Article Source

• Non-linear Causal Discovery for Addtive Noise Models with Multiple Latent Confounders

Non-linear Causal Discovery for Addtive Noise Models with Multiple Latent Confounders

Abstract

Hello lovely people!! This is a technical talk about my research work in Causal Science and AI: https://xuanzhichen.github.io/work/pa… Thank you for watching my video :)

The following shows the common introduction for this work.

“Could we teach AI in brain science spectrum to manoeuvre causation via the specific identification entailed by data? How should we further appreciate ‘causal structures’ underneath the data in a complicate learning environment? An environment in which ‘generic data relations’ are prone to be non-linear, and even impacts from the multiple unknown factors are persisting.

Existing solutions towards the issue might be either theoretically elusive in formal representation or notoriously difficult in algorithmic computation. Such motivations have driven us to a theory-guided and effective causal discovery algorithm.”