Most of my work involves the application and development of Bayesian models and Reinforcement Learning. I believe that the generalized linear mixed model is an undervalued tool which has the potential to bridge many gaps in a variety of contexts: Correct specification of design and contrasts is an elegant and efficient route to resolve many practical problems.

 

Bayesian models can be used in tandem with Machine Learning methods such as Reinforcement Learning as well, in any case where optimal decision-making is needed. Inference in the model-based decision-making framework, of course, is causal. My forthcoming book on Optimal Control Using Causal Agents outlines the deep mathematical synergy between these two revolutionary fields, which have largely remained in their respective silos with few cross-citations between them. The book is under contract with CRC press, with target completion around May 2027. I am also organizing a workshop on optimal decision-making this July. 

 

For speaking engagements and seminar requests, for fastest response please send me an email with the location and date in the subject line.

 

Email: MaryLena.Bleile [at] TechDataService.com

X: ML_does_ML

Linkedin

 

 

 

 

Speaking

 

1. "Decision-relevant Observability in Stochastic Optimal Control: A Relativistic Information Criterion for Reinforcement Learning Convergence"
   • Workshop on Partial Differential Equations, Mathematical Physics, and Numerics, 16-17 August 2025
   • NxtAI, 3-5 November 2025
2. "Relative Information Theory in Cell Type Deconvolution: When More Accurate Methods Give Worse Actionable Insights" BioTechX, 16-17 September 2025

3. "Using bioinformatics + InSilco to optimise research and development" BioTechX (panel discussion), 16-17 September 2025

4. "Biomarkers for Neurodegenerative Disorders and Cancers" BioTechX, 16-17 September 2024
5. "Cell State Heterogeneity In Spatial Transcriptomics Deconvolution" NYU Langone Health, 13 May 2024
6. "A Domain-Oriented Analysis Pipeline for Spatial Transcriptomics with Application to Melanoma Data". Levin Lecture at Columbia University, 11 November 2023
7. Statistical and Machine Learning Methods for Adaptive Radiotherapy Treatment Plans
   • Memorial Sloan-Kettering Cancer Center, March 2023
   • Onc.AI, January 2023
8. "Animal Sacrifice as a Potential Cause of Decelerating Tumor Growth in Late-State Xenograft
   Experiments" Berry Consulting, February 2023