Challenge
Recent developments in Israel have highlighted the impact of cognitive bias and, at times, confirmation bias on intelligence data analysis. In the late 1970s, the CIA developed a methodology to mitigate cognitive bias during intelligence analysis. This approach, known as ACH (Analysis of Competing Hypotheses), breaks down high-stakes decisions into a series of smaller, arguably simpler decisions based on individual pieces of information. However, the process remains manual, time-consuming, and reliant on human judgment
RnD Objectives
In this project, we aim to utilize the emerging LLM-based multiagent paradigm for analyzing competing hypotheses in intelligence analysis and other applications. Drawing inspiration from the conceptual methodology of ACH, we will investigate different agent compositions and their specializations to enhance the reliable and efficient analysis of incoming intelligence data streams
Intelligence Data Analysis
Share a link using:
https://www.afeka.ac.il/en/industry-relations/research-centers/afeka-interdisciplinary-center-for-social-good-generative-ai/intelligence-data-analysis/WhatsApp
Facebook
Twitter
Email
https://www.afeka.ac.il/en/industry-relations/research-centers/afeka-interdisciplinary-center-for-social-good-generative-ai/intelligence-data-analysis/