Learning Automata from Ordered Examples

Connectionist learning models have had considerable empirical success, but it is hard to characterize exactly what they learn. The learning of finite-state languages (FSL) from example strings is a domain which has been extensively studied and might provide an opportunity to help understand connectionist learning. A major problem is that traditional FSL learning assumes the storage of all examples and thus violates connectionist principles. This paper presents a provably correct algorithm for inferring any minimum-state deterministic finite-state automata (FSA) from a complete ordered sample using limited total storage and without storing example strings. The algorithm is an iterative strategy that uses at each stage a current encoding of the data considered so far, and one single sample string. One of the crucial advantages of our algorithm is that the total amount of space used in the course of learning for encoding any finite prefix of the sample is polynomial in the size of the inferred minimum state deterministic FSA. The algorithm is also relatively efficient in time and has been implemented. More importantly, there is a connectionist version of the algorithm that preserves these properties. The connectionist version requires much more structure than the usual models and has been implemented using the Rochester Connectionist Simulator. We also show that no machine with finite working storage can iteratively identify the FSL from arbitrary presentations.     

Performance analysis of a class of transient detection algorithms-aunified framework

The detection of short-duration nonstationary signals, which are commonly referred to as transients, is often performed in the time-frequency-transform domain. An analytical framework within which the performance of different detectors based on linear transforms can be easily compared, for different classes of signals, is developed. A given class of transients is modeled as a signal existing in a linear subspace, plus a mismatch signal. Closed-form expressions are derived for the best and the worst detection performance for all possible transients in a given class. Using this framework, the performance of detectors based on the Gabor transform and on the short-time Fourier transform is evaluated and compared     

You said it before and you'll say it again: Expectations of consistency in communication.

Repeated reference creates strong expectations in addressees that a speaker will continue to use the same expression for the same object. The authors investigate the root reason for these expectations by comparing a cooperativeness-based account (Grice, 1975) with a simpler consistency-based account. In two eye-tracking experiments, the authors investigated the expectations underlying the effect of precedents on comprehension. The authors show that listeners expect speakers to be consistent in their use of expressions even when these expectations cannot be motivated by the assumption of cooperativeness. The authors conclude that though this phenomenon seems to be motivated by cooperativeness, listeners' expectation that speakers be consistent in their use of expressions is governed by a general expectation of consistency. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Modeling the Combining Site of the Human Asialoglycoprotein Receptor

A model for the carbohydrate recognition domain (CRD) and combining site of the human asialoglycoprotein (ASGP) receptor has been computed on the basis of the close sequence homology with the mannose-binding lectin (MBP), whose three-dimensional structure in complex with a ligand has been determined by crystallographic methods (Weis, W.I.; Drickamer, K.; Hendrickson, W.A. Nature 1992, 360: 127). Within the limitations of modeling methods, the model is compatible with data on ligand binding of the family C-type lectins, of which the MBP and the ASGP receptor are members. The model derived can serve as a guide for designing site-directed mutagenesis experiments in order to further elucidate the origins of specificity of the ASGP receptor toward galactose. In particular, the model focuses attention on the possible role of position 207 (MBP sequence numbering) in promoting galactose binding.     

Estimation of frequency in the presence of nonrandom interference

This paper develops a method of error analysis for Fourier transform-based sinusoidal frequency estimation in the presence of nonrandom interferences. A general error formula is derived and then specialized to the cases of additive and multiplicative interferences. Approximate error formulas are derived for the case of additive polynomial-phase interference. Finally, an application to error analysis in estimating the parameters of multiple polynomial-phase signals is discussed in detail     

Efficient pebbling for list traversal synopses with application to program rollback

We show how to support efficient back traversal in a unidirectional list, using small memory and with essentially no slowdown in forward steps. Using O(lgn)$O(lgn)$ memory for a list of size n$n$, the i$i$’th back-step from the farthest point reached so far takes O(lgi)$O(lgi)$ time in the worst case, while the overhead per forward step is at most ?$?$ for arbitrary small constant ?>0$?>0$. An arbitrary sequence of forward and back steps is allowed. A full trade-off between memory usage and time per back-step is presented: k$k$ vs. kn^1/k$k{n}^{1/k}$ and vice versa. Our algorithms are based on a novel pebbling technique which moves pebbles on a virtual binary, or n^1/k$n^{1/k}$-ary, tree that can only be traversed in a pre-order fashion.