Optimizing Optimizers: Regret-optimal gradient descent algorithms

• URL: http://arxiv.org/abs/2101.00041v2
• Date: Tue, 19 Jan 2021 22:50:56 GMT
• Title: Optimizing Optimizers: Regret-optimal gradient descent algorithms
• Authors: Philippe Casgrain, Anastasis Kratsios
• Abstract summary: We study the existence, uniqueness and consistency of regret-optimal algorithms. By providing first-order optimality conditions for the control problem, we show that regret-optimal algorithms must satisfy a specific structure in their dynamics. We present fast numerical methods for approximating them, generating optimization algorithms which directly optimize their long-term regret.
• Score: 9.89901717499058
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The need for fast and robust optimization algorithms are of critical importance in all areas of machine learning. This paper treats the task of designing optimization algorithms as an optimal control problem. Using regret as a metric for an algorithm's performance, we study the existence, uniqueness and consistency of regret-optimal algorithms. By providing first-order optimality conditions for the control problem, we show that regret-optimal algorithms must satisfy a specific structure in their dynamics which we show is equivalent to performing dual-preconditioned gradient descent on the value function generated by its regret. Using these optimal dynamics, we provide bounds on their rates of convergence to solutions of convex optimization problems. Though closed-form optimal dynamics cannot be obtained in general, we present fast numerical methods for approximating them, generating optimization algorithms which directly optimize their long-term regret. Lastly, these are benchmarked against commonly used optimization algorithms to demonstrate their effectiveness.

Related papers

• A Pure Quantum Approximate Optimization Algorithm Based on CNR Operation [0.0]
  We propose a general-purpose pure quantum approximate optimization algorithm. The algorithm is constructed to a $p$-level divide-and-conquer structure. We show the algorithm performance in detail when the required qubits number of the two optimization problems is 10.
  arXiv  Detail & Related papers  (2023-10-27T06:54:39Z)
• Performance Evaluation of Evolutionary Algorithms for Analog Integrated Circuit Design Optimisation [0.0]
  An automated sizing approach for analog circuits is presented in this paper. A targeted search of the search space has been implemented using a particle generation function and a repair-bounds function. The algorithms are tuned and modified to converge to a better optimal solution.
  arXiv  Detail & Related papers  (2023-10-19T03:26:36Z)
• First-Order Dynamic Optimization for Streaming Convex Costs [0.0]
  We develop an approach to track the optimal solution with a bounded error. Our algorithm is executed only by using the first-order derivatives of the cost function.
  arXiv  Detail & Related papers  (2023-10-11T22:41:00Z)
• Accelerating Cutting-Plane Algorithms via Reinforcement Learning Surrogates [49.84541884653309]
  A current standard approach to solving convex discrete optimization problems is the use of cutting-plane algorithms. Despite the existence of a number of general-purpose cut-generating algorithms, large-scale discrete optimization problems continue to suffer from intractability. We propose a method for accelerating cutting-plane algorithms via reinforcement learning.
  arXiv  Detail & Related papers  (2023-07-17T20:11:56Z)
• Accelerated First-Order Optimization under Nonlinear Constraints [73.2273449996098]
  We exploit between first-order algorithms for constrained optimization and non-smooth systems to design a new class of accelerated first-order algorithms. An important property of these algorithms is that constraints are expressed in terms of velocities instead of sparse variables.
  arXiv  Detail & Related papers  (2023-02-01T08:50:48Z)
• An Empirical Evaluation of Zeroth-Order Optimization Methods on AI-driven Molecule Optimization [78.36413169647408]
  We study the effectiveness of various ZO optimization methods for optimizing molecular objectives. We show the advantages of ZO sign-based gradient descent (ZO-signGD) We demonstrate the potential effectiveness of ZO optimization methods on widely used benchmark tasks from the Guacamol suite.
  arXiv  Detail & Related papers  (2022-10-27T01:58:10Z)
• Provably Faster Algorithms for Bilevel Optimization [54.83583213812667]
  Bilevel optimization has been widely applied in many important machine learning applications. We propose two new algorithms for bilevel optimization. We show that both algorithms achieve the complexity of $mathcalO(epsilon-1.5)$, which outperforms all existing algorithms by the order of magnitude.
  arXiv  Detail & Related papers  (2021-06-08T21:05:30Z)
• Convergence of adaptive algorithms for weakly convex constrained optimization [59.36386973876765]
  We prove the $mathcaltilde O(t-1/4)$ rate of convergence for the norm of the gradient of Moreau envelope. Our analysis works with mini-batch size of $1$, constant first and second order moment parameters, and possibly smooth optimization domains.
  arXiv  Detail & Related papers  (2020-06-11T17:43:19Z)
• Private Stochastic Convex Optimization: Optimal Rates in Linear Time [74.47681868973598]
  We study the problem of minimizing the population loss given i.i.d. samples from a distribution over convex loss functions. A recent work of Bassily et al. has established the optimal bound on the excess population loss achievable given $n$ samples. We describe two new techniques for deriving convex optimization algorithms both achieving the optimal bound on excess loss and using $O(minn, n2/d)$ gradient computations.
  arXiv  Detail & Related papers  (2020-05-10T19:52:03Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.