Research map / Learning theory and algorithms / Sample complexity
Learning a 1-layer conditional generative model in total variation
Ajil Jalal, Justin Kang, Ananya Uppal, Kannan Ramchandran, Eric Price
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace x ~ N(0, I_k) with x drawn from i.i.d. standard Laplace | MLE for 1-layer conditional generative model · synthetic, k=5 and k=20 | TV distance | No clear effect |
| increase number of samples n | MLE · synthetic, TV vs n | TV distance | Improved |
| increase input dimension k | MLE · synthetic, n=10^3 and 10^4 | TV distance | Hurt |
| make bias b more negative so more samples are truncated, MLE using truncated samples | MLE · d=3, x=1, y=phi(eta+b1) | TV distance | Improved |
| use baseline of [35] instead of MLE as bias b decreases | baseline of [35] · d=3, truncated outputs | TV distance | Hurt |
| increase condition number of covariance Sigma* | MLE · d=3, b=1 | TV distance | No clear effect |
| increase condition number of covariance Sigma* | baseline of [35] · d=3, b=1 | TV distance | Hurt |
Closest papers by meaning
- A Likelihood Based Approach to Distribution Regression Using Conditional Deep Generative Models (ICML 2025 · similarity 0.88)
- A Theory for Conditional Generative Modeling on Multiple Data Sources (ICML 2025 · similarity 0.87)
- Learning Distributions Generated by One-Layer ReLU Networks (NeurIPS 2019 · similarity 0.86)
- Generative Conditional Distributions by Neural (Entropic) Optimal Transport (ICML 2024 · similarity 0.85)
- Towards Theoretical Understandings of Self-Consuming Generative Models (ICML 2024 · similarity 0.85)
- Attack Deterministic Conditional Image Generative Models for Diverse and Controllable Generation (AAAI 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
