Causal Inductive Synthesis Corpus
Abstract
We introduce the Causal Inductive Synthesis Corpus (CISC) -- a manually constructed collection of interactive domains. CISC domains abstract core causal concepts present in real world mechanisms and environments. We formulate two synthesis challenges of causal model discovery: the passive discovery of a model of a CISC domain from observed data, and active discovery while interacting with the domain. CISC problems are expressed in Autumn, a Turing-complete programming language for specifying causal probabilistic models. Autumn allows succinct expression for models that vary dynamically through time, respond to external input, have internal state and memory, exhibit probabilistic non-determinism, and have complex causal dependencies between variables.
Cite
Text
Tavares et al. "Causal Inductive Synthesis Corpus." NeurIPS 2020 Workshops: CAP, 2020.Markdown
[Tavares et al. "Causal Inductive Synthesis Corpus." NeurIPS 2020 Workshops: CAP, 2020.](https://mlanthology.org/neuripsw/2020/tavares2020neuripsw-causal/)BibTeX
@inproceedings{tavares2020neuripsw-causal,
title = {{Causal Inductive Synthesis Corpus}},
author = {Tavares, Zenna and Das, Ria and Weeks, Elizabeth and Lin, Kate and Tenenbaum, Joshua B. and Solar-Lezama, Armando},
booktitle = {NeurIPS 2020 Workshops: CAP},
year = {2020},
url = {https://mlanthology.org/neuripsw/2020/tavares2020neuripsw-causal/}
}