POPL 2025 (series) / LAFI 2025 (series) / LAFI 2025 /
Partially Evaluating Higher-Order Probabilistic Programs without Stochastic Recursion to Graphical Models (Remote)
Probabilistic programming languages are often distinguished by their support for higher-order functions. We demonstrate that a more nuanced distinction is possible if the focus is set on whether the program induces stochastic control flow in recursions used by higher-order functions. In the case where higher-order functionality is only applied to deterministically known values we can compile a higher-order program into a probabilistic graphical model without restricting expressivity. We provide the implementation of this approach in our probabilistic programming compiler that previously only supported first-order functional abstractions.
Sun 19 JanDisplayed time zone: Mountain Time (US & Canada) change
Sun 19 Jan
Displayed time zone: Mountain Time (US & Canada) change
16:00 - 17:30 | |||
16:00 15mTalk | Partially Evaluating Higher-Order Probabilistic Programs without Stochastic Recursion to Graphical Models (Remote) LAFI | ||
16:16 15mTalk | State Space Model Programming in Turing.jl LAFI Tim Hargreaves Department of Engineering, University of Cambridge, Qing Li Department of Engineering, University of Cambridge, Charles Knipp Federal Reserve Board of Governors, USA, Frederic Wantiez , Simon J. Godsill Department of Engineering, University of Cambridge, Hong Ge University of Cambridge File Attached | ||
16:32 55mOther | Poster session LAFI | ||
17:28 2mDay closing | Closing remarks LAFI Atılım Güneş Baydin University of Oxford |