---
title: "Nottingham Researchers Cut Neural Network Training Updates by 80% Using Mistake Gating"
subtitle: "A biologically inspired training rule skips updates on correctly classified samples, cutting synaptic writes by 50 to 80 percent with no extra hyperparameters."
url: "https://alphasignal.ai/news/nottingham-researchers-cut-neural-network-training-updates-by-80-using-mistake"
author: "AlphaSignal Newsroom"
author_url: "https://alphasignal.ai/editorial-team"
company: "AlphaSignal"
company_url: "https://alphasignal.ai/company/alphasignal"
published: "2026-04-15T00:00:00.000Z"
updated: "2026-08-23T20:46:08.303Z"
type: "Paper"
topics:
  - "Post Training"
  - "Training Infra"
subtopics:
  - "Fine Tuning"
  - "Distributed Training"
  - "Scaling Laws"
upvotes: 1955
source_url: "https://arxiv.org/abs/2604.14336"
comments: 0
truncated: true
---

# Nottingham Researchers Cut Neural Network Training Updates by 80% Using Mistake Gating

> A biologically inspired training rule skips updates on correctly classified samples, cutting synaptic writes by 50 to 80 percent with no extra hyperparameters.

Every time a neural network sees a training sample, it updates its weights, even when the prediction was already correct. Brains do not work that way. A pair of researchers at the University of Nottingham argue this behavior is biologically odd and wasteful, and they have released a training rule that ignores samples the network already gets right, then adds a small twist to keep generalization intact.

Their [paper](https://arxiv.org/abs/2604.14336), titled Mistake gating leads to energy and memory efficient continual learning, proposes what the authors call memorized mistake-gated learning. The rule gates synaptic updates strictly by current and past classification errors, cutting the number of updates by 50 to 80 percent. Code is on [GitHub](https://github.com/vanrossumlab/mistake-gating26).

## The cost of updating on everything

Standard backprop uses a differentiable surrogate loss like cross-entropy, so the gradient is almost never exactly zero, even on samples the network classifies correctly. Weights get nudged on every example forever. Biological learning is metabolically costly: flies subjected to aversive conditioning and then starved died 20 percent sooner than controls, and flies doubled sucrose consumption after long-term memory formation. Evolution seems to have pushed brains toward learning frugally.

Prior ML work tried to address this with core-sets, importance sampling, and example forgetting scores, but finding the core-set is hard and can require a separate model trained beforehand to identify them. None of these approaches work cleanly in an online, continual setting where you see each sample once and cannot pre-analyze the dataset.

_Preview only. Full text requires an AlphaSignal Pro account._

## Sources

- [Mistake gating leads to energy and memory efficient continual learning](https://arxiv.org/abs/2604.14336)
- [mistake-gating GitHub repository](https://github.com/vanrossumlab/mistake-gating26)
- [HTML version of the paper (arXiv)](https://arxiv.org/html/2604.14336)