H3M-SSMoEs: Hypergraph-based Multimodal Learning with LLM Reasoning and Style-Structured Mixture of Experts

2025-10-28

AI Summary

Stock price forecasting faces challenges including complex temporal dependencies, heterogeneous modalities, and dynamically evolving inter-stock relationships, making it difficult for existing methods to jointly model structural, semantic, and market-state information within a scalable framework. To address this, we propose a scalable hypergraph neural network framework: (1) a multi-context hypergraph architecture explicitly capturing both local and global dynamic inter-stock relationships; (2) an LLM-enhanced cross-modal semantic alignment module leveraging frozen large language models with lightweight adapters; and (3) a style-vector-driven sparse mixture-of-experts system enabling market-regime-aware adaptation and industry-specific modeling. Evaluated on three major stock markets, our method significantly outperforms state-of-the-art approaches in prediction accuracy and investment returns, while maintaining robust risk control. The code and models are publicly available.

Technology Category

• Machine Learning: Mixture of Experts (MoE)
• Natural Language Processing: (Large) Language Models
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Stock movement prediction remains fundamentally challenging due to complex temporal dependencies, heterogeneous modalities, and dynamically evolving inter-stock relationships. Existing approaches often fail to unify structural, semantic, and regime-adaptive modeling within a scalable framework. This work introduces H3M-SSMoEs, a novel Hypergraph-based MultiModal architecture with LLM reasoning and Style-Structured Mixture of Experts, integrating three key innovations: (1) a Multi-Context Multimodal Hypergraph that hierarchically captures fine-grained spatiotemporal dynamics via a Local Context Hypergraph (LCH) and persistent inter-stock dependencies through a Global Context Hypergraph (GCH), employing shared cross-modal hyperedges and Jensen-Shannon Divergence weighting mechanism for adaptive relational learning and cross-modal alignment; (2) a LLM-enhanced reasoning module, which leverages a frozen large language model with lightweight adapters to semantically fuse and align quantitative and textual modalities, enriching representations with domain-specific financial knowledge; and (3) a Style-Structured Mixture of Experts (SSMoEs) that combines shared market experts and industry-specialized experts, each parameterized by learnable style vectors enabling regime-aware specialization under sparse activation. Extensive experiments on three major stock markets demonstrate that H3M-SSMoEs surpasses state-of-the-art methods in both superior predictive accuracy and investment performance, while exhibiting effective risk control. Datasets, source code, and model weights are available at our GitHub repository: https://github.com/PeilinTime/H3M-SSMoEs.

Problem

• Predicting stock movements with complex temporal dependencies
• Integrating heterogeneous modalities and dynamic inter-stock relationships
• Unifying structural, semantic, and regime-adaptive modeling in scalable framework

Innovation

• Hierarchical hypergraph captures spatiotemporal dynamics and dependencies
• LLM with adapters aligns quantitative and textual modalities
• Mixture of experts combines shared and specialized market modeling