Refinitiv logo
An LSEG Business

How to improve the accuracy of your financial language models  
Imagine AI that understands financial data better than you do....
Financial institutions are increasingly aware of the potential value that can be found in unstructured data to spot issues or uncover trends in financial markets. As such, the Refinitiv Labs have been working on projects to apply traditional language models to financial data to unlock this potential.

One example, is a project in which Refinitiv Labs trained a BERT model leveraging Reuters news, which achieves a high-level of accuracy and returns a single document embedding from the text via API.

The API will enable financial data practitioners to fine-tune models across a number of NLP use cases and applications, without having to do upfront and expensive pre-training and embedding. The first users of Labs’ prototype model have seen better accuracy scores on their models thanks to the high-quality representation of the text generated by Labs’ underlying model.

In this hands-on lab, you will takeaway:

  • Practical steps and learnings from building and training the state-of-the-art BERT language model
  • Key financial use cases and applications that Refinitiv Labs’ new language model will enable
  • How domain-specific language models can be integrated into NLP training pipelines 

Complete the form to access on-demand now.
Event info
Duration:  30 minutes
Login:   Shared after registration


(Please note you can manage and update your preferences at any time.)
(please note you can manage and update your preferences at any time.) (Optional)

By submitting your details, you are agreeing to receive occasional communications about Refinitiv resources, events, products, or services. You also acknowledge that you have read and understood our Privacy Statement.By submitting your details, you are acknowledging that you have read and understood our Privacy Statement.
Meet Our Speakers
Laura Sartenaer
Strategy & Partnerships Manager, Refinitiv Labs

Laura manages Strategy and Partnerships, EMEA at Refinitiv Labs. She assesses market and emerging technology trends to help inform the Labs strategy, as well as initiates and develops key industry relationships with customers, big tech firms, start-ups, governments, trade bodies, universities and research institutes. She is also a Member of the World Economic Forum’s Working Group for AI in Financial Services. Prior to this, Laura spent 2.5 years in the Corporate Strategy and M&A team, where she led the development of the Financial Crime Risk Management strategy and AI strategy. Before joining Refinitiv in September 2016, Laura spent two years as a European Equities Research Analyst at Third Bridge, a primary research firm in London.
Stanimir Vichev
Senior Software Engineer,  Refinitiv Labs

A senior software engineer passionate about all things AI and currently building MVPs to test how cutting edge tech (AI, ML, cloud, graph, alt data) can be used to solve real business problems at Refinitiv Labs. Skilled in machine learning, Python, Node.js, AWS, Docker, graph data, I am constantly looking for opportunities to expand my expertise in AI and software engineering. I am particularly interested in how AI can be used for social good, as well as how to bring it from a prototype to a real product. I am part of DataKind UK, a volunteering organization that helps charities get more value from their data through AI. I have a Master of Information and Data Science (MIDS) degree from UC Berkeley School of Information (completed part-time while working), and a bachelor’s degree in Computer and Business Studies from the University of Warwick. 
Ievgen Goichuk
Senior Software Engineer,  Refinitiv Labs

Software engineer and architect with 15+ years of experience in enterprise software, distributed systems, real-time analytics, big data and machine learning. Agile practitioner. Delivered numerous solution for my clients in financial and e-commerce domains. Leading distributed teams. My motivation is when my products are in use and users are satisfied. Never stop learning. 

Discover what we do

See what we're working on

Experience Refinitiv Labs
Work with our data and APIs