CaltechTHESIS
  A Caltech Library Service

Automated Macro-scale Causal Hypothesis Formation Based on Micro-scale Observation

Citation

Chalupka, Krzysztof (2017) Automated Macro-scale Causal Hypothesis Formation Based on Micro-scale Observation. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/Z9MW2F4P. http://resolver.caltech.edu/CaltechTHESIS:11292016-204528802

Abstract

This book introduces new concepts at the intersection of machine learning, causal inference and philosophy of science: the macrovariable cause and effect. Methods for learning such from microvariable data are introduced. The learning process proposes a minimal number of guided experiments that recover the macrovariable cause from observational data.

Mathematical definitions of a micro- and macro- scale manipulation, an observational and causal partition, and a subsidiary variable are given. These concepts provide a link to previous work in causal inference and machine learning.

The main theoretical result is the Causal Coarsening Theorem, a new insight into the measure-theoretic structure of probability spaces and structural equation models. The theorem provides grounds for automatic causal hypothesis formation from data. Other results concern the minimality and sufficiency of representations created in accordance with the theorem.

Finally, this book proposes the first algorithms for supervised and unsupervised causal macrovariable discovery. These algorithms bridge large-scale, multidimensional machine learning and causal inference. In an application to climate science, the algorithms re-discover a known causal mechanism as a viable causal hypothesis. In a psychophysical experiment, the algorithms learn to minimally change visual stimuli to achieve a desired effect on human perception.

Item Type:Thesis (Dissertation (Ph.D.))
Subject Keywords:machine learning, causality, causal inference, causal discovery, causation
Degree Grantor:California Institute of Technology
Division:Engineering and Applied Science
Major Option:Computation and Neural Systems
Thesis Availability:Public (worldwide access)
Research Advisor(s):
  • Perona, Pietro (advisor)
  • Eberhardt, Frederick (co-advisor)
Thesis Committee:
  • Perona, Pietro (chair)
  • Eberhardt, Frederick D.
  • Yue, Yisong
  • Tsao, Doris Y.
Defense Date:7 November 2016
Record Number:CaltechTHESIS:11292016-204528802
Persistent URL:http://resolver.caltech.edu/CaltechTHESIS:11292016-204528802
DOI:10.7907/Z9MW2F4P
Related URLs:
URLURL TypeDescription
http://vision.caltech.edu/˜kchalupk/code.htmlAuthorImplementations of the algorithms in the thesis
Default Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:9987
Collection:CaltechTHESIS
Deposited By: Krzysztof Chalupka
Deposited On:12 Dec 2016 23:04
Last Modified:12 Dec 2016 23:04

Thesis Files

[img]
Preview
PDF - Final Version
See Usage Policy.

11Mb

Repository Staff Only: item control page