRODUCTION Brain-computer interface (BCI) is a communication device that allows individuals to control computers or other devices using their brain signals, thereby enabling communication or control without the need for muscle movement [1], [2]. The primary objective of research and advancement in the field of BCIs is to develop an innovative communication tool that can be used specifically for people who have experienced stroke, neurological disorders, muscle impairment, or spinal cord injuries. These individuals face significant challenges in effectively engaging with their environment using conventional communication methods, necessitating the development of alternative means to facilitate their communication capabilities. BCIs enable individuals to issue commands directly to a computer, robot, or smart prosthesis through their brain signals. The important application of BCI performed by machine learning (ML) is device control such as wheelchairs [3], robots [4], car [5], cursor [6], and spellers [7]. Various techniques are used to measure brain activity in BCIs, predominantly relying on the acquisition of electrical signals via invasive or non-invasive approaches. Invasive methods involve the insertion of

show abstract

Deep reinforcement learning in real-time strategy games: a systematic literature review

Barros e Sá,

Madeira

2024

Appl Intell

View full text Add to dashboard Cite

A State-of-the-Art Review of Deep Reinforcement Learning Techniques for Real-Time Strategy Games

Cited by 4 publications

References 26 publications

Global remapping emerges as the mechanism for renewal of context-dependent behavior in a reinforcement learning model

Global remapping emerges as the mechanism for renewal of context-dependent behavior in a reinforcement learning model

A Novel Fast ICA-FBCCA Algorithm and Convolutional Neural Network for Single-Flicker SSVEP-Based BCIs

Deep reinforcement learning in real-time strategy games: a systematic literature review

Contact Info

Product

Resources

About