Chip Design and Implementation of High Enerigy-Efficiency Reconfigurable Neuromorphic Computing for Deep Learning Appications

Project Details


The current research on the circuit implementation of deep learning neuralnetwork engines can be divided into two categories: digital circuit implementationand ultra-low-power analog circuit implementation (ex., neuromorphiccomputing). In order to achieve more reduction in energy consumption, somestudies have also proposed the use of emerging technologies to implementDNNs. The literature shows that the energy efficiency of analog neuromorphiccomputing for object recognition applications is about 100,000 times higher thanthat achieved by RISC software. The energy efficiency of analog neuromorphiccomputing is more than a hundred times higher than that of the CMOS digitalcircuit implementation, and the required area is only about 0.16 times that of theCMOS digital circuit implementation.The computer can process more complicated system due to the evolution ofscience and technology in recent years, and it also making the ArtificialIntelligence (AI) get a second wind.As electronic devices continue to shrink in size, the latest advances in computingtechnology have enabled the design of wearable devices to achieve long-termcontinuous monitoring tasks, and have the potential to promote timely medicalmeasures for treatment and care. In the biomedical signal, doctors often useElectrocardiogram (ECG) and Phonocardiogram (PCG) as the reference forjudging heart disease. In addition, with the evolution of science and technology inrecent years, the system can handle a larger number of operations, makingArtificial Intelligence (AI) once again a research hotspot. Applying artificial intelligence to speech processing includes significantly reducing the Word ErrorRate (WER) of speech recognition. It will improve automatic speech recognition(ASR) for more accurate and wide range by the applications through the DeepNeural Network (DNN).The main research direction of this project is to develop high-energy-efficientreconfigurable architecture and chip circuits for neuromorphic computingtechnology. It will focus on the design and specifications of basic neuron-relatedcircuit architecture, neuromorphic computing architecture and modules. Thisproject will propose a neuron cell model, which will form a Dendritic NeuronModel (DNM) architecture and use Particle Swarm Optimization (PSO) toimplement deep learning for ECG/PCG recognition and speech recognitionsystem, using the above-mentioned applications as a test and inspection platformto show the results of research and development.
Effective start/end date1/08/2030/09/21

UN Sustainable Development Goals

In 2015, UN member states agreed to 17 global Sustainable Development Goals (SDGs) to end poverty, protect the planet and ensure prosperity for all. This project contributes towards the following SDG(s):

  • SDG 3 - Good Health and Well-being
  • SDG 7 - Affordable and Clean Energy
  • SDG 8 - Decent Work and Economic Growth
  • SDG 12 - Responsible Consumption and Production
  • SDG 17 - Partnerships for the Goals


  • Neuron Cell Model
  • Dendritic Neuron Model (DNM)
  • Particle Swarm Optimization (PSO)
  • Electrocardiogram (ECG)
  • Phonocardiogram (PCG)
  • Automatic Speech Recognition (ASR)


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.