Akarsh Kumar

Effective Mutation Rate Adaptation through Group Elite Selection

Accepted at GECCO 2022

Evolutionary algorithms are sensitive to the mutation rate (MR); no single value of this parameter works well across domains. Self-adaptive MR approaches have been proposed but they tend to be brittle; for example, they sometimes decay the MR to zero, thus halting evolution. To make self-adaptive MR robust, this paper introduces the Group Elite Selection of Mutation Rates (GESMR) algorithm. GESMR co-evolves a population of solutions and a population of MRs, such that each MR is assigned to a group of solutions. The resulting best mutational change in the group, instead of average mutational change, is used for MR selection during evolution, thus avoiding the vanishing MR problem. With the same number of function evaluations and with almost no overhead, GESMR converges faster and to better solutions than previous approaches on a wide range of continuous test optimization problems. GESMR also scales well to high-dimensional neuroevolution for supervised image-classification tasks and for reinforcement learning control tasks. Analysis of the distribution of function changes during mutation explains why self-adaptation is prone to premature convergence and how GESMR overcomes this issue. Empirically, GESMR produces MRs that are optimal in the long-term, as demonstrated through a comprehensive look-ahead grid search. GESMR and the analysis have theoretical and practical implications for the fields of artificial life and evolutionary computation.

Physically Plausible Pose Refinement using Fully Differentiable Forces

Presented at EPIC@CVPR 2021

All hand-object interaction is controlled by forces that the two bodies exert on each other, but little work has been done in modeling these underlying forces when doing pose and contact estimation from RGB/RGB-D data. Given the pose of the hand and object from any pose estimation system, we propose an end-to-end differentiable model that refines pose estimates by learning the forces experienced by the object at each vertex in its mesh. By matching the learned net force to an estimate of net force based on finite differences of position, this model is able to find forces that accurately describe the movement of the object, while resolving issues like mesh interpenetration and lack of contact. Evaluating on the ContactPose dataset, we show this model successfully corrects poses and finds contact maps that better match the ground truth, despite not using any RGB or depth image data.

FIRM Project

Presented at Intel ISEF 2018

alt text 

In a vastly increasing energy dependent society, there is need for the transportation of energy quickly and efficiently. Laser power beaming and transferring energy with fiber optic cables are two methods that accomplish this. The low efficiency of converting laser light energy to electrical energy via photovoltaic cells is a problem. This project aims to optimize the efficiency of a PV cell for laser light using a theoretical approach. Shockley and Queisser’s methods of a optimizing a photovoltaic cell for sunlight were used, and the blackbody spectrum of the sun was replaced with a laser spectrum. The mathematical calculations were done for this laser-PV cell system given the condition that a photon with an energy above the band gap of the PV cell may create an exciton with energy equivalent to the band gap. By considering the current-voltage characteristics of the cell, the practical efficiency equation was found as a function of the laser frequency, the band gap frequency, and 5 other independent variables. It was used to calculate the efficiencies of over 600 laser-PV cell combinations and to create a color map representing the three dimensional efficiency function. Most efficiencies were low, however, when the laser frequency was only slightly over the band gap frequency, practical efficiencies of over 80% were reached. These findings can be the start of experimentation using the laser-cell combinations highlighted in the paper and can result in a new era of energy transfer.

Advisor: Dr. Brian Monson