Joel Nicholas Jennings

Title: Joel Nicholas Jennings and Causal Machine Learning

Introduction

Joel Nicholas Jennings is an inventor based in Cambridge, Great Britain. His patent work focuses on methods for modelling causation in machine learning, including approaches that account for confounders and relationships across time. He is associated with Microsoft Technology Licensing, LLC, and has 3 patents.

Latest Patents

Jennings’s latest patent work includes two inventions titled “Modelling causation in machine learning.”

One method uses sampled causal graphs to identify a parent variable that causes a selected variable, as well as a confounder that may influence both. A machine learning model encodes the parent and information about the confounded pair into embeddings, combines those embeddings, and decodes them to reconstruct a value. The approach may support model training or treatment-effect estimation.

A second method models causal relationships across time. It samples a temporal causal graph to identify both a present-time parent and a parent from a preceding time step. Separate encoders produce embeddings for these variables, which are combined and passed to a decoder to reconstruct the selected variable’s value.

Career Highlights

Jennings’s disclosed inventions address causal structure in machine learning, with particular attention to confounding and temporal dependencies. These methods describe ways to incorporate causal relationships into model representations and reconstruction.

Collaborations

The provided information identifies Cheng Zhang and Marife Defante as Jennings’s coworkers. His company affiliation is Microsoft Technology Licensing, LLC.

Conclusion

Joel Nicholas Jennings’s patent work explores machine learning methods that represent causal relationships, including confounders and causes spanning successive time steps.