Two papers on neuro-AI published by Michmizos Lab won the 2nd Best Paper Awards in two prominent conferences, in Brain Informatics and Robotics.
The first paper presents for the first time a principled theoretical analysis of the practical advantages for enhancing neural networks with astrocytes. It demonstrates how astrocytes can learn memory sequences in associative networks.
The second paper introduces the notion of informing neural networks through neuroscience that results to brain-optimized structures that minimize training. It demonstrates how a spiking neural network emulating the brain's oculomotor system can be used as a bio-inspired algorithm to control an in-house built robotic head.
The presentations and the papers for both papers follow.
1. Sequence Learning in Associative Neuronal-Astrocytic Networks
Authors: Leo Kozachkov, Konstantinos Michmizos
2020 ACM Proceedings of the International Conference on Brain Informatics
Place: 2nd Best Paper Award
Presentation Link: https://youtu.be/4rSNQqvnROc
Paper: https://link.springer.com/chapter/10.1007/978-3-030-59277-6_32
2. A Spiking Neural Network Emulating the Structure of the Oculomotor System Requires No Learning to Control a Biomimetic Robotic Head
Authors: Praveenram Balachandar, Konstantinos Michmizos
2020 IEEE Proceedings of the International Conference on Biomedical Robotics and Biomechatronics (BioRob)
Place: 2nd Best Paper Award
Presentation Link: https://www.youtube.com/watch?v=4dMlFg3HiFw&pbjreload=101
