点击蓝字 · 关注我们
本期主题
Out of the (Black)Box: AI as Conditional Probability
时 间：2025年3月11日（Tue.）
15:30 p.m. - 17:00 p.m.
地点：同德楼204
主讲：Antoine Didisheim(UniMelb)
报告摘要
Antoine Didisheim is an Assistant Professor in the Department of Finance at the Faculty of Business and Economics, University of Melbourne. His research is using machine learning techniques to explore questions related to asset pricing. He received his Ph.D. in finance from the Swiss Finance Institute in 2022. He holds a master's degree in data science with machine learning from University College London and a master's degree in finance from the University of Lausanne.
报告人简介
The core technology powering modern Large Language Models (LLMs) estimates the distribution of probable answers conditional on the prompt. Using a financial news and returns dataset, we find that these conditional probabilities are interpretable and contain valuable economic information. Conversely, measures of declared confidence used in the literature are opaque, structurally biased, unstable, and more model-dependent, indicating that LLMs cannot assess their own confidence. Using conditional probabilities, we analyze LLM biases and provide insights into the internal mechanisms driving model decisions. Our results indicate that conditional probabilities provide a reliable and transparent reflection of LLM priors, particularly for economic applications.
欢迎各位老师和同学们前来参加！
