Inference compilation and universal probabilistic programming
We introduce a method for using deep neural networks to amortize the cost of inference in models from the family induced by universal probabilistic programming languages, establishing a framework that combines the strengths of probabilistic programming and deep learning methods. We call what we do “...
| Asıl Yazarlar: | , , |
|---|---|
| Materyal Türü: | Conference item |
| Baskı/Yayın Bilgisi: |
Journal of Machine Learning Research
2017
|
| _version_ | 1826292746135011328 |
|---|---|
| author | Le, T Baydin, A Wood, F |
| author_facet | Le, T Baydin, A Wood, F |
| author_sort | Le, T |
| collection | OXFORD |
| description | We introduce a method for using deep neural networks to amortize the cost of inference in models from the family induced by universal probabilistic programming languages, establishing a framework that combines the strengths of probabilistic programming and deep learning methods. We call what we do “compilation of inference” because our method transforms a denotational specification of an inference problem in the form of a probabilistic program written in a universal programming language into a trained neural network denoted in a neural network specification language. When at test time this neural network is fed observational data and executed, it performs approximate inference in the original model specified by the probabilistic program. Our training objective and learning procedure are designed to allow the trained neural network to be used as a proposal distribution in a sequential importance sampling inference engine. We illustrate our method on mixture models and Captcha solving and show significant speedups in the efficiency of inference. |
| first_indexed | 2024-03-07T03:19:27Z |
| format | Conference item |
| id | oxford-uuid:b6f09ebf-9f66-433b-b71b-3980b577047c |
| institution | University of Oxford |
| last_indexed | 2024-03-07T03:19:27Z |
| publishDate | 2017 |
| publisher | Journal of Machine Learning Research |
| record_format | dspace |
| spelling | oxford-uuid:b6f09ebf-9f66-433b-b71b-3980b577047c2022-03-27T04:44:43ZInference compilation and universal probabilistic programmingConference itemhttp://purl.org/coar/resource_type/c_5794uuid:b6f09ebf-9f66-433b-b71b-3980b577047cSymplectic Elements at OxfordJournal of Machine Learning Research2017Le, TBaydin, AWood, FWe introduce a method for using deep neural networks to amortize the cost of inference in models from the family induced by universal probabilistic programming languages, establishing a framework that combines the strengths of probabilistic programming and deep learning methods. We call what we do “compilation of inference” because our method transforms a denotational specification of an inference problem in the form of a probabilistic program written in a universal programming language into a trained neural network denoted in a neural network specification language. When at test time this neural network is fed observational data and executed, it performs approximate inference in the original model specified by the probabilistic program. Our training objective and learning procedure are designed to allow the trained neural network to be used as a proposal distribution in a sequential importance sampling inference engine. We illustrate our method on mixture models and Captcha solving and show significant speedups in the efficiency of inference. |
| spellingShingle | Le, T Baydin, A Wood, F Inference compilation and universal probabilistic programming |
| title | Inference compilation and universal probabilistic programming |
| title_full | Inference compilation and universal probabilistic programming |
| title_fullStr | Inference compilation and universal probabilistic programming |
| title_full_unstemmed | Inference compilation and universal probabilistic programming |
| title_short | Inference compilation and universal probabilistic programming |
| title_sort | inference compilation and universal probabilistic programming |
| work_keys_str_mv | AT let inferencecompilationanduniversalprobabilisticprogramming AT baydina inferencecompilationanduniversalprobabilisticprogramming AT woodf inferencecompilationanduniversalprobabilisticprogramming |