Details

GAIA is proud to present yet another meetup!

This time it will take place at Recorded Future with a talk on how the NLP team has been able to scale up supervised model development with active learning. Note that this meetup takes place during working hours. (Our hope is to enable more of the GAIA community to join. Feedback on this is welcome and we'll evaluate the concept afterwards.)

Schedule

15:30 Mingle and fika

16:00 Presentation

17:00 Mingle and snacks

Streamlining Rapid Development of Supervised NLP Models

Recorded Future finds itself at the crux of processing high-volume text data churned out by the internet every single day. To make sense of this copious data and filter out valuable insights, we rely heavily on machine learning models to sift through, categorize, and pinpoint noteworthy pieces of information.

In this talk, we will traverse the journey of our model development process, honed to facilitate rapid creation of effective models. This workflow allows us to make quick, data-informed decisions, efficiently sorting through the vast expanse of text data and bringing significant information to the forefront.

The machine learning cycle extends well beyond the deployment of a model into production. This talk will also delve into our procedures for maintaining optimal model performance, discussing how we monitor active models to detect any decline in performance and identify potential enhancement opportunities.

Every process inherently possesses potential for refinement. Thus, we'll discuss how we are, and to some extent already have been, improving this process by incorporating Large Language Models into our workflow.

You're welcome to join us for hopefully an inspiering afternoon with valuable learnings and engaging discussions. Hope to see you there!