Timezone: »
Poster
PLUGIn: A simple algorithm for inverting generative models with recovery guarantees
Babhru Joshi · Xiaowei Li · Yaniv Plan · Ozgur Yilmaz
We consider the problem of recovering an unknown latent code vector under a known generative model. For a $d$layer deep generative network $\mathcal{G}:\mathbb{R}^{n_0}\rightarrow \mathbb{R}^{n_d}$ with ReLU activation functions, let the observation be $\mathcal{G}(x)+\epsilon$ where $\epsilon$ is noise. We introduce a simple novel algorithm, Partially Linearized Update for Generative Inversion (PLUGIn), to estimate $x$ (and thus $\mathcal{G}(x)$). We prove that, when weights are Gaussian and layer widths $n_i \gtrsim 5^i n_0$ (up to log factors), the algorithm converges geometrically to a neighbourhood of $x$ with high probability. Note the inequality on layer widths allows $n_i>n_{i+1}$ when $i\geq 1$. To our knowledge, this is the first such result for networks with some contractive layers. After a sufficient number of iterations, the estimation errors for both $x$ and $\mathcal{G}(x)$ are at most in the order of $\sqrt{4^dn_0/n_d} \\epsilon\$. Thus, the algorithm can denoise when the expansion ratio $n_d/n_0$ is large. Numerical experiments on synthetic data and real data are provided to validate our theoretical results and to illustrate that the algorithm can effectively remove artifacts in an image.
Author Information
Babhru Joshi (University of British Columbia)
Xiaowei Li (University of British Columbia)
Yaniv Plan (University of British Columbia)
Ozgur Yilmaz (The University of British Columbia)
Related Events (a corresponding poster, oral, or spotlight)

2021 Spotlight: PLUGIn: A simple algorithm for inverting generative models with recovery guarantees »
Dates n/a. Room
More from the Same Authors

2019 Poster: Global Guarantees for Blind Demodulation with Generative Priors »
Paul Hand · Babhru Joshi 
2018 Poster: Nearly tight sample complexity bounds for learning mixtures of Gaussians via sample compression schemes »
Hassan Ashtiani · Shai BenDavid · Nicholas Harvey · Christopher Liaw · Abbas Mehrabian · Yaniv Plan 
2018 Oral: Nearly tight sample complexity bounds for learning mixtures of Gaussians via sample compression schemes »
Hassan Ashtiani · Shai BenDavid · Nicholas Harvey · Christopher Liaw · Abbas Mehrabian · Yaniv Plan 
2016 Poster: Averagecase hardness of RIP certification »
Tengyao Wang · Quentin Berthet · Yaniv Plan