Data Curation for Fine-Tuning Language Models in Capital Markets

Time to put the spotlight on what really drives the success of machine learning models: data curation.

From domain discovery, through designing annotation schemas, to adapting to constantly changing edge-cases, data curation specialists refine the fuel needed to build practical AI systems.

Join us for an exciting webinar that dives deep into how annotations can effectively tackle the unpredictability of the human language. Karolina Drabik, Head of Data Curation at Sense Street, shares her experience working with challenging datasets and discusses innovative strategies to unlock the full potential of your data.

In the webinar:

• Discover the power of descriptive labels - We're aware that not everything can be squeezed into a tag. For highly nuanced, intepretation-driven tasks, descriptive labels will allow you to structure the data without falling into oversimplification.
• Learn about the impact of granularity in annotations - We'll explore how breaking down data into finer categories helps to capture critical nuances and opens up new possibilities for data processing.
• Learn how to ensure quality of annotations - From detailed guidelines to review loops, consistent annotating is a multi-stage process and a truly team effort. We’ll dive into strategies that really make the outcome of this effort "gold standard".
• Level up your prompting game - Get insights into how prompting can support the annotation process, boost dataset growth and reduce the workload from your team.

This will be a great webinar packed with actionable information.

We look forward to seeing you there!

Sense Street