The Importance of AI and LLMs in Finance
The recent resurgence of interest in artificial intelligence (AI), particularly through large language models (LLMs) like OpenAI’s ChatGPT and Google’s Gemini, is revolutionizing numerous industries, including finance. These models utilize advanced transformer architectures that enable them to perform language-based tasks with remarkable accuracy. As sentiment analysis becomes increasingly vital for understanding market dynamics, it is crucial for financial professionals to engage with these technologies.
Traditional sentiment classification methods often fall short of achieving human-level benchmarks. LLMs, on the other hand, have demonstrated significant effectiveness in sentiment analysis by being fine-tuned on human-labeled data. This advancement is particularly relevant in finance, where understanding emotional tones can greatly influence decision-making and market behavior.
The growing field of Conviction Narrative Theory (CNT) emphasizes the role of narratives and emotions in decision-making processes, challenging conventional economic thought that often downplays emotional factors. By leveraging LLMs to capture approach and avoidance emotions in financial news, this project aims to enhance the accuracy of sentiment analysis, thereby providing deeper insights into market movements.
Your Expertise Is Essential!
We invite experienced financial risk professionals to participate in a human-labeled tuning exercise where your input will be crucial in shaping the model. You’ll:
- Evaluate financial phrases.
- Score them based on specific risk (BUY/SELL) and general market risk (RISK ON/OFF).
- Provide invaluable insights that will directly influence the development of a model that understands the complex language of financial risk.
What to Expect
- The task will take 1 hour of your time.
- Participation is fully anonymous.
- An instructional video will guide you through the process.
- You’ll have the opportunity for a one-on-one meeting with the research team to ask any questions about the task.
- Your feedback will help refine the model for future applications, and we’ll follow up with a results meeting to discuss your insights and findings.
You’ll be collaborating with a distinguished team of researchers, including:
- Prof. David Tuckett (UCL)
- Prof. Ali Kabiri (UCL, University of Buckingham)
- Prof. John Landon-Lane (Chair, Department of Economics, Rutgers University)
- Dr. David Vinson (UCL Experimental Psychology)
- Dr. Jacob Turton (UCL Centre for the Study of Decision Making Uncertainty)
- Prof. Harold James (Princeton University)