<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=648880075207035&amp;ev=PageView&amp;noscript=1">
Judicial Reasoning with BayesiaLab

Free Seminar for Law Professionals:

Judicial Reasoning, Bayes' Rule, and Artificial Intelligence

Friday, November 22, 2019, 2:00 p.m.–5:00 p.m.
Spaces—Chrysler Building, 9th Floor, 405 Lexington Avenue, New York, NY 10174

Seminar Overview

An Introduction for Law Practitioners

Resolving uncertainty given conflicting evidence and contradictory explanations is central to judicial reasoning. In this three-hour seminar, we introduce Bayes' Rule, which provides a general-purpose procedure for reasoning under uncertainty.

While universally recognized as normative for over 200 years, Bayesian reasoning has remained difficult to utilize for non-statisticians and, thus, its use in jurisprudence is still in its infancy.

We present Bayesian networks as a practical and intuitive type of Artificial Intelligence, which is ideally suited for performing Bayesian reasoning in a legal context. And, by employing the BayesiaLab software platform, we demonstrate that Bayesian reasoning can be as straightforward as doing arithmetic with a spreadsheet.


The Limits of Logic & Cognitive Challenges

The motivation for this seminar consists of three main points:

  1. Given their inherent uncertainty, legal questions are mostly probabilistic and not deterministic in nature, which means that formal deductive logic cannot be applied. Probabilistic reasoning requires the application of Bayes' Rule.
  2. Human intuition is deeply flawed when performing probabilistic reasoning. As a result, human inference is generally inconsistent with normative Bayesian reasoning. The "Prosecutor's Fallacy" is perhaps the best-known problem of this kind.
  3. Except in trivial cases, standard computational tools, such as spreadsheets, are unable to perform the calculations involving Bayes Rule. Until recently, there has been no practical "computational aid" for reasoning.
Dimensions of Reasoning

Seminar Agenda

  • Dimensions of Reasoning:
    • Prediction vs. Causation
    • Theory vs. Data
    • Probabilistic vs. Deterministic
  • Introduction of the Bayesian network formalism
  • The BayesiaLab 9 software platform
  • Evidential Reasoning
    • Encoding uncertain domain knowledge in a Bayesian network.
    • Reasoning from effect to cause, e.g., evaluating a witness statement.
      • Prosecutor's Fallacy or the Fallacy of the Transposed Conditional.
    • Inter-causal reasoning ("explaining away").
    • Using Mutual Information to measure the "strength" of information.
    • Assess uncertainties in a case to compute overall risk.
    • Measuring the "conflict" between multiple pieces of evidence using the Bayes Factor.
    • Using the "Most Relevant Explanation" function in BayesiaLab to establish the most probable cause.
  • Causal Reasoning
    • Simulating interventions to determine causal effects.
      • Resolving Simpson's Paradox.
      • Estimating biases and quantifying potential discrimination.
  • Contribution & Attribution
    • Using causal counterfactuals to attribute observed effects to its causes, e.g., for proportionally assigning liability to multiple parties in litigation.
  • Knowledge Elicitation from Stakeholders

Seminar Format, Technology, and Materials

  • BayesiaLab SeminarIn this seminar, we alternate slides presentations and group discussions of case studies using BayesiaLab as the reasoning platform.
  • The number of participants is limited to 15.
  • There will be one 10-minute break at approx. 3:30 p.m.
  • You will have access to a 30-day license of BayesiaLab 9 (Standard Edition) for independent practice with the examples presented in the seminar.
  • You will receive all presentation slides in PDF format plus a screen recording of the seminar.
  • You can download all Bayesian network models used in the seminar.


  • This seminar is intended for law practitioners, plus students and faculty in the field of law.
  • No mathematical, statistical, or programming skills are required. 
  • Even though the seminar is free of mathematical formulas and statistical jargon, it is a fast-paced and intellectually challenging program. So, your full concentration over three hours will be required.

Seminar Registration

Location Map

About the Instructor

Stefan ConradyStefan Conrady has over 20 years of corporate experience in with leading automotive brands, such as Mercedes-Benz, BMW, and Rolls-Royce Motor Cars. Stefan is a native of Ulm, Germany, but his career has spanned the globe, having lived and worked in Chicago, New York, Munich, and Singapore, just to name a few. In his most recent corporate assignment, he was heading the Analytics & Forecasting group at Nissan North America.

Today, in his role as Managing Partner of Bayesia USA and Bayesia Singapore, he is recognized as a thought leader in applying Artificial Intelligence for research, analytics, and reasoning. Stefan's tutorials, seminars, and lectures on Bayesian Networks are widely followed by scientists who embrace AI innovations to improve decision-making. In this context, Stefan has recently co-authored a book with Lionel Jouffe, Bayesian Networks & BayesiaLab — A Practical Introduction for Researchers.

Related Materials