Almost Uniform Sampling From Neural Networks

12/10/2019
by   Changlong Wu, et al.
0

Given a length n sample from R^d and a neural network with a fixed architecture with W weights, k neurons, linear threshold activation functions, and binary outputs on each neuron, we study the problem of uniformly sampling from all possible labelings on the sample corresponding to different choices of weights. We provide an algorithm that runs in time polynomial both in n and W such that any labeling appears with probability at least (W/2ekn)^W for W<n. For a single neuron, we also provide a random walk based algorithm that samples exactly uniformly.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
07/23/2020

Sampling connected subgraphs: nearly-optimal mixing time bounds, nearly-optimal ε-uniform sampling, and perfect uniform sampling

We study the connected subgraph sampling problem: given an integer k ≥ 3...
research
02/16/2022

Learning a Single Neuron for Non-monotonic Activation Functions

We study the problem of learning a single neuron 𝐱↦σ(𝐰^T𝐱) with gradient...
research
07/03/2023

First Steps towards a Runtime Analysis of Neuroevolution

We consider a simple setting in neuroevolution where an evolutionary alg...
research
11/27/2018

Knots in random neural networks

The weights of a neural network are typically initialized at random, and...
research
03/05/2022

On the Error of Random Sampling: Uniformly Distributed Random Points on Parametric Curves

Given a parametric polynomial curve γ:[a,b]→ℝ^n, how can we sample a ran...
research
12/07/2021

fastball: A fast algorithm to sample binary matrices with fixed marginals

Many applications require randomly sampling binary graphs with fixed deg...
research
01/18/2021

Stable Recovery of Entangled Weights: Towards Robust Identification of Deep Neural Networks from Minimal Samples

In this paper we approach the problem of unique and stable identifiabili...

Please sign up or login with your details

Forgot password? Click here to reset