Advances in Neuromorphic Computing Devices: Insights on Both Conventional and Unconventional Architectures

Area of Science:

• Computer Science
• Electrical Engineering
• Neuroscience

Background:

• Neuromorphic circuits and devices aim to mimic biological neurons for efficient computation.
• Digital computing faces energy and time bottlenecks, driving interest in alternative paradigms.
• The term "neuromorphic" is broadly applied, leading to potential misunderstandings across technologies.

Purpose of the Study:

• To provide an overview of diverse technological strategies for developing neuromorphic computing systems.
• To analyze the specific neuromorphic features implemented by different technologies.
• To focus on technological strategies and patents utilizing unconventional computing paradigms.

Main Methods:

• Literature review of neuromorphic computing strategies.
• Analysis of technological approaches and their implemented features.
• Examination of patents related to unconventional neuromorphic computing.

Main Results:

• Identification of various technological strategies in neuromorphic computing.
• Categorization of implemented neuromorphic features across different systems.
• Highlighting of key patents in unconventional neuromorphic paradigms.

Conclusions:

• Neuromorphic computing encompasses a wide range of technologies, often misunderstood.
• Understanding diverse strategies and features is crucial for advancing the field.
• Unconventional computing paradigms show significant patent activity and potential.