NOVAS: Non-convex Optimization via Adaptive Stochastic Search for End-to-End Learning and Control

2020/06/22 by Ioannis Exarchos, Exarchos, Ioannis, Marcus A. Pereira +5
Computer Science · #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #FOS: Electrical engineering #FOS: Mathematics #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine Learning and Algorithms #Optimization and Control (math.OC) #Stochastic Gradient Optimization Techniques #Systems and Control (eess.SY) #electronic engineering #information engineering

paper · pdf · doi:10.48550/arxiv.2006.11992

openalex publication_date 2020/06/22 · openalex created_date 2021/03/01 · openalex updated_date 2026/07/28

Abstract

In this work we propose the use of adaptive stochastic search as a building block for general, non-convex optimization operations within deep neural network architectures. Specifically, for an objective function located at some layer in the network and parameterized by some network parameters, we employ adaptive stochastic search to perform optimization over its output. This operation is differentiable and does not obstruct the passing of gradients during backpropagation, thus enabling us to incorporate it as a component in end-to-end learning. We study the proposed optimization module's properties and benchmark it against two existing alternatives on a synthetic energy-based structured prediction task, and further showcase its use in stochastic optimal control applications.

Citations

• On Differentiating Parameterized Argmin and Argmax Problems with Application to Bi-level Optimization
• Unrolled Generative Adversarial Networks
• Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
• MPC-Inspired Neural Network Policies for Sequential Decision Making
• Universal Planning Networks
• On First-Order Meta-Learning Algorithms
• Meta-Learning with Implicit Gradients
• Model-Predictive Control via Cross-Entropy and Gradient-Based Optimization

Related

• SANIA: Polyak-type Optimization Framework Leads to Scale Invariant Stochastic Algorithms
• Adaptive Strategies in Non-convex Optimization
• Learning to Optimize Neural Nets
• Stochastic Nonconvex Optimization with Large Minibatches
• Appropriate Learning Rates of Adaptive Learning Rate Optimization\n Algorithms for Training Deep Neural Networks