Learning Bayesian Statistics

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 theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and Mega Ran). Check out his awesome work!Visit our Patreon page to unlock exclusive Bayesian swag ;)Thank you to my Patrons for making this episode possible!Yusuke Saito, Avi Bryant, 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, Bradley Rode, Patrick Kelley, Rick Anderson, Casper de Bruin, Michael Hankin, Cameron Smith, Tomáš Frýda, Ryan Wesslen, Andreas Netti, Riley King, Yoshiyuki Hamajima, Sven De Maeyer, Michael DeCrescenzo, Fergal M, Mason Yahr, Naoya Kanai, Aubrey Clayton, Omri Har Shemesh, Scott Anthony Robson, Robert Yolken, Or Duek, Pavel Dusek, Paul Cox, Andreas Kröpelin, Raphaël R, Nicolas Rode, Gabriel Stechschulte, Arkady, Kurt TeKolste, Marcus Nölke, Maggi Mackintosh, Grant Pezzolesi, Joshua Meehl, Javier Sabio, Kristian Higgins, Matt Rosinski, Luis Fonseca, Dante Gates, Matt Niccolls, Maksim Kuznecov, Michael Thomas, Luke Gorrie, Cory Kiser, Julio, Edvin Saveljev, Frederick Ayala, Jeffrey Powell, Gal Kampel, Adan Romero, Will Geary, Blake Walters, Jonathan Morgan, Francesco Madrisotti, Ivy Huang, Gary Clarke, Robert Flannery, Rasmus Hindström, Stefan, Corey Abshire, Mike Loncaric, David McCormick, Ronald Legere, Sergio Dolia, Michael Cao, Yiğit Aşık, Suyog Chandramouli and Guillaume Berthon.Takeaways:AI is reshaping the workplace, but we're still in early stages.Networking is crucial for job applications in top firms.AI tools can augment work but are not replacements for skilled labor.Understanding the tech landscape requires continuous learning.Timing and cultural readiness are key for tech innovations.Expertise can be gained without formal education.Bayesian statistics is a valuable skill for tech professionals.The importance of personal branding in the job market. You just need to know 1% more than the person you're talking to.Sharing knowledge can elevate your status within a company.Embracing chaos in tech can create new opportunities.Investing in people leads to a more engaged workforce.Navigating corporate culture requires understanding your role and relationships.M&A trends in AI reflect the evolving landscape of technology.High compensation packages are not a new phenomenon in tech.Career growth often requires stepping outside your comfort zone.Soft skills are essential for effective communication in the workplace.Understanding the dynamics of M&A can provide insights into industry trends. AI is creating real economic value in customer service.Speculative activity often overshadows real economic activity in tech.Memorable M&A experiences can have a profound impact on people's lives.Chapters:10:39 The Impact of AI on Work and Culture18:11 Understanding the AI Revolution30:05 Career Advice in the Age of AI38:08 Innovative Company Culture and Experimentation41:04 Interview Dynamics and Performance Bias42:29 Augmenting Work with AI and Learning46:46 Navigating Organizational Boundaries51:33 The Importance of Soft Skills in Tech01:01:21 Mergers and Acquisitions in the Tech Industry01:15:08 The Reality of Tech Salaries01:18:18 The Impact of AI on Customer Service01:20:36 Speculative vs. Real Economic Activity in Tech01:22:33 The Cycle of Tech Booms and Busts01:24:22 Heartfelt Stories from the M&A World01:31:36 A Personal Vendetta Against CancerLinks from the show:Jordan on YouTube: https://www.youtube.com/@SVICPodcastJordan on LinkedIn: https://www.linkedin.com/in/jwthib/Jordan on Twitter: https://twitter.com/jordansvicLBS #124 State Space Models & Structural Time Series, with Jesse Grabowski: https://learnbayesstats.com/episode/124-state-space-models-structural-time-series-jesse-grabowskiUsing AI Tools to Automate Busywork and Scale Productivity | Sully Omar: https://www.youtube.com/watch?v=sk9YkYOZH8YSupercommunicators – How to Unlock the Secret Language of Connection: https://www.amazon.com/Supercommunicators-Unlock-Secret-Language-Connection/dp/0593243919The Charisma Myth – How Anyone Can Master the Art and Science of Personal Magnetism: https://www.amazon.com/Charisma-Myth-Science-Personal-Magnetism/dp/1591845947The 48 Laws of Power: https://www.amazon.com/48-Laws-Power-Robert-Greene/dp/0140280197/There Will Be Blood: https://www.imdb.com/title/tt0469494/How to Invent Everything – A Survival Guide for the Stranded Time Traveler: https://www.amazon.com/How-Invent-Everything-Survival-Stranded/dp/073522014XTranscriptThis is an automatic transcript and may therefore contain errors. Please get in touch if you're willing to correct them.

Join Maurizio Filippone, a Bayesian machine learning researcher specializing in Gaussian processes, as he unpacks the magic of deep Gaussian processes. He explains how composing GPs enhances flexibility and offers insights into modeling complex data. Discover practical approximations for implementing Deep GPs in TensorFlow, and learn when to use them over traditional deep neural networks. Maurizio also shares how to map neural networks to GP-like behavior for better interpretability and uncertainty quantification. It's a fascinating dive into the future of machine learning!

Maurizio Filippone, an associate professor at KAUST and leader of the Bayesian Deep Learning Group, dives into the fascinating world of Bayesian function estimation. He explains why Gaussian Processes are still crucial for function estimation and how deep Gaussian Processes introduce flexibility for complex tasks. Maurizio discusses practical strategies like Monte Carlo Dropout for uncertainty quantification in neural networks, the trade-offs between model complexity and interpretability, and the role of Bayesian methods in modern generative models.

Christoph Bamberg, a health psychology researcher, dives into the intriguing world of Bayesian statistics and its applications in appetite regulation. He discusses how the framing of dietary claims affects cognition, revealing modest influences on performance. Christoph shares insights on the challenges of using Bayesian models, especially in small-sample studies, and emphasizes the importance of communication in health contexts. He also highlights the potential of positive framing in therapeutic settings, merging scientific research with practical implications.

In this discussion, Christoph Bamberg, a researcher in cognitive science and health psychology, explores the impact of Bayesian methods on nutrition science. He shares insights on how dietary framing can influence cognition, revealing that effects of intermittent fasting depend on context and individual rhythms. Christoph emphasizes the importance of clear definitions in research and how small effects can have significant public health implications. He also highlights the challenges of converting collaborators to Bayesian methods and offers advice for students diving into this complex field.

Gabriel Stechschulte, a Bayesian software developer known for his work with PyMCBART, dives into the re-implementation of Bayesian Additive Regression Trees (BART) in Rust. He discusses the technical hurdles and enhanced performance achieved through this project. Gabriel explains the value of BART in uncertainty quantification and how it contrasts with other tree-based methods. The conversation also covers practical aspects, like integrating BART with Python and balancing open-source contributions with a full-time job, all while exploring the innovative features of PyMCBART.

Gabriel Stechschulte is a software engineer specializing in Bayesian methods and optimization. He discusses the power of Bayesian Additive Regression Trees (BART) for uncertainty quantification and its re-implementation in Rust, enhancing performance for big data. Gabriel explores how BART contrasts with other models, its strengths in avoiding overfitting, and its integration into optimization frameworks for decision-making. He also emphasizes the importance of open-source communities, encouraging newcomers to contribute actively.

In this engaging discussion, Sam Witty, a researcher from the Cairo project, dives into the fascinating world of causal inference. He explains the differences between do-calculus and Cairo’s parametric Bayesian methods, and how regression discontinuity designs enable causal estimation. Sam also explores how Cairo automates the construction of interventions, providing users easy access to complex statistical tools. The talk highlights the significance of efficient estimators, making causal queries more accessible without needing extensive expertise.

In this engaging discussion, Sam Whitty, the founder of Sorbus AI and a pioneer in causal probabilistic programming, dives into the intricacies of causal inference. He explores his journey from engineering to developing ChiRho, a language that merges mechanistic and data-driven models. Listeners will learn about counterfactual reasoning, the significance of modular design, and practical applications in science and engineering. Sam emphasizes the need for collaboration in transforming causal questions into actionable insights, while also looking ahead at the future of causal AI.

In this discussion, Ron Yurko, an expert in sports analytics, shares insights on the complexities of modeling player contributions in soccer and football. He highlights the significance of understanding replacement levels and introduces the Going Deep framework for analyzing player performance. They touch on the challenges of teaching Bayesian modeling, particularly how students struggle with model writing. The conversation underscores the importance of using advanced tracking data for better predictions and the necessity of viewing entire distributions in utility function modeling.