Sign in Subscribe
Artificial neural networks

Breakthrough in Neuromorphic Computing: Tsinghua University Researchers Develop Bio-Realistic and Versatile Artificial Dendrites

Strategic Insights

— 1 min read

Neuromorphic computing, a rapidly emerging field that draws inspiration from the brain, has made significant strides in recent years. Researchers at Tsinghua University, led by Yifei Yang, have made a groundbreaking discovery that challenges traditional understanding of neural networks. According to a recent study published in the journal Neuromorphic Computing and Engineering, the researchers developed a bio-realistic and versatile artificial dendrite made of WSe2/MoS2 heterojunction-channel anti-ambipolar (AAB) transistor.

Key Takeaways:

  • The study demonstrates a breakthrough in neuromorphic computing by developing a bio-realistic and versatile artificial dendrite that can closely mimic the experimentally recorded non-monotonic dendritic Ca2+ action potential.
  • The artificial dendrite, made of WSe2/MoS2 heterojunction-channel AAB transistor, enhances the robustness and representational capability of spiking neural networks in non-stationary environments.
  • The study challenges traditional understanding of neural networks, where downplayed dendrites were found to be active, dynamically unique, and computationally powerful.
  • The invention of dendritic AAB transistor is a significant advance in diversifying the functionality of transistors.
  • The research has implications for developing more sophisticated and efficient neuromorphic computing systems.
  • The study's findings are supported by computational simulations that show the incorporation of the artificial dendrite enhances the robustness and representational capability of the network.
  • The research team, led by Yifei Yang, includes Mingkun Xu, Tianhang Zhang, Chenfei Miao, Juan Lyu, Jing Pei, Peng Li, Guoqi Li, Si Wu, and Huanglong Li.

Statistics:

  • 2025: The year the research was published in the journal Neuromorphic Computing and Engineering.
  • 5 (2):024020: The volume and page numbers of the journal article where the research was published.
  • 10.1088/2634-4386/ade428: The DOI number of the journal article, accessible through https://doi-org.sdpl.idm.oclc.org/.
  • 2025, NewsRx LLC: The year and publisher of the news report announcing the research breakthrough.
  • 7481: The page number of the news report, as published in Life Science Weekly.

Sources:

  • Bio-realistic and versatile artificial dendrites made of anti-ambipolar transistors. Neuromorphic Computing and Engineering, 2025, 5(2):024020.
  • NewsRx. Tsinghua University Researchers Have Provided New Study Findings on Neuromorphic Computing and Engineering (Bio-realistic and versatile artificial dendrites made of anti-ambipolar transistors). Life Science Weekly. July 8, 2025; p 7481.

Read more