基于前瞻性学习的供应链中断预测

摘要

在企业与政策制定者面临的核心挑战中，供应链中断的预判始终位居前列。关键难点在于如何从嘈杂的非结构化数据中，对低频高影响事件进行可靠推演——这一场景下通用模型若未经任务适配往往表现不佳。我们提出一种端到端框架，通过已发生的中断结果作为监督信号，训练大语言模型生成经过校准的概率预测。实验表明，该模型在准确性、校准度和精确度上显著优于包括GPT-5在内的强基线模型。研究还发现，训练过程能诱导出更结构化、更可靠的概率推理能力，且无需显式提示。这些成果为训练领域专用预测模型提供了通用路径，使其能生成可直接支撑决策的信号。为促进研究透明性，我们开源了本研究的评估数据集。数据集地址：https://huggingface.co/datasets/LightningRodLabs/supply-chain-predictions

English

Anticipating supply chain disruptions before they materialize is a core challenge for firms and policymakers alike. A key difficulty is learning to reason reliably about infrequent, high-impact events from noisy and unstructured inputs - a setting where general-purpose models struggle without task-specific adaptation. We introduce an end-to-end framework that trains LLMs to produce calibrated probabilistic forecasts using realized disruption outcomes as supervision. The resulting model substantially outperforms strong baselines - including GPT-5 - on accuracy, calibration, and precision. We also show that training induces more structured and reliable probabilistic reasoning without explicit prompting. These results suggest a general pathway for training domain-specific forecasting models that produce decision-ready signals. To support transparency we open-source the evaluation dataset used in this study. Dataset: https://huggingface.co/datasets/LightningRodLabs/supply-chain-predictions

基于前瞻性学习的供应链中断预测

Forecasting Supply Chain Disruptions with Foresight Learning

摘要

Support