arXiv subjects

John Maxwell

Publications and source records attributed to John Maxwell.

3 recordsLinked to original sources

arXiv2023

An optimization-based approach to automated design

We propose a model-based, automated, bottom-up approach for design, which is applicable to various physical domains, but in this work we focus on the electrical domain. This bottom-up approach is based on a meta-topology in which each link is described by a universal component that can be instantiated as basic components (e.g., resistors, capacitors) or combinations of basic components via discrete switches. To address the combinatorial explosion often present in mixed-integer optimization problems, we present two algorithms. In the first algorithm, we convert the discrete switches into continuous switches that are physically realizable and formulate a parameter optimization problem that learns the component and switch parameters while inducing design sparsity through an $L_1$ regularization term. The second algorithm uses a genetic-like approach with selection and mutation steps guided by ranking of requirements costs, combined with continuous optimization for generating optimal parameters. We improve the time complexity of the optimization problem in both algorithms by reconstructing the model when components become redundant and by simplifying topologies through collapsing components and removing disconnected ones. To demonstrate the efficacy of these algorithms, we apply them to the design of various electrical circuits.

math.OC↗

arXiv2022

Improving the Efficiency of Gradient Descent Algorithms Applied to Optimization Problems with Dynamical Constraints

We introduce two block coordinate descent algorithms for solving optimization problems with ordinary differential equations (ODEs) as dynamical constraints. The algorithms do not need to implement direct or adjoint sensitivity analysis methods to evaluate loss function gradients. They results from reformulation of the original problem as an equivalent optimization problem with equality constraints. The algorithms naturally follow from steps aimed at recovering the gradient-decent algorithm based on ODE solvers that explicitly account for sensitivity of the ODE solution. In our first proposed algorithm we avoid explicitly solving the ODE by integrating the ODE solver as a sequence of implicit constraints. In our second algorithm, we use an ODE solver to reset the ODE solution, but no direct are adjoint sensitivity analysis methods are used. Both algorithm accepts mini-batch implementations and show significant efficiency benefits from GPU-based parallelization. We demonstrate the performance of the algorithms when applied to learning the parameters of the Cucker-Smale model. The algorithms are compared with gradient descent algorithms based on ODE solvers endowed with sensitivity analysis capabilities, for various number of state size, using Pytorch and Jax implementations. The experimental results demonstrate that the proposed algorithms are at least 4x faster than the Pytorch implementations, and at least 16x faster than Jax implementations. For large versions of the Cucker-Smale model, the Jax implementation is thousands of times faster than the sensitivity analysis-based implementation. In addition, our algorithms generate more accurate results both on training and test data. Such gains in computational efficiency is paramount for algorithms that implement real time parameter estimations, such as diagnosis algorithms.

cs.LG↗

arXiv2020

Characterizing an Analogical Concept Memory for Architectures Implementing the Common Model of Cognition

Architectures that implement the Common Model of Cognition - Soar, ACT-R, and Sigma - have a prominent place in research on cognitive modeling as well as on designing complex intelligent agents. In this paper, we explore how computational models of analogical processing can be brought into these architectures to enable concept acquisition from examples obtained interactively. We propose a new analogical concept memory for Soar that augments its current system of declarative long-term memories. We frame the problem of concept learning as embedded within the larger context of interactive task learning (ITL) and embodied language processing (ELP). We demonstrate that the analogical learning methods implemented in the proposed memory can quickly learn a diverse types of novel concepts that are useful not only in recognition of a concept in the environment but also in action selection. Our approach has been instantiated in an implemented cognitive system \textsc{Aileen} and evaluated on a simulated robotic domain.

cs.AI↗