Paper Info

Field Content
Title The Forward-Forward Algorithm: Some Preliminary Investigations
Authors Geoffrey Hinton
Venue Google Brain
Year 2022
Link arxiv

One-line Summary

Spiking Neural Network의 미분불가능성으로 인한 학습의 한계를 Surrogate Gradient Learning을 사용해서 극복했다.


Problem Statement

Challenges in Training SNNs

Objective of the Paper

  • hidden layer를 가진 SNN 훈련의 어려움을 논의하고, 이를 해결하기 위한 다양한 전략을 소개하려 함


Key Idea

Understanding SNNs as RNNs

Biological neuron

Methods

Overall Architecture

(figure description or simple diagram)

Key Equations

\[여기에 수식\]

Algorithm Summary

1. 2. 3.

Results

  • Dataset:
  • Baseline:
  • Key Results:

Thoughts / Questions

  • What’s interesting:
  • What I don’t understand yet:
  • Follow-up papers to read:

References

-