The principal investigator (PI)-IN-A-BOX knowledge based system helps astronauts perform science experiments in space. These experiments are typically costly to devise and build and often are difficult to perform. Further, the space laboratory environment is unique; ever changing; hectic; and, therefore, stressful. The environment requires quick, correct reactions to events over a wide range of experiments and disciplines, including ones distant from an astronaut's main science specialty. This environment suggests the use of advanced techniques for data collection, analysis, and decision making to maximize the value of the research performed. PI-IN-A-BOX aids astronauts with quick-look data collection, reduction, and analysis as well as equipment diagnosis and troubleshooting, procedural reminders, and suggestions for high-value departures from the preplanned experiment protocol. The astronauts have direct access to the system, which is hosted on a portable computer in the Space Lab module. The system is in use on the ground for mission training and was used in flight during the October 1993 space life sciences 2 (SLS-2) shuttle mission.

The First covered genetic algorithms, neural elements called alpha helices and International Conference on Intelligent networks, and linguistic methods for beta sheets can be predicted with Systems for Molecular Biology sequence analysis. Invited speakers, some success, using algorithms that (ISMB-93), held 6-9 July 1993 at the who opened each day of the conference, examine a window of surrounding Lister Hill Center of the National were Temple Smith (Boston primary sequence.

The Association for the Advancement of Artificial Intelligence held its 1993 Fall Symposium Series on October 22-24 in Raleigh, North Carolina. This article contains summaries of the six symposia that were conducted: Automated Deduction in Nonstandard Logics; Games: Planning and Learning; Human-Computer Collaboration: Reconciling Theory, Synthesizing Practice; Instantiating Intelligent Agents; and Machine Learning and Computer Vision: What, Why, and How?

Science deals with knowledge of the material world based on objective reality. It is under constant attack by those who need magic, that is, concepts based on imagination and desire, with no basis in objective reality. A convenient target for such people is speculation on the machinery and method of operation of the human mind, questions that are still obscure in 1994. In The Emperor's New Mind, Roger Penrose attempts to look beyond objective reality for possible answers, using, in his argument, the theory that computers will never be able to duplicate the human experience. This article attempts to show where Penrose is in error by reviewing the evolution of men and computers and, based on this review, speculates about where computers might and might not imitate human perception. It then warns against the dangers of passive acceptance when respected scientists venture into the occult.

The theory revision problem is the problem of how best to go about revising a deficient domain theory using information contained in examples that expose inaccuracies. In this paper we present our approach to the theory revision problem for propositional domain theories. The approach described here, called PTR, uses probabilities associated with domain theory elements to numerically track the ``flow'' of proof through the theory. This allows us to measure the precise role of a clause or literal in allowing or preventing a (desired or undesired) derivation for a given example. This information is used to efficiently locate and repair flawed elements of the theory. PTR is proved to converge to a theory which correctly classifies all examples, and shown experimentally to be fast and accurate even for deep theories.

In this work we apply a texture classification network to remote sensing image analysis. The goal is to extract the characteristics of the area depicted in the input image, thus achieving a segmented map of the region. We have recently proposed a combined neural network and rule-based framework for texture recognition. The framework uses unsupervised and supervised learning, and provides probability estimates for the output classes. We describe the texture classification network and extend it to demonstrate its application to the Landsat and Aerial image analysis domain. 1 INTRODUCTION In this work we apply a texture classification network to remote sensing image analysis. The goal is to segment the input image into homogeneous textured regions and identify each region as one of a prelearned library of textures, e.g.

The real time computation of motion from real images using a single chip with integrated sensors is a hard problem. We present two analog VLSI schemes that use pulse domain neuromorphic circuits to compute motion. Pulses of variable width, rather than graded potentials, represent a natural medium for evaluating temporal relationships.

We investigate the use of information from all second order derivatives of the error function to perfonn network pruning (i.e., removing unimportant weights from a trained network) in order to improve generalization, simplify networks, reduce hardware or storage requirements, increase the speed of further training, and in some cases enable rule extraction. Our method, Optimal Brain Surgeon (OBS), is Significantly better than magnitude-based methods and Optimal Brain Damage [Le Cun, Denker and Sol1a, 1990], which often remove the wrong weights. OBS permits the pruning of more weights than other methods (for the same error on the training set), and thus yields better generalization on test data. Crucial to OBS is a recursion relation for calculating the inverse Hessian matrix HI from training data and structural information of the net. OBS permits a 90%, a 76%, and a 62% reduction in weights over backpropagation with weighL decay on three benchmark MONK's problems [Thrun et aI., 1991]. Of OBS, Optimal Brain Damage, and magnitude-based methods, only OBS deletes the correct weights from a trained XOR network in every case. Finally, whereas Sejnowski and Rosenberg [1987J used 18,000 weights in their NETtalk network, we used OBS to prune a network to just 1560 weights, yielding better generalization.

Inst., 19600 NW vonNeumann Dr, Beaverton, OR 97006 Abstract We propose a very simple, and well principled way of computing the optimal step size in gradient descent algorithms. The online version is very efficient computationally, and is applicable to large backpropagation networks trained on large data sets. The main ingredient is a technique for estimating the principal eigenvalue(s) and eigenvector(s) of the objective function's second derivative matrix (Hessian),which does not require to even calculate the Hessian. Severalother applications of this technique are proposed for speeding up learning, or for eliminating useless parameters. 1 INTRODUCTION Choosing the appropriate learning rate, or step size, in a gradient descent procedure such as backpropagation, is simultaneously one of the most crucial and expertintensive partof neural-network learning. We propose a method for computing the best step size which is both well-principled, simple, very cheap computationally, and, most of all, applicable to online training with large networks and data sets.

Typical methods for gradient descent in neural network learning involve calculation of derivatives based on a detailed knowledge of the network model. This requires extensive, time consuming calculations for each pattern presentation and high precision that makes it difficult to implement in VLSI. We present here a perturbation technique that measures, not calculates, the gradient. Since the technique uses the actual network as a measuring device, errors in modeling neuron activation and synaptic weights do not cause errors in gradient descent. The method is parallel in nature and easy to implement in VLSI. We describe the theory of such an algorithm, an analysis of its domain of applicability, some simulations using it and an outline of a hardware implementation.