LightTag Blog

2018

Efficiently Labeling Data for NLP

Deep learning applied to NLP has allowed practitioners understand their data less, in exchange for more labeled data. Thus, labeled data has become the bottleneck and cost center of many NLP efforts.

3 min read

Embrace the noise: A case study of text annotation for medical imaging

In this post we’ll discuss the recent paper TextRay: Mining Clinical Reports to Gain a Broad Understanding of Chest X-rays focusing on the best practices the paper exemplifies with regards to labeling text data for NLP. 

What is TextRay ? 

TextRay uses NLP to find what text should be red and then figure out what labels the red text indicates. Doing that, they can produce enough labeled XRays to train a deep learning model. 
TextRay uses NLP to find what text should be red and then figure out what labels the red text indicates. Doing that, they can produce enough labeled XRays to train a deep learning model. 
10 min read

Bug Postmortem: Wrong image deployed on Docker Swarm

On Thursday June 7th 2018 a new customer emailed us reporting that they were unable to submit the work they’d done on our platform. The cause of this error was the deployment of an outdated version of our frontend. Other customer sites and customers who use our on-premise offering were not affected.

4 min read

Active Learning: Optimization != Improvement

Labeled data has become paramount to the success of many business ventures and research projects. But obtaining labeled data remains a costly exercise. Active Learning is a technique that promises to make obtaining labeled data more efficient and has recently been hyped by a number of companies.

11 min read
Back to Top ↑