Traceable Multi-Agent System for Knowledge-Based Forecasting

2026-08-04

AI Summary

This work addresses the lack of traceability in multi-agent systems for enterprise forecasting, where it is often difficult to explain the rationale behind prediction changes. The authors propose TraceMAS, a novel system that integrates causal loop diagrams (CLDs) with multi-agent forecasting for the first time. By organizing agent outputs through an ideal CLD anchored in empirical data, TraceMAS links textual evidence, data selection, and model revisions, enabling end-to-end traceability from domain knowledge and data modeling to final predictions. The system combines multi-agent collaboration, document parsing, data-agent mapping, feature construction, and interactive visualization. Evaluated on crude oil price forecasting, TraceMAS allows users to interactively inspect prediction iterations, trace agents’ decision revisions, and connect market narratives with causal structures to support context-aware forecasting.

Abstract

Enterprise forecasting increasingly relies on autonomous agents that interpret documents, search for data, generate code, and revise models. While this autonomy helps build adaptive forecasting pipelines, it also makes it difficult for practitioners to inspect why a forecast changed, which evidence supported the change, and how data and modeling choices were revised. We present TraceMAS, an interactive demo system for traceable multi-agent forecasting. TraceMAS organizes agent outputs around two causal-loop representations: an Ideal Causal Loop Diagram (Ideal CLD), which captures key factors and their causal relations extracted from domain documents, and a Data-Grounded Causal Loop Diagram (Data-Grounded CLD), which links those factors to internal variables, external data, or documented proxies. The Data-Grounded CLD guides feature construction and model design while preserving the connection between textual evidence, data choices, and model revisions. We demonstrate TraceMAS on crude oil price forecasting. The demo interface allows users to compare forecasting iterations, inspect agent-level revisions, explore causal maps, review feature-data mappings and model architecture, and connect scenario forecasts to market narratives. This demonstration shows how autonomous forecasting agents can retain flexibility while making the evidence-to-forecast process inspectable.

Problem

• traceability
• multi-agent system
• knowledge-based forecasting
• causal loop diagram
• forecasting interpretability

Innovation

• Traceable Multi-Agent System
• Causal Loop Diagram
• Knowledge-Based Forecasting
• Interpretable AI
• Autonomous Forecasting Agents