(CMAC) [1, 2] is a neural network that models the structure and function of the part of the brain known as the cerebellum. The cerebellum provides precise coordination of motor control for such body parts as the eyes, arms, fingers, legs, and wings. It stores and retrieves information required to control thousands of muscles in producing coordinated behavior as a function of time. CMAC was designed to provide this kind of motor control for robotic manipulators. CMAC is a kind of memory, or table look-up mechanism, that is capable of learning motor behavior. It exhibits properties such as generalization, learning interference, discrimination, and forgetting that are characteristic of motor learning in biological creatures. In a biological motor system, the drive signal for each

Abstract—The problem of finding the closest point in high-dimensional spaces is common in pattern recognition. Unfortunately, the complexity of most existing search algorithms, such as k-d tree and R-tree, grows exponentially with dimension, making them impractical for dimensionality above 15. In nearly all applications, the closest point is of interest only if it lies within a user-specified distance e. We present a simple and practical algorithm to efficiently search for the nearest neighbor within Euclidean distance e. The use of projection search combined with a novel data structure dramatically improves performance in high dimensions. A complexity analysis is presented which helps to automatically determine e in structured problems. A comprehensive set of benchmarks clearly shows the superiority of the proposed algorithm for a variety of structured and unstructured search problems. Object recognition is demonstrated as an example application. The simplicity of the algorithm makes it possible to construct an inexpensive hardware search engine which can be 100 times faster than its software equivalent. A C++ implementation of our algorithm is available upon request to search@cs.columbia.edu/CAVE/.

Abstract: Stack processing, and in particular stack processing for the least recently used replacement algorithms, may present com-putational problems when it is applied to a sequence of page references with many different pages. This paper describes a new tech-nique for LRU stack processing that permits efficient processing of these sequences. An analysis of the algorithm and a comparison of its running times with those of the conventional stack processing algorithms are presented. Finally we discuss a multipass implementa-tion, which was found necessary to process trace data from a large data base system.

The automatic sequence is the central concept at the intersection of formal language theory and number theory. It was introduced by Cobham, and has been extensively studied by Christol, Kamae, Mendes France and Rauzy, and other writers. Since the range of automatic sequences is nite, however, their descriptive power is severely limited.

Skip lists are a probabilistic data structure that seem likely to supplant balanced trees as the implementation method of choice for many applications. Skip list algorithms have the same asymptotic expected time bounds as balanced trees and are simpler, faster and use less space. The original paper on skip lists only presented algorithms for search, insertion and deletion. In this paper, we show that skip lists are as versatile as balanced trees. We describe and analyze algorithms to use search fingers, merge, split and concatenate skip lists, and implement linear list operations using skip lists. The skip list algorithms for these actions are faster and simpler than their balanced tree cousins. The merge algorithm for skip lists we describe has better asymptotic time complexity than any previously described merge algorithm for balanced trees.

A new algorithm for text recognition that corrects character substitution errors in words of text is presented. The search for a correct word effectively integrates three knowledge sources: channel characteristics, bottom-up context, and top-down context. Channel characteristics are used in the form of probabilities that observed letters are corruptions of other letters; bottom-up context is in the form of the probability of a letter when the previous letters of the word are known; and top-down context is in the form of a lexicon. A one-pass algorithm is obtained by merging a previously known dynamic programming algorithm to compute the maximum a posteriori probability string (known as the Viterbi algorithm) with searching a lexical trie. Analysis of the computational compexity of the algorithm and results of experimentation with a PASCAL implementation are presented.

Cauchy-Euler differential equations surfaced naturally in a number of sorting and searching problems, notably in quicksort and binary search trees and their variations. Asymptotics of coefficients of functions satisfying such equations has been studied for several special cases in the literature. We study in this paper the most general framework for Cauchy-Euler equations and propose an asymptotic theory that covers almost all applications where Cauchy-Euler equations appear. Our approach is very general and requires almost no background on differential equations. Indeed the whole theory can be stated in terms of recurrences instead of functions. Old and new applications of the theory are given. New phase changes of limit laws of new variations of quicksort are systematically derived. We apply our theory to about a dozen of diverse examples in quicksort, binary search trees, urn models, increasing trees, etc.

We consider the planar point location problem from the perspective of expected search time. We are given a planar polygonal subdivision S and for each polygon of the subdivision the probability that a query point lies within this polygon. The goal is to compute a search structure to determine which cell of the subdivision contains a given query point, so as to minimize the expected search time. This is a generalization of the classical problem of computing an optimal binary search tree for one-dimensional keys. In the one-dimensional case it has long been known that the entropy H of the distribution is the dominant term in the lower bound on the expected-case search time, and further there exist search trees achieving expected search times of at most H + 2. Prior to this work, there has been no known structure for planar point location with an expected search time better than 2H, and this result required strong assumptions on the nature of the query point distribution. Here we present a data structure whose expected search time is nearly equal to the entropy lower bound, namely H + o(H). The result holds for any polygonal subdivision in which the number of sides of each of the polygonal cells is bounded, and there are no assumptions on the query distribution within each cell. We extend these results to subdivisions with convex cells, assuming a uniform query distribution within each cell.

The ability to model time-varying natures is essential to many database applications such as data warehousing and mining. However, the temporal aspects provide many unique characteristics and challenges for query processing and optimization. Among the challenges is computing temporal aggregates, which is complicated by having to compute temporal grouping. In this paper, we introduce a variety of temporal aggregation algorithms that overcome major drawbacks of previous work. First, for small-scale aggregations, both the worst-case and average-case processing time have been improved significantly. Second, for large-scale aggregations, the proposed algorithms can deal with a database that is substantially larger than the size of available memory. Third, the parallel algorithm designed on a shared-nothing architecture achieves scalable performance by delivering nearly linear scale-up and speed-up, even at the presence of data skew. The contributions made in this paper are particularly important because the rate of increase in database size and response time requirements has out-paced advancements in processor and mass storage technology.

The purpose of this paper is to introduce frameworks based on data-flow equations which provide for estimating the worst-case execution time (WCET) of (real-time) programs. These frameworks allow several different WCET analysis techniques, which range from nave approaches to exact analysis, provided exact knowledge on the program behaviour is available. However, data-flow frameworks can also be used for symbolic analysis based on information derived automatically from the source code of the program. As a byproduct we show that slightly modified elimination methods can be employed for solving WCET data-flow equations, while iteration algorithms cannot be used for this purpose.