Engineering topics

DeCoste, D.

Publications and source records attributed to DeCoste, D..

12 records

NASA NTRS2006

Automated Knowledge Discovery from Simulators

In this paper, we explore one aspect of knowledge discovery from simulators, the landscape characterization problem, where the aim is to identify regions in the input/ parameter/model space that lead to a particular output behavior. Large-scale numerical simulators are in widespread use by scientists and engineers across a range of government agencies, academia, and industry; in many cases, simulators provide the only means to examine processes that are infeasible or impossible to study otherwise. However, the cost of simulation studies can be quite high, both in terms of the time and computational resources required to conduct the trials and the manpower needed to sift through the resulting output. Thus, there is strong motivation to develop automated methods that enable more efficient knowledge extraction.

landscapes↗

NASA NTRS2003

Fast query-optimized kernel machine classification via incremental approximate nearest support vectors

Support vector machines (and other kernel machines) offer robust modern machine learning methods for nonlinear classfication.

vector machines kernel machines query nonlinear cl↗

NASA NTRS2003

An automated approach for acquiring onboard rover science

Rover traverse distances are increasing at a faster rate than downlink capacity is increasing. As this trend continues, the quantity of data that can be returned to Earth per meter traversed is reduced. We have developed an onboard science analysis technology for increasing science return from missions.

Dohm, J.↗

NASA NTRS2002

Anytime query-tuned kernel machine classifiers via Cholesky factorization

We recently demonstrated 2 to 64-fold query-time speedups of Support Vector Machine and Kernel Fisher classifiers via a new computational geometry method for anytime output bounds (DeCoste,2002). This new paper refines our approach in two key ways. First, we introduce a simple linear algebra formulation based on Cholesky factorization, yielding simpler equations and lower computational overhead. Second, this new formulation suggests new methods for achieving additional speedups, including tuning on query samples. We demonstrate effectiveness on benchmark datasets.

Kernel Cholesky MNIST↗

NASA NTRS2002

Anytime interval-valued outputs for kernel machines: fast support vector machine classification via distance geometry

We introduce a computational geometry method for which classification cost becomes roughly proportional to the difficulty of each example.

interval value outputs kernal vector distance geom↗

NASA NTRS2001

Support vector machines and kernel fisher discriminants: a case study using electronic nose data

Tradeoffs between the different kernel methods and performance relative to more traditional methods are discussed.

support vector machines classification↗

NASA NTRS2001

Compiling support vector machines for efficient classification

The time cost of classifying examples with support vector machines is traditionally linear in the number of support vectors, since each example computes dot products with each SV.

support vector machines machine learning classifie↗

NASA NTRS2001

Visualizing Mercer Kernel feature spaces via kernelized locally-linear embeddings

A new technique for projecting high-dimensional data to low-dimensional spaces, called locally linear embedding (LLE) has recently been introduced.

locally-linear embedding kernel methods↗

NASA NTRS2000

Alpha Seeding for Support Vector Machines

A key practical obstacle in applying support vector machines to many large-scale data mining tasks is that SVM's generally scale quadratically (or worse) in the number of examples or support vectors.

Support Vector Machines Alpha Seeding Data Mining↗

NASA NTRS2000

Distortion-Invariant Recognition via Jittered Queries

UNKNOWN

Distortion-Invariant↗

NASA NTRS2000

Learning Envelopes for Fault Detection and State Summarization

This paper discusses a data mining approach for overcoming common problems with the traditional red-line limit-checking approach to fault detection and state summarization.

Envelopes Fault Detection State Summarization↗

NASA NTRS1997

Bounds Estimation Via Regression with Asymmetric Cost Functions

This paper addresses a significant but mostly-neglected class of problems that we call bounds estimation. This includes learning empirical best-case and worst-case algorithmic complexity bounds and red-line bounds on sensor data.

regression-based learning algorithm bounds estimat↗