Combining Bayes and Graph-based Causal Inference
Details
šļø Speaker: Robert Ness | ā° Time: 17:00 UTC / 9am PT / 12pm ET / 6pm Berlin
Graphical causal inference and probabilistic programming share much history. For example, directed probabilistic graphical models were early versions of causal models and d-separation (graphical criteria for conditional independence) provided fundamentals for the do-calculus. Also, directed graphical models drove advancements in Bayesian inference algorithms and were the precursors of probabilistic programming languages like PyTorch. Further, both causal models and probabilistic programming favor explicitly modeling the data generating process. Yet, despite these commonalities, graphical causal inference and probabilistic programming have evolved into separate communities with little cross-talk beyond Bayesian inference of parameters in causal estimators. In this seminar, we discuss how to do causal graphical modeling with probabilistic programming, as well as tools and design patterns for doing so.
š Resources
- Causal AI Book: https://www.manning.com/books/causal-ai
- Related Paper: https://arxiv.org/abs/2102.06626
- Probabilistic Machine Learning Workshop: https://www.altdeep.ai/p/probml
- Causal Modeling in Machine Learning Workshop: https://www.altdeep.ai/p/causalml
š Outline of Talk / Agenda:
- 5 min: Intro to PyMC Labs and speakers
- 45 min: Presentation, panel discussion
- 10 min: Q&A
š¼ About the speaker:
- Robert Ness
Researcher at Microsoft Research, where he focuses on causal reasoning, deep probabilistic modeling, language models and programming languages. He is author of the book Causal AI, and founder of AI learning platform Altdeep.ai. He has worked as a research engineer and received his Ph.D. in statistics from Purdue University. He is a Johns Hopkins SAIS alumnus.
š Connect with Robert Ness:
š LinkedIn: https://www.linkedin.com/in/osazuwa/
š Twitter: https://twitter.com/osazuwa
š GitHub: https://github.com/altdeep/causalML
š MSR: https://www.microsoft.com/en-us/research/people/robertness/
š¼ About the Host:
- Dr. Thomas Wiecki (PyMC Labs)
Dr. Thomas Wiecki is an author of PyMC, the leading platform for statistical data science. To help businesses solve some of their trickiest data science problems, he assembled a world-class team of Bayesian modelers and founded PyMC Labs -- the Bayesian consultancy. He did his PhD at Brown University studying cognitive neuroscience.
š Connect with Thomas Wiecki:
š Website: https://www.pymc-labs.com/
š GitHub: https://github.com/twiecki
š Twitter: https://twitter.com/twiecki
š Blog posts: https://twiecki.io/
š Code of Conduct:
Please note that participants are expected to abide by PyMC's Code of Conduct.
š Connecting with PyMC Labs:
š„ LinkedIn: https://www.linkedin.com/company/pymc-labs/
š¦ Twitter: https://twitter.com/pymc_labs
š„ YouTube: https://www.youtube.com/c/PyMCLabs
š¤ Meetup: https://www.meetup.com/pymc-labs-online-meetup/
š Connecting with PyMC Open Source:
š¬ Q&A/Discussion: https://discourse.pymc.io
š GitHub: https://github.com/pymc-devs/pymc
š¼ LinkedIn: https://www.linkedin.com/company/pymc/mycompany
š„ Twitter: https://twitter.com/pymc_devs
šŗ YouTube: https://www.youtube.com/c/PyMCDevelopers
š Meetup: https://www.meetup.com/pymc-online-meetup/