Abstract:Cryptocurrency investment is inherently difficult due to its shorter history compared to traditional assets, the need to integrate vast amounts of data from various modalities, and the requirement for complex reasoning. While deep learning approaches have been applied to address these challenges, their black-box nature raises concerns about trust and explainability. Recently, large language models (LLMs) have shown promise in financial applications due to their ability to understand multi-modal data and generate explainable decisions. However, single LLM faces limitations in complex, comprehensive tasks such as asset investment. These limitations are even more pronounced in cryptocurrency investment, where LLMs have less domain-specific knowledge in their training corpora. To overcome these challenges, we propose an explainable, multi-modal, multi-agent framework for cryptocurrency investment. Our framework uses specialized agents that collaborate within and across teams to handle subtasks such as data analysis, literature integration, and investment decision-making for the top 30 cryptocurrencies by market capitalization. The expert training module fine-tunes agents using multi-modal historical data and professional investment literature, while the multi-agent investment module employs real-time data to make informed cryptocurrency investment decisions. Unique intrateam and interteam collaboration mechanisms enhance prediction accuracy by adjusting final predictions based on confidence levels within agent teams and facilitating information sharing between teams. Empirical evaluation using data from November 2023 to September 2024 demonstrates that our framework outperforms single-agent models and market benchmarks in classification, asset pricing, portfolio, and explainability performance.

What problem does this paper attempt to address?

The problems that this paper attempts to solve are several key challenges in cryptocurrency investment: 1. **Diversity and complexity of data modalities**: Cryptocurrency investment requires the integration of a large amount of data from different modalities (such as text, images, etc.), which increases the complexity of data analysis. 2. **Limited historical data and evidence of asset pricing**: Compared with traditional assets, cryptocurrencies have a shorter history, resulting in limited available historical data and evidence of asset pricing. 3. **Complex reasoning requirements**: The volatility and uncertainty of the cryptocurrency market require complex reasoning and decision - making. 4. **The "black - box" problem of deep - learning models**: Although deep - learning methods have made certain progress in dealing with these challenges, their "black - box" nature makes trust and interpretability an issue. 5. **Limitations of a single LLM**: Large - language models (LLMs) have limited performance in handling complex and comprehensive tasks, especially in the field of cryptocurrency investment, because they lack domain - specific knowledge in the training corpus. To solve these problems, the paper proposes an **explainable multi - modal multi - agent framework**, which aims to handle cryptocurrency investment tasks through the collaboration among multiple professional agents. Specifically: - **Expert training module**: Use multi - modal historical data and professional investment literature to fine - tune agents to generate high - quality training prompts and explanations. - **Multi - agent investment module**: Use real - time data to make cryptocurrency investment decisions and improve prediction accuracy through intra - team and inter - team collaboration mechanisms. The main contributions of this framework include: - Proposing the first multi - agent framework that combines multi - modal data and professional investment knowledge for portfolio management of large - cap cryptocurrencies. - Designing unique intra - team and inter - team collaboration mechanisms to promote communication and reduce prediction errors among different agents. - Developing a new asset - pricing method that converts the binary up - down price - trend classification into a confidence - level spectrum, thereby constructing portfolios according to the empirical asset - pricing methods in the financial field. - Generating high - quality predictions and explanations that can effectively explain the changes in cryptocurrency returns by fine - tuning learning historical data and relevant academic literature. - The experimental results show that this multi - agent model is not only superior to a single LLM in classification, asset pricing, and scalability, but also outperforms the market benchmark in portfolio performance. In conclusion, this paper aims to overcome the data complexity, limited historical data, and the "black - box" problem of deep - learning models in cryptocurrency investment by introducing a multi - agent system, and provide a more reliable and explainable investment - decision - making tool.

LLM-Powered Multi-Agent System for Automated Crypto Portfolio Management

A Reflective LLM-based Agent to Guide Zero-shot Cryptocurrency Trading

Automate Strategy Finding with LLM in Quant investment

Modelling crypto markets by multi-agent reinforcement learning

Enhancing LLM Trading Performance with Fact-Subjectivity Aware Reasoning

TradingGPT: Multi-Agent System with Layered Memory and Distinct Characters for Enhanced Financial Trading Performance

FinVision: A Multi-Agent Framework for Stock Market Prediction

FinCon: A Synthesized LLM Multi-Agent System with Conceptual Verbal Reinforcement for Enhanced Financial Decision Making

Multiagent-based deep reinforcement learning framework for multi-asset adaptive trading and portfolio management

FinMem: A Performance-Enhanced LLM Trading Agent with Layered Memory and Character Design

Enhancing Financial Question Answering with a Multi-Agent Reflection Framework

Online portfolio management via deep reinforcement learning with high-frequency data

Multi-Agent Deep Reinforcement Learning With Progressive Negative Reward for Cryptocurrency Trading

Learning to Generate Explainable Stock Predictions using Self-Reflective Large Language Models

A Multimodal Foundation Agent for Financial Trading: Tool-Augmented, Diversified, and Generalist

A Deep Reinforcement Learning Framework for the Financial Portfolio Management Problem

Enhancing Anomaly Detection in Financial Markets with an LLM-based Multi-Agent Framework

When AI Meets Finance (StockAgent): Large Language Model-based Stock Trading in Simulated Real-world Environments

Fundamental Multi-factor Deep-learning Strategy For Cryptocurrency Trading

AAPM: Large Language Model Agent-based Asset Pricing Models

UNSURE - A machine learning approach to cryptocurrency trading