On March 30, 2023, Bloomberg announced the introduction of its
50-billion parameter large language model, purpose-built from scratch for finance, termed
BloombergGPT. The model is specifically designed to assist in various financial natural language processing (NLP) tasks such as sentiment analysis, named entity recognition, news classification, and question-answering. This initiative represents a stride towards harnessing the full potential of AI within the financial industry, addressing the complexities and unique terminologies inherent in finance. The
press release articulates how BloombergGPT is set to outperform similarly-sized open models on financial NLP tasks by significant margins.
The complexity of financial terminology often necessitates the development of models specifically tailored to the sector. General-purpose models might not be efficient enough at interpreting the nuances found in financial texts. According to
Shawn Edwards, Bloomberg's Chief Technology Officer, the launch of BloombergGPT aims to harness the power of AI to handle new applications and enhance performance right out of the box—promising faster time-to-market for custom solutions in finance.