New
a weak supervision annotation tool
Learn more

AI research, ideas and product updates.

Blog

Why you should be annotating in-house for NLP

Text labeling is vital if you want to build high quality NLP models. In the past, you used to have to rely on outsourced labour but recently new technologies like transfer learning, Programmatic labeling and active learning mean you can train state of the art models with just a few hundred labels. Now that the volume of data required has been reduced there are huge advantages to labeling in house: you have more control over quality, can iterate faster and don't have to create privacy sharing agreements. Humanloop have built a suite of tools that make annotating data and improving NLP models much easier.

Raza Habib
May 2022
Announcement

Introducing Humanloop Programmatic

Humanloop Programmatic is now available for early access. A powerful weak labelling tool to rapidly annotate your NLP datasets.

Jordan Burgess
March 2022
Blog

Why I changed my mind about weak labelling for ML

Weak labelling can replace manually annotated data with data created from heuristic rules written by domain experts. Here I want to explain what weak supervision is, what I learned and why I think it complements techniques like active learning for data annotation.

Raza Habib
December 2021
Blog

What is Human-in-the-Loop AI?

Human-in-the-loop AI is an automation approach which means you can quickly deploy a working model, with less data and with guaranteed quality predictions. In this post we'll explain what HITL is and how you can make use of this approach in your own AI projects.

Jordan Burgess
November 2021
Blog

Why you should take a data first approach to AI

Machine learning systems are born through the marriage of both code and data. Academia mostly focuses on ways to improve the learning algorithms, the *how* of machine learning. When you come to build practical AI systems though the data-set you're training on has at least as much impact on performance as the choice of algorithm.

Raza Habib
August 2021
Blog

What is an active learning platform?

Data first ML Ops needs new kinds of tools. An Active Learning platform is a single tool that combines data annotation, model training and deployment in one continuous process.

Raza Habib
August 2021
NLP

How good is GPT-3 in practice?

We discuss the real world performance of Open AI's GPT-3 language model and speculate on what the future of large pre-trained language models may hold.

Raza Habib
July 2021
Blog

Four lessons from a year building tools for machine learning

The most surprising lessons about how to build tooling for machine learning, what is needed going forwards and why domain experts will play a much larger role in the future of AI.

Raza Habib
July 2021
Case Study

Measuring Active Learning performance in the real world

Humanloop worked with Black Swan Data, a company at the forefront of applied NLP, to test our active learning platform. Humanloop was able to save at least 40% of the labeling normally required whilst improving performance

Raza Habib
March 2021
Announcement

Humanloop + Hugging Face 🤗 = the fastest way to build custom NLP

You can now train any Huggingface model on your own data just by labeling and curating a small portion of that data, with early access to our Humanloop - Hugging Face integration. Label your data and we'll give you an API to your model, trained on your own data

Raza Habib
February 2021
Machine Learning

Why you should be using active learning to build ML

Data labeling is often the biggest bottleneck in machine learning. Active learning lets you train machine learning models with much less labeled data. The best AI-driven companies, like Tesla, use active learning.

Raza Habib
February 2021
John R.
Chartered Legal Executive

John works at one of the UKs largest legal firms. He’d been tasked with labelling over 35,000 contracts into various categories - not a fun job. But with help from Humanloop, he was able to build and train an AI model that automatically orgsanised the contracts into the correct category with a 98.7% accuracy. He completed the task in under two days, which would usually have taken over 200 hours - that’s a happy John.

Legal services
John R.
Chartered Legal Executive

John works at one of the UKs largest legal firms. He’d been tasked with labelling over 35,000 contracts into various categories - not a fun job. But with help from Humanloop, he was able to build and train an AI model that automatically orgsanised the contracts into the correct category with a 98.7% accuracy. He completed the task in under two days, which would usually have taken over 200 hours - that’s a happy John.

Customer service
John R.
Chartered Legal Executive

John works at one of the UKs largest legal firms. He’d been tasked with labelling over 35,000 contracts into various categories - not a fun job. But with help from Humanloop, he was able to build and train an AI model that automatically orgsanised the contracts into the correct category with a 98.7% accuracy. He completed the task in under two days, which would usually have taken over 200 hours - that’s a happy John.

Community management
John R.
Chartered Legal Executive

John works at one of the UKs largest legal firms. He’d been tasked with labelling over 35,000 contracts into various categories - not a fun job. But with help from Humanloop, he was able to build and train an AI model that automatically orgsanised the contracts into the correct category with a 98.7% accuracy. He completed the task in under two days, which would usually have taken over 200 hours - that’s a happy John.

Social media
Rotating lissajous

Watch how to build a custom AI in minutes

Arrange a demo