The Autonomous Exploration for Gathering Increased Science (AEGIS) system enables automated data collection by planetary rovers. AEGIS software was uploaded to the Mars Exploration Rover (MER) mission’s Opportunity rover in December 2009 and has successfully demonstrated automated onboard targeting based on scientist-specified objectives. Prior to AEGIS, images were transmitted from the rover to the operations team on Earth; scientists manually analyzed the images, selected geological targets for the rover’s remote-sensing instruments, and then generated a command sequence to execute the new measurements. AEGIS represents a significant paradigm shift---by using onboard data analysis techniques, the AEGIS software uses scientist input to select high-quality science targets with no human in the loop. This approach allows the rover to autonomously select and sequence targeted observations in an opportunistic fashion, which is particularly applicable for narrow field-of-view instruments (such as the MER Mini-TES spectrometer, the MER Panoramic camera, and the 2011 Mars Science Laboratory (MSL) ChemCam spectrometer). This article provides an overview of the AEGIS automated targeting capability and describes how it is currently being used onboard the MER mission Opportunity rover.

Physics-based simulation codes are widely used in science and engineering to model complex systems that would be infeasible to study otherwise. Such codes provide the highest-fidelity representation of system behavior, but are often so slow to run that insight into the system is limited. For example, conducting an exhaustive sweep over a d-dimensional input parameter space with k-steps along each dimension requires kd simulation trials (translating into kd CPU-days for one of our current simulations). An alternative is directed exploration in which the next simulation trials are cleverly chosen at each step. Given the results of previous trials, supervised learning techniques (SVM, KDE, GP) are applied to build up simplified predictive models of system behavior. These models are then used within an active learning framework to identify the most valuable trials to run next. Several active learning strategies are examined including a recently-proposed information-theoretic approach. Performance is evaluated on a set of thirteen synthetic oracles, which serve as surrogates for the more expensive simulations and enable the experiments to be replicated by other researchers.

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. Unlike traditional data mining, knowledge discovery from simulators is not limited to a static, pre-determined dataset; instead, the simulator itself can be used as an oracle to generate new data of our own choosing. We exploit this opportunity by employing active learning and support vector machines (SVMs) to choose which are the most valuable simulation trials to run next. On two realworld scientific simulators, one for asteroid collisions and one for magnetospheric modeling, we demonstrate twofold and sixfold reductions, respectively, in the number of simulator trials required to achieve a particular level of fidelity in landscape characterization as compared with a standard grid-based sampling approach.

In this paper, an algorithm for extracting information from raw, surveillance-style video of an outdoor scene containing a mix of people, bicycles, and motorized vehicles is presented. A feature extraction algorithm based on background estimation and subtraction followed by spatial clustering and multi-object tracking is used to process sequences of video frames into a track set. The resulting track set, which encodes the positions, velocities, and appearances of the various objects as a function of time, are mined to answer user-generated queries that are potentially relevant for surveillance applications and for input to public planning processes.

Arrays of polymer films embedded with conductive or resistive material have attracted significant attention as “electronic noses”. Sorption of a vapor into the polymer films causes physical swelling, which leads to a change in the DC electrical resistance of the film. Like the receptors of the mammalian olfactory system, each polymer-based detector responds to more than one analyte and each analyte elicits a response from more than one detector. The DC resistance across the array of detectors is sampled versus time producing a multivariate time-series. Except when rapid response is critical, this multivariate time-series can be converted into a vector representation by focusing on the steady-state behavior and calculating the relative change in resistance in each channel relative to an initial baseline. In this paper, we provide an overview of our efforts to mine useful information from electronic nose data. Four case studies are presented: rapid detection of DNT (dinitrotoluene) vapors at very low concentrations (< 1 part-per-billion), pairwise discrimination between chemically similar analytes (e.g., H2O and D2O; heptane and hexane), prediction of human percepts of odor quality from electronic nose detector responses, and classification of vapors into the appropriate chemical family (e.g., saying methanol is an alcohol without having previously smelled methanol).

This paper presents a new approach for achieving distortion-invariant recognition and classification. A test example to be classified is viewed as a query intended to find similar examples in the training set (or to find similar class models that represent a compression of the training set). The key idea is that instead of querying with a single pattern, we construct a more robust query, based on the family of patterns formed by distorting the test example. Although query execution is slower than if the invariances were successfully pre-compiled during training, there are significant advantages in several contexts: (i) providing invariances in memory-based learning, (ii) in model selection, where reducing training time at the expense of test time is a desirable trade-off, and (iii) in enabling robust, ad hoc searches based on a single example. Preliminary tests for memory-based learning on the NIST handwritten digit database with a limited set of shearing and translation distortions produced an error rate of 1.35%.

Diamond Eye is a new image mining system that enables users (scientists) to locate and catalog objects of interest in large image collections. This system provides a platform-independent interface to novel image mining algorithms, as well as to computational and database resources that allow scientists to browse, annotate and search through images and analyze the resulting object catalogs.

Different instances of a handwritten word consist of the same basic features (humps, cusps, crossings, etc.) arranged in a deformable spatial pattern. Thus, keywords in cursive text can be detected by looking for the appropriate features in the "correct" spatial configuration. A keyword can be modeled hierarchically as a set of word fragments, each of which consists of lower-level features. To allow flexibility, the spatial configuration of keypoints within a fragment is modeled using a Dryden-Mardia (DM) probability density over the shape of the configuration. In a writer-dependent test on a transcription of the Declaration of Independence (/spl sim/1300 words, /spl sim/7500 characters), the method detected all eleven instances of the keyword "government" with only four false positives.

Under a weak perspective camera model, the image plane coordinates in different views of a planar object are related by an affine transformation. Because of this property, researchers have attempted to use affine invariants for recognition. However, there are two problems with this approach: (1) objects or object classes with inherent variability cannot be adequately treated using invariants; and (2) in practice the calculated affine invariants can be quite sensitive to errors in the image plane measurements. In this paper we use probability distributions to address both of these difficulties. Under the assumption that the feature positions of a planar object can be modeled using a jointly Gaussian density, we have derived the joint density over the corresponding set of affine coordinates. Even when the assumptions of a planar object and a weak perspective camera model do not strictly hold, the results are useful because deviations from the ideal can be treated as deformability in the underlying object model.

We present a new framework for recognizing planar object classes, which is based on local feature detectors and a probabilistic model of the spatial arrangement of the features. The allowed object deformations are represented through shape statistics, which are learned from examples. Instances of an object in an image are detected by finding the appropriate features in the correct spatial configuration. The algorithm is robust with respect to partial occlusion, detector false alarms, and missed features. A 94% success rate was achieved for the problem of locating quasi-frontal views of faces in cluttered scenes.

An algorithm for locating quasi-frontal views of human faces in cluttered scenes is presented. The algorithm works by coupling a set of local feature detectors with a statistical model of the mutual distances between facial features it is invariant with respect to translation, rotation (in the plane), and scale and can handle partial occlusions of the face. On a challenging database with complicated and varied backgrounds, the algorithm achieved a correct localization rate of 95% in images where the face appeared quasi-frontally.>

This paper considers the decision-making problem of selecting a strategy from a set of alternatives on the basis of incomplete information (e.g. a finite number of observations). At any time the system can adopt a particular strategy or decide to gather additional information at some cost. Balancing the expected utility of the new information against the cost of acquiring the information is the central problem that the authors address. In the authors' approach, the cost and utility of applying a particular strategy to a given problem are represented as random variables from a parametric distribution. By observing the performance of each strategy on a randomly selected sample of problems, one can use parameter estimation techniques to infer statistical models of performance on the general population of problems. These models can then be used to estimate: (1) the utility and cost of acquiring additional information and (2) the desirability of selecting a particular strategy from a set of choices. Empirical results are presented that demonstrate the effectiveness of the hypothesis evaluation techniques for tuning system parameters in a NASA antenna scheduling application.>

Our long-term goal is to develop a trainable tool for locating patterns of interest in large image databases. Toward this goal we have developed a prototype system, based on classical filtering and statistical pattern recognition techniques, for automatically locating volcanoes in the Magellan SAR database of Venus. Training for the specific volcano-detection task is obtained by synthesizing feature templates (via normalization and principal components analysis) from a small number of examples provided by experts. Candidate regions identified by a focus of attention (FOA) algorithm are classified based on correlations with the feature templates. Preliminary tests show performance comparable to trained human observers.>

Lack of verifiable ground truth is a common problem in remote sensing image analysis. For example, consider the synthetic aperture radar (SAR) image data of Venus obtained by the Magellan spacecraft. Planetary scientists are interested in automatically cataloging the locations of all the small volcanoes in this data set; however, the problem is very difficult and cannot be performed with perfect reliability even by human experts. Thus, training and evaluating the performance of an automatic algorithm on this data set must be handled carefully. We discuss the use of weighted free-response receiver-operating characteristics (wFROCs) for evaluating detection performance when the "ground truth" is subjective. In particular, we evaluate the relative detection performance of humans and automatic algorithms. Our experimental results indicate that proper assessment of the uncertainty in "ground truth" is essential in applications of this nature.>

In remote sensing applications "ground-truth" data is often used as the basis for training pattern recognition algorithms to gener(cid:173) ate thematic maps or to detect objects of interest. In practical situations, experts may visually examine the images and provide a subjective noisy estimate of the truth. Calibrating the reliability and bias of expert labellers is a non-trivial problem. In this paper we discuss some of our recent work on this topic in the context of detecting small volcanoes in Magellan SAR images of Venus. Empirical results (using the Expectation-Maximization procedure) suggest that accounting for subjective noise can be quite signifi(cid:173) cant in terms of quantifying both human and algorithm detection performance.