#142 Bayesian Trees & Deep Learning for Optimization & Big Data, with Gabriel Stechschulte

Sign up for Alex's first live cohort, about Hierarchical Model building!Get 25% off "Building AI Applications for Data Scientists and Software Engineers"Proudly sponsored by PyMC Labs, the Bayesian Consultancy. Book a call, or get in touch!Our&nbsp;theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and&nbsp;Mega Ran). Check out his awesome work!Visit our Patreon page to unlock exclusive Bayesian swag ;)Takeaways:Why GPs still matter: Gaussian Processes remain a go-to for function estimation, active learning, and experimental design – especially when calibrated uncertainty is non-negotiable.Scaling GP inference: Variational methods with inducing points (as in GPflow) make GPs practical on larger datasets without throwing away principled Bayes.MCMC in practice: Clever parameterizations and gradient-based samplers tighten mixing and efficiency; use MCMC when you need gold-standard posteriors.Bayesian deep learning, pragmatically: Stochastic-gradient training and approximate posteriors bring Bayesian ideas to neural networks at scale.Uncertainty that ships: Monte Carlo dropout and related tricks provide fast, usable uncertainty – even if they’re approximations.Model complexity ≠ model quality: Understanding capacity, priors, and inductive bias is key to getting trustworthy predictions.Deep Gaussian Processes: Layered GPs offer flexibility for complex functions, with clear trade-offs in interpretability and compute.Generative models through a Bayesian lens: GANs and friends benefit from explicit priors and uncertainty – useful for safety and downstream decisions.Tooling that matters: Frameworks like GPflow lower the friction from idea to implementation, encouraging reproducible, well-tested modeling.Where we’re headed: The future of ML is uncertainty-aware by default – integrating UQ tightly into optimization, design, and deployment.Chapters:08:44 Function Estimation and Bayesian Deep Learning10:41 Understanding Deep Gaussian Processes25:17 Choosing Between Deep GPs and Neural Networks32:01 Interpretability and Practical Tools for GPs43:52 Variational Methods in Gaussian Processes54:44 Deep Neural Networks and Bayesian Inference01:06:13 The Future of Bayesian Deep Learning01:12:28 Advice for Aspiring Researchers<p...

Sign up for Alex's first live cohort, about Hierarchical Model buildingSoccer Factor Model DashboardToday’s clip is from episode 143 of the podcast, with Christoph Bamberg.Christoph shares his journey into Bayesian statistics and computational modeling, the challenges faced in academia, and the technical tools used in research. Alex and Christoph delve into a specific study on appetite regulation and cognitive performance, exploring the implications of framing in psychological research and the importance of careful communication in health-related contexts.Get the full discussion here.Intro to Bayes Course (first 2 lessons free)Advanced Regression Course (first 2 lessons free)Our&nbsp;theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and&nbsp;Mega Ran). Check out his awesome work!Visit our Patreon page to unlock exclusive Bayesian swag ;)TranscriptThis is an automatic transcript and may therefore contain errors. Please get in touch if you're willing to correct them.

Sign up for Alex's first live cohort, about Hierarchical Model building!Proudly sponsored by PyMC Labs, the Bayesian Consultancy. Book a call, or get in touch!Intro to Bayes Course (first 2 lessons free)Advanced Regression Course (first 2 lessons free)Our&nbsp;theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and&nbsp;Mega Ran). Check out his awesome work!Visit our Patreon page to unlock exclusive Bayesian swag ;)Takeaways:Bayesian mindset in psychology: Why priors, model checking, and full uncertainty reporting make findings more honest and useful.Intermittent fasting &amp; cognition: A Bayesian meta-analysis suggests effects are context- and age-dependent – and often small but meaningful.Framing matters: The way we frame dietary advice (focus, flexibility, timing) can shape adherence and perceived cognitive benefits.From cravings to choices: Appetite, craving, stress, and mood interact to influence eating and cognitive performance throughout the day.Define before you measure: Clear definitions (and DAGs to encode assumptions) reduce ambiguity and guide better study design.DAGs for causal thinking: Directed acyclic graphs help separate hypotheses from data pipelines and make causal claims auditable.Small effects, big implications: Well-estimated “small” effects can scale to public-health relevance when decisions repeat daily.Teaching by modeling: Helping students write models (not just run them) builds statistical thinking and scientific literacy.Bridging lab and life: Balancing careful experiments with real-world measurement is key to actionable health-psychology insights.Trust through transparency: Openly communicating assumptions, uncertainty, and limitations strengthens scientific credibility.Chapters:10:35 The Struggles of Bayesian Statistics in Psychology22:30 Exploring Appetite and Cognitive Performance29:45 Research Methodology and Causal Inference36:36 Understanding Cravings and Definitions39:02 Intermittent Fasting and Cognitive Performance42:57 Practical Recommendations for Intermittent Fasting49:40 Balancing Experimental Psychology and Statistical Modeling55:00 Pressing Questions in Health Psychology01:04:50 Future Directions in ResearchThank you to my Patrons for...

Soccer Factor Model DashboardUnveiling True Talent: The Soccer Factor Model for Skill EvaluationLBS #91, Exploring European Football Analytics, with Max GöbelGet early access to Alex's next live-cohort courses!Today’s clip is from episode 142 of the podcast, with Gabriel Stechschulte.Alex and Garbriel explore the re-implementation of BART (Bayesian Additive Regression Trees) in Rust, detailing the technical challenges and performance improvements achieved.They also share insights into the benefits of BART, such as uncertainty quantification, and its application in various data-intensive fields.Get the full discussion here.Intro to Bayes Course (first 2 lessons free)Advanced Regression Course (first 2 lessons free)Our&nbsp;theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and&nbsp;Mega Ran). Check out his awesome work!Visit our Patreon page to unlock exclusive Bayesian swag ;)TranscriptThis is an automatic transcript and may therefore contain errors. Please get in touch if you're willing to correct them.

Proudly sponsored by PyMC Labs, the Bayesian Consultancy. Book a call, or get in touch!Get early access to Alex's next live-cohort courses!Intro to Bayes Course (first 2 lessons free)Advanced Regression Course (first 2 lessons free)Our&nbsp;theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and&nbsp;Mega Ran). Check out his awesome work!Visit our Patreon page to unlock exclusive Bayesian swag ;)Takeaways:BART as a core tool: Gabriel explains how Bayesian Additive Regression Trees provide robust uncertainty quantification and serve as a reliable baseline model in many domains.Rust for performance: His Rust re-implementation of BART dramatically improves speed and scalability, making it feasible for larger datasets and real-world IoT applications.Strengths and trade-offs: BART avoids overfitting and handles missing data gracefully, though it is slower than other tree-based approaches.Big data meets Bayes: Gabriel shares strategies for applying Bayesian methods with big data, including when variational inference helps balance scale with rigor.Optimization and decision-making: He highlights how BART models can be embedded into optimization frameworks, opening doors for sequential decision-making.Open source matters: Gabriel emphasizes the importance of communities like PyMC and Bambi, encouraging newcomers to start with small contributions.Chapters:05:10 – From economics to IoT and Bayesian statistics18:55 – Introduction to BART (Bayesian Additive Regression Trees)24:40 – Re-implementing BART in Rust for speed and scalability32:05 – Comparing BART with Gaussian Processes and other tree methods39:50 – Strengths and limitations of BART47:15 – Handling missing data and different likelihoods54:30 – Variational inference and big data challenges01:01:10 – Embedding BART into optimization and decision-making frameworks01:08:45 – Open source, PyMC, and community support01:15:20 – Advice for newcomers01:20:55 – Future of BART, Rust, and probabilistic programmingThank you to my Patrons for making this episode possible!Yusuke Saito, Avi Bryant, Ero Carrera, Giuliano Cruz, James Wade, Tradd Salvo, William Benton, James Ahloy, Robin Taylor,, Chad Scherrer, Zwelithini Tunyiswa, Bertrand Wilden, James Thompson, Stephen Oates, Gian Luca Di Tanna, Jack Wells, Matthew Maldonado, Ian...

Get early access to Alex's next live-cohort courses!Today’s clip is from episode 141 of the podcast, with Sam Witty.Alex and Sam discuss the ChiRho project, delving into the intricacies of causal inference, particularly focusing on Do-Calculus, regression discontinuity designs, and Bayesian structural causal inference. They explain ChiRho's design philosophy, emphasizing its modular and extensible nature, and highlights the importance of efficient estimation in causal inference, making complex statistical methods accessible to users without extensive expertise.Get the full discussion here.Intro to Bayes Course (first 2 lessons free)Advanced Regression Course (first 2 lessons free)Our&nbsp;theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and&nbsp;Mega Ran). Check out his awesome work!Visit our Patreon page to unlock exclusive Bayesian swag ;)TranscriptThis is an automatic transcript and may therefore contain errors. Please get in touch if you're willing to correct them.

Proudly sponsored by PyMC Labs, the Bayesian Consultancy. Book a call, or get in touch!Get early access to Alex's next live-cohort courses!Enroll in the Causal AI workshop, to learn live with Alex (15% off if you're a Patron of the show)Our&nbsp;theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and&nbsp;Mega Ran). Check out his awesome work!Visit our Patreon page to unlock exclusive Bayesian swag ;)Takeaways:Causal inference is crucial for understanding the impact of interventions in various fields.ChiRho is a causal probabilistic programming language that bridges mechanistic and data-driven models.ChiRho allows for easy manipulation of causal models and counterfactual reasoning.The design of ChiRho emphasizes modularity and extensibility for diverse applications.Causal inference requires careful consideration of assumptions and model structures.Real-world applications of causal inference can lead to significant insights in science and engineering.Collaboration and communication are key in translating causal questions into actionable models.The future of causal inference lies in integrating probabilistic programming with scientific discovery.Chapters:05:53 Bridging Mechanistic and Data-Driven Models09:13 Understanding Causal Probabilistic Programming12:10 ChiRho and Its Design Principles15:03 ChiRho’s Functionality and Use Cases17:55 Counterfactual Worlds and Mediation Analysis20:47 Efficient Estimation in ChiRho24:08 Future Directions for Causal AI50:21 Understanding the Do-Operator in Causal Inference56:45 ChiRho’s Role in Causal Inference and Bayesian Modeling01:01:36 Roadmap and Future Developments for ChiRho01:05:29 Real-World Applications of Causal Probabilistic Programming01:10:51 Challenges in Causal Inference Adoption01:11:50 The Importance of Causal Claims in Research01:18:11 Bayesian Approaches to Causal Inference01:22:08 Combining Gaussian Processes with Causal Inference01:28:27 Future Directions in Probabilistic Programming and Causal InferenceThank you to my Patrons for making this episode possible!Yusuke Saito, Avi Bryant, Ero Carrera, Giuliano Cruz, James Wade, Tradd Salvo, William Benton, James Ahloy, Robin Taylor,, Chad...

Get early access to Alex's next live-cohort courses!Today’s clip is from episode 140 of the podcast, with Ron Yurko.Alex and Ron discuss the challenges of model deployment, and the complexities of modeling player contributions in team sports like soccer and football.They emphasize the importance of understanding replacement levels, the Going Deep framework in football analytics, and the need for proper modeling of expected points. Additionally, they share insights on teaching Bayesian modeling to students and the difficulties they face in grasping the concepts of model writing and application.Get the full discussion here.Intro to Bayes Course (first 2 lessons free)Advanced Regression Course (first 2 lessons free)Our&nbsp;theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and&nbsp;Mega Ran). Check out his awesome work!Visit our Patreon page to unlock exclusive Bayesian swag ;)TranscriptThis is an automatic transcript and may therefore contain errors. Please get in touch if you're willing to correct them.

Get early access to Alex's next live-cohort courses!Proudly sponsored by PyMC Labs, the Bayesian Consultancy. Book a call, or get in touch!Intro to Bayes Course (first 2 lessons free)Advanced Regression Course (first 2 lessons free)Our&nbsp;theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and&nbsp;Mega Ran). Check out his awesome work!Visit our Patreon page to unlock exclusive Bayesian swag ;)Takeaways:Teaching students to write out their own models is crucial.Developing a sports analytics portfolio is essential for aspiring analysts.Modeling expectations in sports analytics can be misleading.Tracking data can significantly improve player performance models.Ron encourages students to engage in active learning through projects.The importance of understanding the dependency structure in data is vital.Ron aims to integrate more diverse sports analytics topics into his teaching.Chapters:03:51 The Journey into Sports Analytics15:20 The Evolution of Bayesian Statistics in Sports26:01 Innovations in NFL WAR Modeling39:23 Causal Modeling in Sports Analytics46:29 Defining Replacement Levels in Sports48:26 The Going Deep Framework and Big Data in Football52:47 Modeling Expectations in Football Data55:40 Teaching Statistical Concepts in Sports Analytics01:01:54 The Importance of Model Building in Education01:04:46 Statistical Thinking in Sports Analytics01:10:55 Innovative Research in Player Movement01:15:47 Exploring Data Needs in American Football01:18:43 Building a Sports Analytics PortfolioThank you to my Patrons for making this episode possible!Yusuke Saito, Avi Bryant, Ero Carrera, Giuliano Cruz, James Wade, Tradd Salvo, William Benton, James Ahloy, Robin Taylor,, Chad Scherrer, Zwelithini Tunyiswa, Bertrand Wilden, James Thompson, Stephen Oates, Gian Luca Di Tanna, Jack Wells, Matthew Maldonado, Ian Costley, Ally Salim, Larry Gill, Ian Moran, Paul Oreto, Colin Caprani, Colin Carroll, Nathaniel Burbank, Michael Osthege, Rémi Louf, Clive Edelsten, Henri Wallen, Hugo Botha, Vinh Nguyen, Marcin Elantkowski, Adam C. Smith, Will Kurt, Andrew Moskowitz, Hector Munoz, Marco Gorelli, Simon Kessell,...

Today’s clip is from episode 139 of the podcast, with with Max Balandat.Alex and Max discuss the integration of BoTorch with PyTorch, exploring its applications in Bayesian optimization and Gaussian processes. They highlight the advantages of using GPyTorch for structured matrices and the flexibility it offers for research. The discussion also covers the motivations behind building BoTorch, the importance of open-source culture at Meta, and the role of PyTorch in modern machine learning.Get the full discussion here.Attend Alex's tutorial at PyData Berlin: A Beginner's Guide to State Space Modeling&nbsp;Intro to Bayes Course (first 2 lessons free)Advanced Regression Course (first 2 lessons free)Our&nbsp;theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and&nbsp;Mega Ran). Check out his awesome work!Visit our Patreon page to unlock exclusive Bayesian swag ;)TranscriptThis is an automatic transcript and may therefore contain errors. Please get in touch if you're willing to correct them.