What question did this study set out to answer?

To develop a brain-inspired reinforcement learning system utilizing α-In2Se3 ferroelectric transistors.

February 23, 2026Nature Communications0 citationsOpen Access

Brain-Inspired Architecture for Energy-Efficient Reinforcement Learning in Autonomous Driving

Brain-inspired synaptic transistors for in-situ spiking reinforcement learning with eligibility trace

View Full Paper

Ask AI

Bookmark

Authors

Demonstrates brain-inspired reinforcement learning with an energy-efficient transistor system, indicating advancements in autonomous driving tasks.

To develop a brain-inspired reinforcement learning system utilizing α-In2Se3 ferroelectric transistors.
Implemented a spiking neural network-based reinforcement learning architecture using α-In2Se3 transistors.
Leveraged polarization and relaxation properties of the ferroelectric semiconductor for conductance modulation.
Conducted in-situ updates without external memory during autonomous driving tasks.
Showed that the system achieved in-situ weight updates in reinforcement learning tasks.
Demonstrated improved processing capability through biological eligibility trace decay.
Achieved energy efficiency in the implementation of spiking neural networks for autonomous driving.

Wang et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark

February 23, 2026Nature Communications0 citationsOpen Access

Brain-inspired synaptic transistors for in-situ spiking reinforcement learning with eligibility trace

View Full Paper

Ask AI

Bookmark

Authors

Demonstrates brain-inspired reinforcement learning with an energy-efficient transistor system, indicating advancements in autonomous driving tasks.

To develop a brain-inspired reinforcement learning system utilizing α-In2Se3 ferroelectric transistors.
Implemented a spiking neural network-based reinforcement learning architecture using α-In2Se3 transistors.
Leveraged polarization and relaxation properties of the ferroelectric semiconductor for conductance modulation.
Conducted in-situ updates without external memory during autonomous driving tasks.
Showed that the system achieved in-situ weight updates in reinforcement learning tasks.
Demonstrated improved processing capability through biological eligibility trace decay.
Achieved energy efficiency in the implementation of spiking neural networks for autonomous driving.

Wang et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark