About half the pyramidal neurons in layer 5 of neocortex have long apical dendrites that arborize extensively in layers 1-3. There the dendrites receive synaptic input from the inter-areal feedback projections (Felleman and van Essen, 1991) that play an important role in many models of brain function (Rockland and Virga, 1989). At first sight this seems to be an unsatisfactory arrangement. In light of traditional passive models of dendritic function the distant inputs cannot have a significant effect on the output discharge of the pyramidal cell. The distal inputs are at least one to two space constants removed from the soma in layer 5 and so only a small fraction of the voltage signal will reach there.

A gradient descent algorithm for parameter estimation which is similar to those used for continuous-time recurrent neural networks was derived for Hodgkin-Huxley type neuron models. Using membrane potentialtrajectories as targets, the parameters (maximal conductances, thresholds and slopes of activation curves, time constants) weresuccessfully estimated. The algorithm was applied to modeling slow non-spike oscillation of an identified neuron in the lobster stomatogastric ganglion. A model with three ionic currents was trained with experimental data. It revealed a novel role of A-current for slow oscillation below -50 mY. 1 INTRODUCTION Conductance-based neuron models, first formulated by Hodgkin and Huxley [10], are commonly used for describing biophysical mechanisms underlying neuronal behavior.

In this paper we shall be concerned with Mealy machines, although our approach can easily be extended to other formulations to yield equivalent results.

Spotting tasks require detection of target patterns from a background of richly varied non-target inputs. The performance measure of interest for these tasks, called the figure of merit (FOM), is the detection rate for target patterns when the false alarm rate is in an acceptable range. A new approach to training spotters is presented which computes the FOM gradient for each input pattern and then directly maximizes the FOM using backpropagation. This eliminates the need for thresholds during training. It also uses network resources to model Bayesian a posteriori probability functions accurately only for patterns which have a significant effect on the detection accuracy over the false alarm rate of interest.

Terence D. Sanger Jet Propulsion Laboratory MS 303-310 4800 Oak Grove Drive Pasadena, CA 91109 Abstract The Singular Value Decomposition (SVD) is an important tool for linear algebra and can be used to invert or approximate matrices. Although many authors use "SVD" synonymously with "Eigenvector Decomposition"or "Principal Components Transform", it is important to realize that these other methods apply only to symmetric matrices, while the SVD can be applied to arbitrary nonsquare matrices. This property is important for applications to signal transmission and control. I propose two new algorithms for iterative computation of the SVD given only sample inputs and outputs from a matrix. Although there currently exist many algorithms for Eigenvector Decomposition (Sanger1989, for example), these are the first true samplebased SVDalgorithms. 1 INTRODUCTION The Singular Value Decomposition (SVD) is a method for writing an arbitrary nons quare matrix as the product of two orthogonal matrices and a diagonal matrix.

Reciprocal circuit elements facilitate such an implementation since they 882 The "Softmax" Nonlinearity 883 can be combined with other reciprocal elements to form an analog network having Lyapunov-like functions: the network content or co-content. In this paper, we show a reciprocal implementation of the "softmax" nonlinearity that is usually used to enforce local competition between neurons [Peterson, 1989]. We show that the circuit ispassive and incrementally passive, and we explicitly compute its content and co-content functions. This circuit adds a new element to the library of the analog circuit designer that can be combined with reciprocal constraint boxes [Harris, 1988] and nonlinear resistive fuses [Harris, 1989] to form fast, analog VLSI optimization networks.

The recurrent network,containing six continuous-time analog neurons and 42 free parameters (connection strengths and thresholds), is trained to generate time-varying outputs approximating given periodic signals presented to the network. The chip implements a stochastic perturbative algorithm,which observes the error gradient along random directions in the parameter space for error-descent learning. In addition tothe integrated learning functions and the generation of pseudo-random perturbations, the chip provides for teacher forcing andlong-term storage of the volatile parameters. The network learns a 1 kHz circular trajectory in 100 sec. The chip occupies 2mm x 2mm in a 2JLm CMOS process, and dissipates 1.2 mW. 1 Introduction Exact gradient-descent algorithms for supervised learning in dynamic recurrent networks [1-3]are fairly complex and do not provide for a scalable implementation in a standard 2-D VLSI process. We have implemented a fairly simple and scalable ·Present address: Johns Hopkins University, ECE Dept., Baltimore MD 21218-2686.

Alessandro Sperduti* Department of Computer Science Pisa University Corso Italia 40, 56125 Pisa, Italy Abstract In this paper we propose an extension to the RAAM by Pollack. This extension, the Labeling RAAM (LRAAM), can encode labeled graphswith cycles by representing pointers explicitly. Data encoded in an LRAAM can be accessed by pointer as well as by content. Direct access by content can be achieved by transforming theencoder network of the LRAAM into an analog Hopfield network with hidden units. Different access procedures can be defined depending on the access key. Sufficient conditions on the asymptotical stability of the associated Hopfield network are briefly introduced. 1 INTRODUCTION In the last few years, several researchers have tried to demonstrate how symbolic structures such as lists, trees, and stacks can be represented and manipulated in a connectionist system, while still preserving all the computational characteristics of connectionism (and extending them to the symbolic representations) (Hinton, 1990; Plate, 1991; Pollack, 1990; Smolensky, 1990; Touretzky, 1990).

Motivated by mathematical modeling, analog implementation and distributed simulation of neural networks, we present a definition of asynchronous dynamics of general CT dynamical systems defined by ordinary differential equations, based on notions of local times and communication times. We provide some preliminary results on globally asymptotical convergence of asynchronous dynamics for contractive and monotone CT dynamical systems. When applying theresults to neural networks, we obtain some conditions that ensure additive-type neural networks to be asynchronizable.

Many well known neural network techniques for adaptive pattern classification and function approximation are inherently highly parallel, and thus have proven difficult toimplement for real-time applications at a reasonable cost.