Summary of Learning to Generate Explainable Stock Predictions Using Self-reflective Large Language Models, by Kelvin J.l. Koa et al.
Learning to Generate Explainable Stock Predictions using Self-Reflective Large Language Models
by Kelvin J.L. Koa, Yunshan Ma, Ritchie Ng, Tat-Seng Chua
First submitted to arxiv on: 6 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computation and Language (cs.CL); Statistical Finance (q-fin.ST)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper proposes a novel framework called Summarize-Explain-Predict (SEP) to enable Large Language Models (LLMs) to generate explainable stock predictions. Traditional non-generative deep learning models struggle to provide explanations for their decision-making process, while LLMs can generate human-readable explanations but require expert-annotated samples for training, which is impractical. The SEP framework uses a self-reflective agent and Proximal Policy Optimization (PPO) to let the LLM teach itself how to generate explainable stock predictions autonomously. The reflective agent learns to explain past stock movements through self-reasoning, while the PPO trainer trains the model to generate the most likely explanations from input texts. Experimental results show that the SEP framework outperforms traditional deep-learning and LLM methods in prediction accuracy and Matthews correlation coefficient for the stock classification task. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine a computer program that can predict the future value of a stock based on what people are saying about it online. This program, called a Large Language Model (LLM), is very good at understanding human language but has trouble explaining why it made certain predictions. The problem is that we need humans to help train these models by showing them examples of what makes sense and what doesn’t. But this takes too much time and money. To solve this problem, the authors propose a new way to train LLMs using a special kind of thinking called “self-reflection.” This allows the model to learn how to explain its predictions without needing human help. The results show that their method is better than existing approaches at predicting stock values and understanding why it made certain decisions. |
Keywords
* Artificial intelligence * Classification * Deep learning * Large language model * Optimization
