缩放尖峰驱动变压器与高效尖峰发射近似训练
概括
尖端神经网络 (SNN) 现在与人工神经网络 (ANN) 的性能相匹配,使用一种新的尖端射击近似 (SFA) 方法. 这种方法提高了培训效率,并减少了低功耗AI应用的功耗.
科学领域:
- 人工智能的人工智能
- 神经形态计算是一种神经形态计算.
- 计算机视觉 计算机视觉
背景情况:
- 尖端神经网络 (SNN) 为人工神经网络 (ANN) 提供了一种低功耗的替代方案.
- 在弥合绩效差距和降低SNNs的培训成本方面仍然存在挑战.
- 尖端神经元中的二元触发机制可能导致性能限制.
研究的目的:
- 解决SNNs的绩效差距和高的培训成本.
- 为了优化尖端神经元的发射模式,以提高效率和性能.
- 为了使SNN能够作为通用视觉骨干.
主要方法:
- 提出了一种使用整数训练和尖端驱动推理的尖端射击近似方法 (SFA).
- 开发了一个高效的尖峰驱动的变压器架构.
- 引入了一种尖峰掩盖的自动编码器,以减轻SNN扩展期间的性能退化.
主要成果:
- 在ImageNet-1k上实现了最先进的top-1精度,在173M参数模型上达到86.2%.
- 在ImageNet.Net上,10M参数SNN模型的性能比现有的SNN高出7.2%.
- 演示了4.5倍的训练时间加速和3.9倍的推断能效改进.
- 在对象检测,语义细分和神经形态视觉任务中验证了有效性.
结论:
- 该SFA方法使SNN能够实现与ANN的竞争性表现.
- 现在SNN可以保持其低功耗优势,同时与ANN性能相匹配.
- 这项工作代表了SNNs作为一般视觉脊柱的重大进展.
相关概念视频
Energy Losses in Transformers
819
In an ideal transformer, it is assumed that there are no energy losses, and, hence, all the power at the primary winding is transferred to the secondary winding. However, in reality, the transformers always have some energy losses, and, hence, the output power obtained at the secondary winding is less than the input power at the primary winding due to energy losses.
There are four main reasons for energy losses in transformers.
The first cause can be the high resistance of the...
There are four main reasons for energy losses in transformers.
The first cause can be the high resistance of the...
819
Transformers with Off-Nominal Turns Ratios
129
In scenarios involving parallel transformers with disparate ratings, developing per-unit models requires accommodating off-nominal turns ratios. This situation arises when the selected base voltages are not proportional to the transformer’s voltage ratings. Consider a transformer where the rated voltages are related by the term a. If the chosen voltage bases satisfy a relationship involving term b, term c is defined as the ratio of these bases. This ratio is then substituted into the...
129
Reducing Line Loss
141
In a three-phase circuit, line loss is an indicator of energy dissipated as heat due to the resistance of transmission lines. To address this, incorporating transformers into the system—a step-up transformer at the source and a step-down transformer at the load—is a strategic solution. Two three-phase transformers are introduced to improve this.
With a step-up transformer at the source, the voltage is increased, thereby reducing the current in the transmission lines since power loss...
With a step-up transformer at the source, the voltage is increased, thereby reducing the current in the transmission lines since power loss...
141
Equivalent Circuits for Practical Transformers
377
The practical equivalent circuits of single-phase two-winding transformers exhibit significant deviations from their idealized versions due to the inherent properties of winding resistance and finite core permeability. These properties result in real and reactive power losses, affecting the transformer's performance. Understanding these deviations is crucial for designing more efficient transformers.
In a practical transformer, each winding exhibits resistance and leakage reactance. The...
In a practical transformer, each winding exhibits resistance and leakage reactance. The...
377
Power System Three-Phase Short Circuits
71
Determining the subtransient fault current in a power system involves representing transformers by their leakage reactances, transmission lines by their equivalent series reactances, and synchronous machines as constant voltage sources behind their subtransient reactances. In this analysis, certain elements are excluded, such as winding resistances, series resistances, shunt admittances, delta-Y phase shifts, armature resistance, saturation, saliency, non-rotating impedance loads, and small...
71
Simplified Synchronous Machine Model
173
The Synchronous Machine Model is a fundamental tool in analyzing and ensuring the transient stability of power systems. This model simplifies the representation of a synchronous machine under balanced three-phase positive-sequence conditions, assuming constant excitation and ignoring losses and saturation. The model is pivotal for understanding the behavior of synchronous generators connected to a power grid, particularly during transient events.
In this model, each generator is connected to a...
In this model, each generator is connected to a...
173


