通过双输入突触晶体管中的信号总和过程来实现强化学习的物理实现:Au的光诱导双极逆转 (I) 复杂的cPVP的电荷陷
Dong Gue Roe1,2, Sungjoon Cheon1, Byung Hak Jhun1
1Department of Chemical and Biomolecular Engineering, Yonsei University, Seoul, Republic of Korea.
Advanced materials (Deerfield Beach, Fla.)
|January 6, 2026
概括
研究人员开发了一种新的双输入突触晶体管,使用光和电压来增强人工智能 (AI) 硬件. 这一创新减少了人工智能算法的计算复杂性,通过实现高效的设备级总和操作.
科学领域:
- 材料科学 材料科学 材料科学
- 计算机工程 计算机工程
- 人工智能的人工智能
背景情况:
- 人工智能 (AI) 的进步需要硬件创新,超越传统的晶体管.
- 当前的人工智能加速器在扩展和热效率方面存在局限性.
- 交联晶体管为模拟和并行计算提供了潜力,但需要进一步优化.
研究的目的:
- 开发一种新的双输入突触晶体管,以提高人工智能硬件中的设备级计算效率.
- 探索光和电压对突触重量调节的协同效应.
- 通过集成的总和操作来降低AI算法中的计算复杂性.
主要方法:
- 一个双输入突触晶体管的制造,使用线性双坐标Au (I) 复合体 (Au (DippPZI) (DPA)) 和cPVP层.
- 复合体的光刺激,以诱导双极逆转,并使光诱导的突触重量调制成为可能.
- 应用电压来利用 ─OH 陷站点进行电压驱动的调制.
- 同时应用光和电压,以实现协同模拟总和突触电流.
主要成果:
- 通过联结体对联结体电荷转移诱导的双极逆转,证明了光诱导的突触重量调制.
- 在cPVP层中通过丰富的 ─OH 陷位实现了电压驱动的调制.
- 成功集成光和电压输入,在单一设备内协同生成突触电流的模拟总和.
- 展示了晶体管在没有复杂的外围单元的情况下在强化学习算法中执行重量总和的能力.
结论:
- 双输入突触晶体管为AI硬件提供了增强的设备级计算效率.
- 这种方法通过将总和操作直接集成到晶体管中来降低计算成本.
- 开发的晶体管为各种依赖于总和的AI学习算法提供了一个多功能平台,为高效的硬件级AI计算铺平了道路.
更多相关视频
09:30Patterned Photostimulation with Digital Micromirror Devices to Investigate Dendritic Integration Across Branch Points
Published on: March 2, 2011
16.2K
10:29Multi-photon Intracellular Sodium Imaging Combined with UV-mediated Focal Uncaging of Glutamate in CA1 Pyramidal Neurons
Published on: October 8, 2014
14.5K
相关概念视频
Integration of Synaptic Events
3.5K
Synaptic integration mainly includes the summation of graded potentials. Graded potentials, regardless of their type, cause subtle alterations in membrane voltage, resulting in either depolarization or hyperpolarization. These incremental changes, when combined or summed, can propel the neuron toward its threshold. Consider, for example, a membrane experiencing a +15 mV shift, causing it to depolarize from -70 mV to -55 mV. In this scenario, graded potentials govern the membrane's ability to...
3.5K
Long-term Potentiation
3.4K
Long-term potentiation, or LTP, is one of the ways by which synaptic plasticity—changes in the strength of chemical synapses—can occur in the brain. LTP is the process of synaptic strengthening that occurs over time between pre and postsynaptic neuronal connections. The synaptic strengthening of LTP works in opposition to the synaptic weakening of long-term depression (LTD) and together are the main mechanisms that underlie learning and memory.
Hebbian LTP
LTP can occur when...
Hebbian LTP
LTP can occur when...
3.4K
Long-term Potentiation
58.2K
Long-term potentiation, or LTP, is one of the ways by which synaptic plasticity—changes in the strength of chemical synapses—can occur in the brain. LTP is the process of synaptic strengthening that occurs over time between pre- and postsynaptic neuronal connections. The synaptic strengthening of LTP works in opposition to the synaptic weakening of long-term depression (LTD) and together are the main mechanisms that underlie learning and memory.
58.2K
Synaptic Signaling
6.5K
Neurons communicate at synapses, or junctions, to excite or inhibit the activity of other neurons or target cells, such as muscles. Synapses may be chemical or electrical.
Most synapses are chemical, meaning an electrical impulse or action potential spurs the release of chemical messengers called neurotransmitters. The neuron sending the signal is called the presynaptic neuron, and the neuron receiving the signal is the postsynaptic neuron.
The presynaptic neuron fires an action potential that...
Most synapses are chemical, meaning an electrical impulse or action potential spurs the release of chemical messengers called neurotransmitters. The neuron sending the signal is called the presynaptic neuron, and the neuron receiving the signal is the postsynaptic neuron.
The presynaptic neuron fires an action potential that...
6.5K
Synaptic Signaling
79.1K
Neurons communicate at synapses, or junctions, to excite or inhibit the activity of other neurons or target cells, such as muscles. Synapses may be chemical or electrical.
79.1K
The Role of Ion Channels in Neuronal Computation
3.6K
A postsynaptic neuron usually receives numerous impulses from several other presynaptic neurons. The axon hillock of the postsynaptic neuron integrates all these signals and determines the likelihood of firing an action potential.
Sometimes a single EPSP is strong enough to induce an action potential in the postsynaptic neuron. However, multiple presynaptic inputs must often create EPSPs around the same time for the postsynaptic neuron to be sufficiently depolarized to fire an action potential....
Sometimes a single EPSP is strong enough to induce an action potential in the postsynaptic neuron. However, multiple presynaptic inputs must often create EPSPs around the same time for the postsynaptic neuron to be sufficiently depolarized to fire an action potential....
3.6K
