Ali A. Minai
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Ali A. Minai has a total of 150 co-authors. Here are the top co-authors by number of shared publications:
Laxmi R. Iyer
Agency for Science, Technology and Research (A*STAR)
10 shared publicationsSimona Doboli
Hofstra University
10 shared publications- Marios M. Polycarpou
University of Cyprus
6 shared publications S. Doboli
Dept. of Comput. Sci., Hofstra Univ., Hempstead, NY, USA
6 shared publicationsKiran V. Byadarhaly
School of Electronic and Computing Systems, University of Cincinnati, Cincinnati OH 45221
5 shared publicationsM.M. Polycarpou
Dept. Electron. Comput. & Eng. Comput. Sci., Cincinnati Univ., OH, USA
5 shared publications
Latest Publications
- Open Access
- 23 July 2024
- Published by Springer Nature in Cognitive Computation
- Vol. 16 (5) , 1-3
Editorial: What AI and Neuroscience Can Learn from Each Other—Open Problems in Models and Theories
- 23 July 2024
- Published by Springer Nature in Cognitive Computation
- Vol. 16 (5) , 1-3
Open AccessEditorial: What AI and Neuroscience Can Learn from Each Other—Open Problems in Models and Theories
- 22 September 2023
- Published by Springer Nature
- p. 1-12
A Comparative Study of Sentence Embedding Models for Assessing Semantic Variation
- 22 September 2023
- Published by Springer Nature
- p. 1-12
A Comparative Study of Sentence Embedding Models for Assessing Semantic Variation
- 6 March 2023
- Published by Springer Nature in Cognitive Computation
- Vol. 16 (5) , 2389-2404
Deep Intelligence: What AI Should Learn from Nature’s Imagination
- 6 March 2023
- Published by Springer Nature in Cognitive Computation
- Vol. 16 (5) , 2389-2404
Deep Intelligence: What AI Should Learn from Nature’s Imagination
- 28 February 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive and Developmental Systems
- Vol. 15 (4) , 2108-2121
DeepCPG Policies for Robot Locomotion
Central Pattern Generators (CPGs) form the neural basis of the observed rhythmic behaviors for locomotion in legged animals. The CPG dynamics organized into networks allow the emergence of complex locomotor behaviors. In this work, we take this inspiration for developing walking behaviors in multi-legged robots. We present novel DeepCPG policies that embed CPGs as a layer in a larger neural network and facilitate end-to-end learning of locomotion behaviors in deep reinforcement learning (DRL) setup. We demonstrate the effectiveness of this approach on physics engine-based insectoid robots. We show that, compared to traditional approaches, DeepCPG policies allow sample-efficient end-to-end learning of effective locomotion strategies even in the case of high-dimensional sensor spaces (vision). We scale the DeepCPG policies using a modular robot configuration and multi-agent DRL. Our results suggest that gradual complexification with embedded priors of these policies in a modular fashion could achieve non-trivial sensor and motor integration on a robot platform. These results also indicate the efficacy of bootstrapping more complex intelligent systems from simpler ones based on biological principles. Finally, we present the experimental results for a proof-of-concept insectoid robot system for which DeepCPG learned policies initially using the simulation engine and these were afterwards transferred to real-world robots without any additional fine-tuning.
- 28 February 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive and Developmental Systems
- Vol. 15 (4) , 2108-2121
DeepCPG Policies for Robot Locomotion
Central Pattern Generators (CPGs) form the neural basis of the observed rhythmic behaviors for locomotion in legged animals. The CPG dynamics organized into networks allow the emergence of complex locomotor behaviors. In this work, we take this inspiration for developing walking behaviors in multi-legged robots. We present novel DeepCPG policies that embed CPGs as a layer in a larger neural network and facilitate end-to-end learning of locomotion behaviors in deep reinforcement learning (DRL) setup. We demonstrate the effectiveness of this approach on physics engine-based insectoid robots. We show that, compared to traditional approaches, DeepCPG policies allow sample-efficient end-to-end learning of effective locomotion strategies even in the case of high-dimensional sensor spaces (vision). We scale the DeepCPG policies using a modular robot configuration and multi-agent DRL. Our results suggest that gradual complexification with embedded priors of these policies in a modular fashion could achieve non-trivial sensor and motor integration on a robot platform. These results also indicate the efficacy of bootstrapping more complex intelligent systems from simpler ones based on biological principles. Finally, we present the experimental results for a proof-of-concept insectoid robot system for which DeepCPG learned policies initially using the simulation engine and these were afterwards transferred to real-world robots without any additional fine-tuning.
- Open Access
DeepCPG Policies for Robot Locomotion
Central Pattern Generators (CPGs) form the neural basis of the observed rhythmic behaviors for locomotion in legged animals. The CPG dynamics organized into networks allow the emergence of complex locomotor behaviors. In this work, we take this inspiration for developing walking behaviors in multi-legged robots. We present novel DeepCPG policies that embed CPGs as a layer in a larger neural network and facilitate end-to-end learning of locomotion behaviors in deep reinforcement learning (DRL) setup. We demonstrate the effectiveness of this approach on physics engine-based insectoid robots. We show that, compared to traditional approaches, DeepCPG policies allow sample-efficient end-to-end learning of effective locomotion strategies even in the case of high-dimensional sensor spaces (vision). We scale the DeepCPG policies using a modular robot configuration and multi-agent DRL. Our results suggest that gradual complexification with embedded priors of these policies in a modular fashion could achieve non-trivial sensor and motor integration on a robot platform. These results also indicate the efficacy of bootstrapping more complex intelligent systems from simpler ones based on biological principles. Finally, we present the experimental results for a proof-of-concept insectoid robot system for which DeepCPG learned policies initially using the simulation engine and these were afterwards transferred to real-world robots without any additional fine-tuning.
Open AccessDeepCPG Policies for Robot Locomotion
Central Pattern Generators (CPGs) form the neural basis of the observed rhythmic behaviors for locomotion in legged animals. The CPG dynamics organized into networks allow the emergence of complex locomotor behaviors. In this work, we take this inspiration for developing walking behaviors in multi-legged robots. We present novel DeepCPG policies that embed CPGs as a layer in a larger neural network and facilitate end-to-end learning of locomotion behaviors in deep reinforcement learning (DRL) setup. We demonstrate the effectiveness of this approach on physics engine-based insectoid robots. We show that, compared to traditional approaches, DeepCPG policies allow sample-efficient end-to-end learning of effective locomotion strategies even in the case of high-dimensional sensor spaces (vision). We scale the DeepCPG policies using a modular robot configuration and multi-agent DRL. Our results suggest that gradual complexification with embedded priors of these policies in a modular fashion could achieve non-trivial sensor and motor integration on a robot platform. These results also indicate the efficacy of bootstrapping more complex intelligent systems from simpler ones based on biological principles. Finally, we present the experimental results for a proof-of-concept insectoid robot system for which DeepCPG learned policies initially using the simulation engine and these were afterwards transferred to real-world robots without any additional fine-tuning.
- 19 January 2023
- Published by Elsevier in Neural Networks
- Vol. 161, 116-128
Rapid learning of spatial representations for goal-directed navigation based on a novel model of hippocampal place fields
- 19 January 2023
- Published by Elsevier in Neural Networks
- Vol. 161, 116-128
Rapid learning of spatial representations for goal-directed navigation based on a novel model of hippocampal place fields
- Open Access
Rapid Learning of Spatial Representations for Goal-Directed Navigation Based on a Novel Model of Hippocampal Place Fields
The discovery of place cells and other spatially modulated neurons in the hippocampal complex of rodents has been crucial to elucidating the neural basis of spatial cognition. More recently, the replay of neural sequences encoding previously experienced trajectories has been observed during consummatory behavior potentially with implications for quick memory consolidation and behavioral planning. Several promising models for robotic navigation and reinforcement learning have been proposed based on these and previous findings. Most of these models, however, use carefully engineered neural networks and are tested in simple environments. In this paper, we develop a self-organized model incorporating place cells and replay, and demonstrate its utility for rapid one-shot learning in non-trivial environments with obstacles.
Open AccessRapid Learning of Spatial Representations for Goal-Directed Navigation Based on a Novel Model of Hippocampal Place Fields
The discovery of place cells and other spatially modulated neurons in the hippocampal complex of rodents has been crucial to elucidating the neural basis of spatial cognition. More recently, the replay of neural sequences encoding previously experienced trajectories has been observed during consummatory behavior potentially with implications for quick memory consolidation and behavioral planning. Several promising models for robotic navigation and reinforcement learning have been proposed based on these and previous findings. Most of these models, however, use carefully engineered neural networks and are tested in simple environments. In this paper, we develop a self-organized model incorporating place cells and replay, and demonstrate its utility for rapid one-shot learning in non-trivial environments with obstacles.
- 15 June 2021
- Published by Springer Nature
- p. 243-259
Efficient Self-organized Feature Maps Through Developmental Learning
- 15 June 2021
- Published by Springer Nature
- p. 243-259
Efficient Self-organized Feature Maps Through Developmental Learning
- 24 May 2021
- Published by Springer Nature
- p. 195-227
IDEA—Itinerant Dynamics with Emergent Attractors: A Neural Model for Conceptual Combination
- 24 May 2021
- Published by Springer Nature
- p. 195-227
IDEA—Itinerant Dynamics with Emergent Attractors: A Neural Model for Conceptual Combination
- 24 May 2021
- Published by Springer Nature
- p. 229-264
Emergence and Diffusion of Novel Associative Ideas in a Multi-Agent Network Model of Implicit Learning
- 24 May 2021
- Published by Springer Nature
- p. 229-264
Emergence and Diffusion of Novel Associative Ideas in a Multi-Agent Network Model of Implicit Learning