Michael Wieck-Sosa

I am a fifth-year PhD student in Statistics at Carnegie Mellon University.

I am very lucky to be advised by Aaditya Ramdas and Cosma Shalizi.

Here is my CV. You can reach me at mwiecksosa AT cmu FULLSTOP edu.

Interests

• Causality (invariant causal prediction, conditional independence testing, causal structure learning, mechanism discovery)
• Dependence (high-dimensional time series, time-varying networks, spatiotemporal data, nonlinear stochastic dynamics)
• Machine learning (deep neural networks, transformers, state space models, foundation models, large language models)
• Simulation-based inference (uncertainty quantification, identification, selection, scalability, random feature methods)
• Sequential anytime-valid inference (confidence sequences, strong approximations, beyond martingale dependence)

Research

Statistics journals

• The dynamic generalized covariance measure for conditional independence testing with nonstationary time series (with Michel Haddad and Aaditya Ramdas). Accepted for publication in the Journal of Business & Economic Statistics. arXiv. Slides. Code.
• Estimating dynamic models by matching random features (with Cosma Shalizi). Submitted to the Journal of the Royal Statistical Society, Series B. arXiv. Code.

Machine learning conferences

• Generative sequence modeling for infinite memory processes via predictive states (with Cosma Shalizi). Submitted to ICLR 2027. arXiv.
• Compressed simulation-based inference: neural posterior estimation for massive data via low-dimensional sketches (with Cosma Shalizi). Available upon request.

Interdisciplinary journals

• Dynamic models with p parameters are identified by 2p+1 random features (with Cosma Shalizi). Major revision at Physical Review E. arXiv. Code.

Current projects

• Distribution-uniform anytime-valid inference for time series (with Aaditya Ramdas).
• Anytime-valid invariance testing for time series (with Aaditya Ramdas).
• Simulation-based inference through random features (with Cosma Shalizi).
• Pretraining large predictive state space models (with Cosma Shalizi).

News

• October 2026: I’ll be attending the Workshop on Information-Theoretic Methods for Learning Dynamics in the Program on Image-Based Scientific Machine Learning for Theories of Biological Dynamics Across Scales at the NSF-Simons National Institute for Theory and Mathematics in Biology during February 8-12, 2027.
• June 2026: I gave an invited talk at the 9th International Workshop in Sequential Methodologies (IWSM 2026) at American University.
• April 2026: I was awarded the DeGroot-Goel Fellowship for 2026 by the CMU Statistics & Data Science Department faculty.
• April 2026: I successfully proposed my thesis. Topics: generative modeling and dependent data.
• October 2025: I presented a poster at the 2025 Workshop on Neural Simulation-Based Inference at Carnegie Mellon University.
• September 2024: I presented a poster at the 2024 NBER-NSF Time Series Conference at the University of Pennsylvania.