Tom Dietterich

Title: Tom Dietterich: Innovator in Machine Learning for Molecular Modeling

Introduction

Tom Dietterich is a prominent inventor based in Corvallis, OR (US). He has made significant contributions to the field of machine learning, particularly in the context of modeling biological activity for molecular structures. His innovative approach combines explicit representations of molecular shapes with advanced neural network learning methods.

Latest Patents

Tom Dietterich holds a patent for a machine-learning approach that models biological activity for molecules. This patent emphasizes the explicit representation of molecular shapes, which is integrated with neural network learning techniques. The methodology he developed allows for high predictive ability and generalization across different chemical classes. It treats structurally diverse molecules with similar surface characteristics as analogous. The new machine-learning methodology can accept multiple representations of objects and construct predictive models for their characteristics. An iterative process is employed to adjust parameters, generating new representations of the objects and retraining the models for improved predictions. This method is particularly applicable to molecules, as they can exhibit various orientations and conformations determined by translation, rotation, and torsion angle parameters.

Career Highlights

Tom Dietterich is currently associated with Arris Pharmaceutical Corporation, where he applies his expertise in machine learning to advance pharmaceutical research. His work has been instrumental in developing predictive models that enhance the understanding of molecular interactions and biological activity.

Collaborations

Throughout his career, Tom has collaborated with notable colleagues, including David Chapman and Rick Lathrop. These partnerships have fostered innovation and contributed to the advancement of machine learning applications in the pharmaceutical industry.

Conclusion

Tom Dietterich's contributions to machine learning and molecular modeling exemplify the intersection of technology and biology. His innovative methodologies continue to influence the field, paving the way for future advancements in drug discovery and molecular research.