CMOS Implementation of Spiking Equilibrium Propagation for Real-Time Learning

Taylor, Brady; Ramos, Nicky; Yeats, Eric; Li, Hai

doi:10.1109/aicas54282.2022.9869989

Cited by 3 publications

(2 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Unlike the backpropagation algorithm, the EP algorithm with Hopfield energy can be trained using the local learning rule, where a weight update can be calculated using information from only neighboring neurons. The hardware for local learning can be implemented using much simpler analog circuits, which can be regarded as being more similar to the real operation of the human brain [ 21 , 22 , 23 , 32 ]. Figure 1 a presents a flowchart of the original EP algorithm [ 28 ].…”

Section: Methodsmentioning

confidence: 99%

“…Furthermore, backpropagation is a nonlocal learning algorithm that requires a significant amount of buffer memory to store all the neuronal and synaptic information from an entire network [ 16 , 17 , 18 , 19 , 20 ]. Alternatively, brain-mimicking learning algorithms, such as spike-timing-dependent plasticity (STDP), can be considered [ 21 , 22 , 23 , 24 , 25 ]. Although STDP requires much simpler hardware than backpropagation, the performance of DNNs trained by STDP is still unsatisfactory compared to backpropagation learning [ 24 , 26 , 27 ].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Memristor Crossbar Circuits Implementing Equilibrium Propagation for On-Device Learning

Cho

et al. 2023

Micromachines

View full text Add to dashboard Cite

Equilibrium propagation (EP) has been proposed recently as a new neural network training algorithm based on a local learning concept, where only local information is used to calculate the weight update of the neural network. Despite the advantages of local learning, numerical iteration for solving the EP dynamic equations makes the EP algorithm less practical for realizing edge intelligence hardware. Some analog circuits have been suggested to solve the EP dynamic equations physically, not numerically, using the original EP algorithm. However, there are still a few problems in terms of circuit implementation: for example, the need for storing the free-phase solution and the lack of essential peripheral circuits for calculating and updating synaptic weights. Therefore, in this paper, a new analog circuit technique is proposed to realize the EP algorithm in practical and implementable hardware. This work has two major contributions in achieving this objective. First, the free-phase and nudge-phase solutions are calculated by the proposed analog circuits simultaneously, not at different times. With this process, analog voltage memories or digital memories with converting circuits between digital and analog domains for storing the free-phase solution temporarily can be eliminated in the proposed EP circuit. Second, a simple EP learning rule relying on a fixed amount of conductance change per programming pulse is newly proposed and implemented in peripheral circuits. The modified EP learning rule can make the weight update circuit practical and implementable without requiring the use of a complicated program verification scheme. The proposed memristor conductance update circuit is simulated and verified for training synaptic weights on memristor crossbars. The simulation results showed that the proposed EP circuit could be used for realizing on-device learning in edge intelligence hardware.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Memristor Crossbar Circuits Implementing Equilibrium Propagation for On-Device Learning

Cho

et al. 2023

Micromachines

View full text Add to dashboard Cite

show abstract

A self-learning magnetic Hopfield neural network with intrinsic gradient descent adaption

Niu,

Zhang,

et al. 2024

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

Physical neural networks (PNN) using physical materials and devices to mimic synapses and neurons offer an energy-efficient way to implement artificial neural networks. Yet, training PNN is difficult and heavily relies on external computing resources. An emerging concept to solve this issue is called physical self-learning that uses intrinsic physical parameters as trainable weights. Under external inputs (i.e., training data), training is achieved by the natural evolution of physical parameters that intrinsically adapt modern learning rules via an autonomous physical process, eliminating the requirements on external computation resources. Here, we demonstrate a real spintronic system that mimics Hopfield neural networks (HNN), and unsupervised learning is intrinsically performed via the evolution of the physical process. Using magnetic texture–defined conductance matrix as trainable weights, we illustrate that under external voltage inputs, the conductance matrix naturally evolves and adapts Oja’s learning algorithm in a gradient descent manner. The self-learning HNN is scalable and can achieve associative memories on patterns with high similarities. The fast spin dynamics and reconfigurability of magnetic textures offer an advantageous platform toward efficient autonomous training directly in materials.

show abstract

Weight Update Scheme for 1T1R Memristor Array Based Equilibrium Propagation

Taylor,

Yang,

2024

2024 IEEE 6th International Conference on AI Circuits and Systems (AICAS)

View full text Add to dashboard Cite

CMOS Implementation of Spiking Equilibrium Propagation for Real-Time Learning

Cited by 3 publications

References 13 publications

Memristor Crossbar Circuits Implementing Equilibrium Propagation for On-Device Learning

Memristor Crossbar Circuits Implementing Equilibrium Propagation for On-Device Learning

A self-learning magnetic Hopfield neural network with intrinsic gradient descent adaption

Weight Update Scheme for 1T1R Memristor Array Based Equilibrium Propagation

Contact Info

Product

Resources

About