Explore the concept of identifiability in causal models and its applications to Perturb-seq data in this comprehensive lecture by Chandler Squires from Valence Labs. Delve into structural causal models, interventions, and learning techniques for unknown-target interventions and unobserved variables. Examine latent factor causal models, causal disentanglement models, and linear causal disentanglement via intervention. Gain insights into ongoing work in the field and participate in a Q&A session. Discover how these advanced concepts are applied to learning gene regulatory networks from Perturb-seq data and predicting the effects of novel perturbations.
Overview
Syllabus
- Identifiability Background
- Structural Causal Models
- Interventions
- Identifiability in Causality
- Learning From Unknown-Target Interventions
- Learning in the Presence of Unobserved Variables
- Treks
- Latent Factor Causal Models LFCMs
- Causal Disentanglement Models
- Linear Causal Disentanglement via Intervention
- Ongoing Work
- Q+A
Taught by
Valence Labs
Related Courses
-
Generative Language Models in Molecular Discovery: Regression Transformer, GT4SD and Beyond
-
Causal Representation Learning via Soft Interventions
-
Large-Scale Differentiable Causal Discovery of Factor Graphs - Lecture and Primer
-
Weakly Supervised Causal Representation Learning
-
Identifying Representations for Intervention Extrapolation
-
Hidden Variables in Linear Non-Gaussian Causal Models
