VISTA: Vision-Language Inference for Training-Free Stock Time-Series Analysis

2025-05-24

AI Summary

Stock price forecasting faces challenges in capturing complementary patterns from heterogeneous data sources using unimodal approaches. This paper proposes a training-free multimodal zero-shot inference framework that jointly processes historical stock price time series (as textual sequences) and their corresponding line charts (as visual inputs), leveraging vision-language models (e.g., LLaVA, Qwen-VL) for end-to-end future price prediction. Our key contributions are: (1) the first training-free multimodal paradigm for financial time-series analysis; and (2) a chain-of-thought prompting mechanism that synergistically extracts temporal patterns from both numerical and visual modalities. Evaluated on standard benchmarks, our method achieves up to an 89.83% improvement over ARIMA and text-only large language models. These results robustly demonstrate the efficacy and generalizability of unsupervised multimodal collaborative reasoning in financial forecasting.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Multi-modal Vision
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Stock price prediction remains a complex and high-stakes task in financial analysis, traditionally addressed using statistical models or, more recently, language models. In this work, we introduce VISTA (Vision-Language Inference for Stock Time-series Analysis), a novel, training-free framework that leverages Vision-Language Models (VLMs) for multi-modal stock forecasting. VISTA prompts a VLM with both textual representations of historical stock prices and their corresponding line charts to predict future price values. By combining numerical and visual modalities in a zero-shot setting and using carefully designed chain-of-thought prompts, VISTA captures complementary patterns that unimodal approaches often miss. We benchmark VISTA against standard baselines, including ARIMA and text-only LLM-based prompting methods. Experimental results show that VISTA outperforms these baselines by up to 89.83%, demonstrating the effectiveness of multi-modal inference for stock time-series analysis and highlighting the potential of VLMs in financial forecasting tasks without requiring task-specific training.

Problem

• Training-free stock price prediction using vision-language models
• Multi-modal forecasting combining text and visual data
• Zero-shot approach outperforming traditional statistical methods

Innovation

• Training-free VLM framework for stock prediction
• Combines numerical and visual data modalities
• Zero-shot chain-of-thought prompting technique