Timezone: »

Reducing Causal Illusions through Deliberate Undersampling
Kseniya Solovyeva · David Danks · Mohammadsajad Abavisani · Sergey Plis
Event URL: https://openreview.net/forum?id=_qBTi54t-we »

Domain scientists interested in the causal mechanisms are usually limited by the frequency at which they can collect the measurements of social, physical, or biological systems.It is a reasonable assumption that higher frequency is more informative of the causal structure.This assumption is a strong driver for designing new, faster instruments.A task that is expensive and often impossible at the current state of technology.In this work, we show that counter to the intuition it is possible for causal systems to improve the estimation of causal graphs from undersampled time-series by augmenting the measurements with those collected at a rate slower than currently available.We present an algorithm able to take advantage of measurement time-scale graphs estimated from data at various sampling rates and lower the underdeterminacy of the system by reducing the equivalence size.We investigate the probability of cases in which deliberate undersampling yields a gain and the size of this gain.

Author Information

Kseniya Solovyeva (Georgia State University)
David Danks (Carnegie Mellon University)
Mohammadsajad Abavisani (Georgia Institute of Technology)
Mohammadsajad Abavisani

I am a third year Ph.D. student in Electrical and Computer Engineering Department of Georgia Institute of Technology. My area of research is Machine Learning, Deep Learning and Causal Learning I am a graduate research assistant in (TReNDS). My areas of interest are developing novel machine learning methods to extract causal relations among complex and big scale data (including fMRI), Self-Supervised Learning and Computer Vision.

Sergey Plis (TReNDS center, GSU)

More from the Same Authors