Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Learning”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 361 records · Page 20

Fast temporal neural learning using teacher forcing

A neural network is trained to output a time dependent target vector defined over a predetermined time interval in response to a time dependent input vector defined over the same time interval by applying corresponding elements of the error vector, or difference between the target vector and the actual neuron output vector, to the inputs of corresponding output neurons of the network as corrective feedback. This feedback decreases the error and quickens the learning process, so that a much smaller number of training cycles are required to complete the learning process. A conventional gradient descent algorithm is employed to update the neural network parameters at the end of the predetermined time interval. The foregoing process is repeated in repetitive cycles until the actual output vector corresponds to the target vector. In the preferred embodiment, as the overall error of the neural network output decreasing during successive training cycles, the portion of the error fed back to the output neurons is decreased accordingly, allowing the network to learn with greater freedom from teacher forcing as the network parameters converge to their optimum values. The invention may also be used to train a neural network with stationary training and target vectors.

Toomarian, Nikzad↗

Learning and diagnosing faults using neural networks

Neural networks have been employed for learning fault behavior from rocket engine simulator parameters and for diagnosing faults on the basis of the learned behavior. Two problems in applying neural networks to learning and diagnosing faults are (1) the complexity of the sensor data to fault mapping to be modeled by the neural network, which implies difficult and lengthy training procedures; and (2) the lack of sufficient training data to adequately represent the very large number of different types of faults which might occur. Methods are derived and tested in an architecture which addresses these two problems. First, the sensor data to fault mapping is decomposed into three simpler mappings which perform sensor data compression, hypothesis generation, and sensor fusion. Efficient training is performed for each mapping separately. Secondly, the neural network which performs sensor fusion is structured to detect new unknown faults for which training examples were not presented during training. These methods were tested on a task of fault diagnosis by employing rocket engine simulator data. Results indicate that the decomposed neural network architecture can be trained efficiently, can identify faults for which it has been trained, and can detect the occurrence of faults for which it has not been trained.

Whitehead, Bruce A.↗

Modular, Hierarchical Learning By Artificial Neural Networks

Modular and hierarchical approach to supervised learning by artificial neural networks leads to neural networks more structured than neural networks in which all neurons fully interconnected. These networks utilize general feedforward flow of information and sparse recurrent connections to achieve dynamical effects. The modular organization, sparsity of modular units and connections, and fact that learning is much more circumscribed are all attractive features for designing neural-network hardware. Learning streamlined by imitating some aspects of biological neural networks.

Baldi, Pierre F.↗

Using Fuzzy Logic for Performance Evaluation in Reinforcement Learning

Current reinforcement learning algorithms require long training periods which generally limit their applicability to small size problems. A new architecture is described which uses fuzzy rules to initialize its two neural networks: a neural network for performance evaluation and another for action selection. This architecture is applied to control of dynamic systems and it is demonstrated that it is possible to start with an approximate prior knowledge and learn to refine it through experiments using reinforcement learning.

Berenji, Hamid R.↗

Learning In networks

Intelligent systems require software incorporating probabilistic reasoning, and often times learning. Networks provide a framework and methodology for creating this kind of software. This paper introduces network models based on chain graphs with deterministic nodes. Chain graphs are defined as a hierarchical combination of Bayesian and Markov networks. To model learning, plates on chain graphs are introduced to model independent samples. The paper concludes by discussing various operations that can be performed on chain graphs with plates as a simplification process or to generate learning algorithms.

Buntine, Wray L.↗

Web-Based Instruction and Learning: Analysis and Needs Assessment

An analysis and needs assessment was conducted to identify kindergarten through grade 14 (K-14) customer needs with regard to using the World Wide Web (WWW) for instruction and to identify obstacles K-14 teachers face in utilizing NASA Learning Technologies products in the classroom. The needs assessment was conducted as part of the Dryden Learning Technologies Project which is a collaboration between Dryden Flight Research Center (DFRC), Edwards, California and Tne Pennsylvania State University (PSU), University Park, Pennsylvania. The overall project is a multiyear effort to conduct research in the development of teacher training and tools for Web-based science, mathematics and technology instruction and learning.

Grabowski, Barbara↗

Technology Development, Implementation, and Assessment: K-16 Pre-Service, In-Service, and Distance Learning Initiatives

This summer 22 kindergarten through 8th grade teachers attended a 3-week Teacher Enhancement Institute (TEI) at NASA Langley Research Center. TEI is funded by NASA Education Division and is a collaborative effort between NASA Langley's Office of Education and Christopher Newport University. Selected teacher teams were drawn from Langley's 5-state precollege service region, which includes Kentucky, North Carolina, South Carolina, Virginia, and West Virginia. The goal of TEI was for teachers to learn aeronautics and the broad application of science and technology through a problem-based learning (PBL) strategy. PBL is an instructional method using a real world problem, also known as an ill-structured problem, as the context for an in-depth investigation. Most real life problems are ill-structured, as are all the really important social, political and scientific problems. The teachers were immediately immersed in an ill-structured problem to design a communication strategy for the White House Commission on Aviation Safety and Security to educate and disseminate aviation information to the general public. Specifically, the communication strategy was to focus on aeronautics principles, technology and design associated with US general aviation revitalization and aviation safety programs. The presented problem addressed NASA's strategic outcome to widely communicate the content, relevancy and excitement of its missions and discoveries to the general population. Further, the PBL scenario addressed the technological challenges being taken up by NASA to revolutionize air travel and the way in which aircraft are designed, built, and operated. It also addressed getting people and freight safely and efficiently to any location in the world at a reasonable cost. With a "real" need-to-know problem facing them, the teachers set out to gather information and to better understand the problem using inquiry-based and scientific methods. The learning in this aeronautics scenario was driven by the direction taken by participants. With the support of the TEI faculty, the teachers quickly identified NASA Langley researchers that served as consultants to help solve the problem. To achieve their goal, the teacher teams developed lesson plans for elementary and middle school students, wrote a newspaper, published a brochure to educate the general public, constructed games for children of all ages, and produced a video. As a second problem, the TEI participants will design their own aeronautic lesson plan and immerse their 1997-98 school year students in the problem. The problem is for the students "to create a traveling hands-on, minds-on aeronautics museum exhibit created for children by children." As a culminating activity, the Virginia Air and Space Center in Hampton, VA, will set up a special display of the exhibits in the Summer 1998. The TEI faculty will visit each TEI teacher's classroom during the academic school year to observe the implementation of the unit. In addition to the classroom observations, electronic follow-up sessions will be conducted during the school year to support the teachers' efforts in developing their PBL units to integrate technology in math and science instruction. These sessions eill be conducted using the Internet. Teachers will be connected through a chat-line to share ideas, ask questions, and generate solutions.

Petersen, Richard↗

Linear System Control Using Stochastic Learning Automata

This paper explains the use of a Stochastic Learning Automata (SLA) to control switching between three systems to produce the desired output response. The SLA learns the optimal choice of the damping ratio for each system to achieve a desired result. We show that the SLA can learn these states for the control of an unknown system with the proper choice of the error criteria. The results of using a single automaton are compared to using multiple automata.

Ziyad, Nigel↗

Learning to Identify Airports from Photographs Versus Navigational Maps Presented at Various Orientations

We previously demonstrated: 1) initial learning of navigational maps produced discriminative reaction times (RTs) that were most rapid when the maps were seen in the same orientation as that in which they were initially learned, 2) RTs were significantly reduced with repeated representations, and 3) information learned from the map was not sufficient for all observers to recognize aerial photographs of the same airports. The present study examined the relative effectiveness of initial training on photographs versus maps.

Cohen, Malcolm M.↗

Development of X-33/X-34 Aerothermodynamic Data Bases: Lessons Learned and Future Enhancements

A synoptic of programmatic and technical lessons learned in the development of aerothermodynamic data bases for the X-33 and X-34 programs is presented in general terms and from the perspective of the NASA Langley Research Center Aerothermodynamics Branch. The format used is that of the "aerothermodynamic chain," the links of which are personnel, facilities, models/test articles, instrumentation, test techniques, and computational fluid dynamics (CFD). Because the aerodynamic data bases upon which the X-33 and X-34 vehicles will fly are almost exclusively from wind tunnel testing, as opposed to CFD, the primary focus of the lessons learned is on ground-based testing. The period corresponding to the development of X-33 and X-34 aerothermodynamic data bases was challenging, since a number of other such programs (e.g., X-38, X-43) competed for resources at a time of downsizing of personnel, facilities, etc., outsourcing, and role changes as NASA Centers served as subcontractors to industry. The impact of this changing environment is embedded in the lessons learned. From a technical perspective, the relatively long times to design and fabricate metallic force and moment models, delays in delivery of models, and a lack of quality assurance to determine the fidelity of model outer mold lines (OML) prior to wind tunnel testing had a major negative impact on the programs. On the positive side, the application of phosphor thermography to obtain global, quantitative heating distributions on rapidly fabricated ceramic models revolutionized the aerothermodynamic optimization of vehicle OMLs, control surfaces, etc. Vehicle designers were provided with aeroheating information prior to, or in conjunction with, aerodynamic information early in the program, thereby allowing trades to be made with both sets of input; in the past only aerodynamic data were available as input. Programmatically, failure to include transonic aerodynamic wind tunnel tests early in the assessment phase led to delays in the optimization phase, as OMLs required modification to provide adequate transonic aerodynamic performance without sacrificing subsonic and hypersonic performance. Funding schedules for industry, based on technical milestones, also presented challenges to aerothermodynamics seeking optimum flying characteristics across the subsonic to hypersonic speed regimes and minimum aeroheating. This paper is concluded with a brief discussion of enhancements in ground-based testing/CFD capabilities necessary to partially/fully satisfy future requirements.

Miller, C. G.↗

Optimal Reward Functions in Distributed Reinforcement Learning

We consider the design of multi-agent systems so as to optimize an overall world utility function when (1) those systems lack centralized communication and control, and (2) each agents runs a distinct Reinforcement Learning (RL) algorithm. A crucial issue in such design problems is to initialize/update each agent's private utility function, so as to induce best possible world utility. Traditional 'team game' solutions to this problem sidestep this issue and simply assign to each agent the world utility as its private utility function. In previous work we used the 'Collective Intelligence' framework to derive a better choice of private utility functions, one that results in world utility performance up to orders of magnitude superior to that ensuing from use of the team game utility. In this paper we extend these results. We derive the general class of private utility functions that both are easy for the individual agents to learn and that, if learned well, result in high world utility. We demonstrate experimentally that using these new utility functions can result in significantly improved performance over that of our previously proposed utility, over and above that previous utility's superiority to the conventional team game utility.

Wolpert, David H.↗

Any Two Learning Algorithms Are (Almost) Exactly Identical

This paper shows that if one is provided with a loss function, it can be used in a natural way to specify a distance measure quantifying the similarity of any two supervised learning algorithms, even non-parametric algorithms. Intuitively, this measure gives the fraction of targets and training sets for which the expected performance of the two algorithms differs significantly. Bounds on the value of this distance are calculated for the case of binary outputs and 0-1 loss, indicating that any two learning algorithms are almost exactly identical for such scenarios. As an example, for any two algorithms A and B, even for small input spaces and training sets, for less than 2e(-50) of all targets will the difference between A's and B's generalization performance of exceed 1%. In particular, this is true if B is bagging applied to A, or boosting applied to A. These bounds can be viewed alternatively as telling us, for example, that the simple English phrase 'I expect that algorithm A will generalize from the training set with an accuracy of at least 75% on the rest of the target' conveys 20,000 bytes of information concerning the target. The paper ends by discussing some of the subtleties of extending the distance measure to give a full (non-parametric) differential geometry of the manifold of learning algorithms.

Wolpert, David H.↗

Demonstrating and Evaluating an Action Learning Approach to Building Project Management Competence

This paper contributes a description of an action-learning approach to building project management competence. This approach was designed, implemented, and evaluated for use with the Dynacs Engineering Development Contract at the Kennedy Space Center. The aim of the approach was to improve three levels of competence within the organization: individual project management skills, project team performance. and organizational capabilities such as the project management process and tools. The overall steps to the approach, evaluation results, and lessons learned are presented. Managers can use this paper to design a specific action-learning approach for their organization.

Kotnour, Tim↗

A Guide to the Literature on Learning Graphical Models

This literature review discusses different methods under the general rubric of learning Bayesian networks from data, and more generally, learning probabilistic graphical models. Because many problems in artificial intelligence, statistics and neural networks can be represented as a probabilistic graphical model, this area provides a unifying perspective on learning. This paper organizes the research in this area along methodological lines of increasing complexity.

Buntine, Wray L.↗

NASA's Lessons Learned and Technical Standards: A Logical Marriage

Lessons Learned have been the basis for our accomplishments throughout the ages. They have been passed down from father to son, mother to daughter, teacher to pupil, and older to younger worker. Lessons Learned have also been the basis for NASA's accomplishments for more than forty years. Both government and industry have long recognized the need to systematically document and utilize the knowledge gained from past experiences in order to avoid the repetition of failures and mishaps. Lessons Learned have formed the foundation for discoveries, inventions, improvements, textbooks, and Technical Standards.

Gill, Paul↗

A Dual-Route Model that Learns to Pronounce English Words

This paper describes a model that learns to pronounce English words. Learning occurs in two modules: 1) a rule-based module that constructs pronunciations by phonetic analysis of the letter string, and 2) a whole-word module that learns to associate subsets of letters to the pronunciation, without phonetic analysis. In a simulation on a corpus of over 300 words the model produced pronunciation latencies consistent with the effects of word frequency and orthographic regularity observed in human data. Implications of the model for theories of visual word processing and reading instruction are discussed.

Remington, Roger W.↗

Automated Decomposition of Model-based Learning Problems

A new generation of sensor rich, massively distributed autonomous systems is being developed that has the potential for unprecedented performance, such as smart buildings, reconfigurable factories, adaptive traffic systems and remote earth ecosystem monitoring. To achieve high performance these massive systems will need to accurately model themselves and their environment from sensor information. Accomplishing this on a grand scale requires automating the art of large-scale modeling. This paper presents a formalization of [\em decompositional model-based learning (DML)], a method developed by observing a modeler's expertise at decomposing large scale model estimation tasks. The method exploits a striking analogy between learning and consistency-based diagnosis. Moriarty, an implementation of DML, has been applied to thermal modeling of a smart building, demonstrating a significant improvement in learning rate.

Williams, Brian C.↗

Orientation During Initial Learning and Subsequent Discrimination of Faces

Discrimination of facial features degrades with stimulus rotation (e.g., the "Margaret Thatcher" effect). Thirty-two observers learned to discriminate between two upright, or two inverted, faces. Images, erect and rotated by +/-45deg, +/-90deg, +/-135deg and 180deg about the line of sight, were presented on a computer screen. Initial discriminative reaction times increased with stimulus rotation only for observers who learned the upright faces. Orientation during learning is critical in identifying faces subsequently seen at different orientations.

Cohen, Malcolm M.↗