Think Build Implement Repeat
London, UK +44 7367 067226
WhatsApp FOLLOW f in X
  1. Home
  2. Blog
  3. Label Noise: Training on History That Was Mislabelled
AI & Machine Learning

Label Noise: Training on History That Was Mislabelled

Models learn from historical outcomes, and those outcomes contain mistakes. How to detect label noise, estimate how much you have, and limit the damage.

Updated 2 min readBy SpiderHunts Technologies

Free estimateNo obligation

Get a free estimate

Tell us what you need. A senior engineer reads every enquiry.

Takes under a minute. We never share your details.

  • Free consultation
  • No commitment
  • NDA on request

Prefer to talk? Book a free 30-minute call →

Quick answer — TL;DR

Historical records contain errors, and a model trained on them learns the errors too. Finding confidently wrong predictions on training data is the quickest way to surface mislabelled records, and the review usually improves the process as well as the dataset.

Your history is not ground truth

Machine learning treats historical outcomes as the truth to learn from. In business data they are a record of what someone entered, which is a different thing.

  • A return categorised as 'faulty' because that was the quickest option in the dropdown
  • A lead marked 'lost' when it was actually never followed up
  • A fault code chosen by an engineer in a hurry from a list of forty
  • A customer type set at onboarding and never updated as the business changed

None of these is negligence. They are ordinary artefacts of systems used by busy people, and they place a ceiling on model quality that no algorithm removes.

Finding the wrong ones with the model itself

A useful trick: train a model, then look at the training examples it gets confidently wrong. A model that has learned the general pattern and strongly disagrees with a specific record is often right - the record is mislabelled.

  1. Train using cross-validation so every record gets a prediction from a model that did not see it.
  2. Rank records where the prediction disagrees with the recorded label, strongest disagreement first.
  3. Have a domain expert review the top of that list.
  4. Record how many were genuinely mislabelled - that gives you an estimate of noise across the dataset.
  5. Correct what you can, and feed what you learn back into the process that created the errors.

That review is often the most valuable hour of the project, because it reveals how the recording process actually behaves rather than how it is documented.

How much noise you can tolerate

Random noise scattered evenly across classes is survivable - models are reasonably robust to it, and more data helps. Systematic noise is the dangerous kind, because the model learns the bias faithfully.

Noise typeExampleEffect
Random errorsOccasional mis-keyingDegrades accuracy, usually gracefully
Systematic by categoryOne team always picks 'other'Model reproduces the habit
Systematic by periodCategories changed in 2024Model learns the wrong era
Systematic by outcomeFailures recorded, successes notModel sees a distorted world

Practical mitigations

Where a full clean is impractical, several things help. Use a recent, cleaner window rather than the full history if the process improved. Reweight towards records with corroborating evidence. Where several sources disagree, treat that as a signal rather than picking one arbitrarily.

And be honest in the reporting. A model measured against noisy labels has a noisy accuracy figure, and the true performance may be better or worse than reported. A small, carefully verified gold-standard test set is worth building for exactly this reason.

A model trained on what people clicked will learn what people clicked, not what happened.

FAQ

Frequently asked questions

The questions readers ask us after this guide.

Still have a question?

Ask us directly — a senior engineer will get back to you.

Ask about your project

How do I know how much label noise we have?

Review a sample where the model disagrees confidently, plus a random sample for comparison. The difference between the two gives a rough estimate.

Should mislabelled records be deleted or corrected?

Correct where you can establish the truth; otherwise exclude them from training but keep them, with the reason recorded.

Can a model be trained to ignore noisy labels?

There are techniques that reduce sensitivity, but none recovers information that was never recorded correctly. Fixing the source is better.

Does more data compensate for noisy labels?

For random noise, largely yes. For systematic noise, no - more data reinforces the bias.

Keep reading

More on AI & Machine Learning

Start here

Want machine learning project details from us?

Tell us what you are trying to predict and roughly what data you hold. We will come back with an honest view on whether machine learning is the right tool, what the work would involve and a realistic cost range. If a spreadsheet would do the job, we will say so.

  1. You tell us what you needTwo minutes on the form, or a message on WhatsApp.
  2. A senior engineer reviews itAnd comes back with questions, a realistic range and an honest view on fit.
  3. Free 30-minute scoping callWe talk through scope, options and a realistic estimate — with no obligation.
Free estimateNo obligation

Talk to someone who builds this

Send a short brief and we will come back with an honest view and a realistic range.

Takes under a minute. We never share your details.

  • Free consultation
  • No commitment
  • NDA on request

Prefer to talk? Book a free 30-minute call →