Showing papers on "Learnability published in 1988"
24 Oct 1988
TL;DR: The authors consider the problem of predicting (0, 1)-valued functions on R/sup n/ and smaller domains, based on their values on randomly drawn points, and construct prediction strategies that are optimal to within a constant factor for any reasonable class F of target functions.
Abstract: The authors consider the problem of predicting (0, 1)-valued functions on R/sup n/ and smaller domains, based on their values on randomly drawn points. Their model is related to L.G. Valiant's learnability model (1984), but does not require the hypotheses used for prediction to be represented in any specified form. The authors first disregard computational complexity and show how to construct prediction strategies that are optimal to within a constant factor for any reasonable class F of target functions. These prediction strategies use the 1-inclusion graph structure from N. Alon et al.'s work on geometric range queries (1987) to minimize the probability of incorrect prediction. They then turn to computationally efficient algorithms. For indicator functions of axis-parallel rectangles and halfspaces in R/sup n/, they demonstrate how their techniques can be applied to construct computational efficient prediction strategies that are optimal to within a constant factor. They compare the general performance of prediction strategies derived by their method to those derived from existing methods in Valiant's learnability theory. >
197 citations
Proceedings Article•
1 Jan 1988
TL;DR: In this article, the authors consider the problem of predicting (0, l)valued functions on R" and smaller domains, based on their values on randomly drawn points, and construct prediction strategies that are optimal to within a constant factor for any reasonable class F of target functions.
Abstract: Summary. We consider the problem of predicting (0,l)valued functions on R" and smaller domains, based on their values on randomly drawn points. Our model is related to Valiant's learnability model, but does not require the hypotheses used for prediction to be represented in any specified form. First we disregard computational complexity and show how to construct prediction strategies that are optimal to within a constant factor for any reasonable class F of target functions. These prediction strategies use the 1-inclusion graph structure bom Non. Haussler and Welzl's work on geometric range queries to minimize the probability of incorrect prediction. We then turn to computationally efficient algorithms. For indicator functions of axis-parallel rectangles and halfspaces in R" , we demonstrate how our techniques can be applied to construct computationally efficient prediction stxategies that are optimal to within a constant factor. Finally. we compare the general performance of prediction strategies derived by OUT method to those derived from existing methods in Valiant's learnability theory.
176 citations
Proceedings Article•
1 Jan 1988
TL;DR: The problem of learning a concept from examples in a distribution-free model is considered.
83 citations
1 Dec 1988
TL;DR: In this article, the authors consider several variants of Valiant's learnability model that have appeared in the literature and give conditions under which these models are equivalent in terms of the polynomially learnable concept classes they define.
Abstract: Abstract In this paper we consider several variants of Valiant's learnability model that have appeared in the literature. We give conditions under which these models are equivalent in terms of the polynomially learnable concept classes they define. These equivalences allow comparisons of most of the existing theorems in Valiant-style learnability and show that several simplifying assumptions on polynomial learning algorithms can be made without loss of generality. We also give a useful reduction of learning problems to the problem of finding consistent hypotheses, and give comparisons and equivalences between Valiant's model and the prediction learning models of Haussler, Littlestone, and Warmuth ( in “29th Annual IEEE Symposium on Foundations of Computer Science,” 1988).
83 citations
1 Dec 1988
75 citations
Journal Article•
TL;DR: Several criteria for traffic sign symbols were examined through a questionnaire survey that allowed determination of the importance, or weighting, that should be assigned to each symbol in the design and evaluation of signs as mentioned in this paper.
Abstract: Several criteria for traffic sign symbols were examined through a questionnaire survey that allowed determination of the importance, or weighting, that should be assigned to each symbol in the design and evaluation of signs. The survey sample included traffic sign experts (members of national traffic control device committees) and practicing traffic engineers from Australia, New Zealand, Canada, and the United States. Separate ratings were assembled for symbols in general and for warning, regulatory, and information symbols in particular. Understandability was the factor rated most important, with conspicuity second. Learnability was considered least important, while reaction time, legibility distance, and glance legibility were rated equally but were determined to be more important than learnability.
43 citations
24 Oct 1988
TL;DR: The notion of dynamic sampling, wherein the number of examples examined can increase with the complexity of the target concept, is introduced and is used to establish the learnability of various concept classes with an infinite Vapnik-Chervonenkis (VC) dimension.
Abstract: The problem of learning a concept from examples in a distribution-free model is considered. The notion of dynamic sampling, wherein the number of examples examined can increase with the complexity of the target concept, is introduced. This method is used to establish the learnability of various concept classes with an infinite Vapnik-Chervonenkis (VC) dimension. An important variation on the problem of learning from examples, called approximating from examples, is also discussed. The problem of computing the VC dimension of a finite concept set defined on a finite domain is considered. >
32 citations
1 Dec 1988
22 citations
22 Aug 1988
TL;DR: A novel, nontrivial constraint on the degree of "locality" of grammars which allows a rich class of mildly context sensitive languages to be feasibly learnable is presented.
Abstract: We propose to apply a complexity theoretic notion of feasible learnability called "polynomial learnability" to the evaluation of grammatical formalisms for linguistic description. Polynomial learnability was originally defined by Valiant in the context of boolean concept learning and subsequently generalized by Blumer et al. to infinitary domains. We give a clear, intuitive exposition of this notion of learnability and what characteristics of a collection of languages may or many not help feasible learn ability under this paradigm. In particular, we present a novel, nontrivial constraint on the degree of "locality" of grammars which allows a rich class of mildly context sensitive languages to be feasibly learnable. We discuss possible implications of this observation to the theory of natural language acquisition.
16 citations
TL;DR: In this article, a model that allows children to control the complexity of the speech they hear within conversations on a moment-to-moment basis is presented. But, the model is not suitable for children to learn to control their own speech.
1 Jan 1988
TL;DR: In this paper, the authors focus on the status of core and peripheral constructions in the interlanguage (IL) data of adult native speakers of English learning Spanish and make the assumption that determining that status may help to delimit the domain of core grammar as distinct from the marked periphery.
Abstract: This paper focuses on questions related to issues of learnability in second language acquisition. Specifically, in this paper I attempt to clarify the relationship between a theory of grammar and the mechanisms that are responsible for the development of second language (L2) competence. Within that relationship, the specific though very general mechanism that I am concerned with is the status of core and peripheral constructions in the interlanguage (IL) data of adult native speakers of English learning Spanish. One basic assumption that I make is that determining that status may help to delimit the domain of core grammar as distinct from the marked periphery. Such a delimitation is one of the problems faced by researchers working on Universal Grammar (UG).
TL;DR: It is argued that appropriate metaphors in instructional materials could serve a mnemonic function, helping users to learn and remember the available operations on a computer system, and it is suggested that metaphorical models for device semantics may effect the learnability of command languages.
11 Jul 1988
12 Jun 1988
TL;DR: An in-depth analysis of the tractability of learning functions from determinations, a particular form of incomplete domain theory, and introduces the notion of “exceptions,” which is used to identify sufficient conditions for the learnability of function families consistent with partial and extended determinations.
Abstract: One well-known limitation of the explanation-based approach to concept learning is the need for a domain theory strong enough to deductively entail training examples of the concept. As such a theory may be unavailable in many situations, the problem of learning from incomplete domain theories must be addressed. The aim of this paper is to use the Valiant/Natarajan theoretical formalizations of concept learning to study the tractability of learning from incomplete domain theories. In particular, we present an in-depth analysis of the tractability of learning functions from determinations , a particular form of incomplete domain theory[3]. We show that only two of the five function families consistent with the five total determinations are polynomial-time learnable. We introduce the notion of “exceptions, and use it to identify sufficient conditions for the learnability of function families consistent with partial and extended determinations. While our results are specific to determinations, we believe that the underlying approach can be used to analyze other forms of incomplete theories.
TL;DR: In this paper, a game-theoretic approach to testing in the limit has been proposed, analogous to similar concepts in inductive inference and learnability theory, and feasibility and infeasibility results have been obtained.
5 Oct 1988
TL;DR: The author describes how learnability can be designed into software products focusing on the online documentation and hypermedia learning products that can help novice users become expert users faster and more easily.
Abstract: The author describes how learnability can be designed into software products focusing on the online documentation and hypermedia learning products that can help novice users become expert users faster and more easily. The author discusses how to choose the online documentation and hypermedia learning products which will best fit into the learning model for a product's particular users. To illustrate this she explains how such learning products are being mapped into the user learning model for Hewlett-Packard's Printed Circuit Design System. >
7 Jun 1988
TL;DR: It is shown that a novel, nontrivial constraint on the degree of "locality" of grammars allows not only context free languages but also a rich class of mildy context sensitive languages to be polynomially learnable.
Abstract: We apply a complexity theoretic notion of feasible learnability called "polynomial learnability" to the evaluation of grammatical formalisms for linguistic description. We show that a novel, nontrivial constraint on the degree of "locality" of grammars allows not only context free languages but also a rich class of mildy context sensitive languages to be polynomially learnable. We discuss possible implications of this result to the theory of natural language acquisition.
TL;DR: In this article, it was shown that human beings, whose contacts with the world are brief and personal and limited, are nevertheless able to know as much as they do know, and they are able to understand as much of the world as they know.
Abstract: … how comes it that human beings, whose contacts with the world are brief and personal and limited, are nevertheless able to know as much as they do know? – Russell in Human Knowledge: Its Scope and Limits
5 Oct 1988
TL;DR: In spring 1987 the Learning Products Group of Hewlett-Packard's Electronic Design Division began a project to design and write tutorials for the Hewlett -Packard Printed Circuit Design System, and tested the final tutorials for usability in November, 1987, before releasing them in January, 1988.
Abstract: In spring 1987 the Learning Products Group of Hewlett-Packard's Electronic Design Division began a project to design and write tutorials for the Hewlett-Packard Printed Circuit Design System. After a review of the literature on tutorial design, two approaches were selected to develop into short prototypes. These prototypes were tested to compare learnability and learning retention. After the results of the prototype testing were studied, one of the prototypes was selected to develop into a product. They tested the final tutorials for usability in November, 1987, before releasing them in January, 1988. >
16 Jan 1988
TL;DR: It is shown that for PUPS-style generalization analogical structure can be imposed on an arbitrary system (within a broad class he calls command systems.), so that there are learnable and unlearnable structures for this method.
Abstract: : Progress has been made in characterizing formally the capabilities and performance of inductive learning algorithms. Similar characterizations are needed for recently-proposed methods that produce generalizations from small numbers of analyzed examples. The author considers one class of such methods, based on the analogical generalization technique in Anderson and Thompson's PUPS system. It might appear that some to-be-learned structures can be learned by analogy, while others are too chaotic or inconsistent. It is shown that this intuition is correct for a simple form of analogical generalization, so that there are learnable and unlearnable structures for this method. In contrast, the author shows that for PUPS-style generalization analogical structure can be imposed on an arbitrary system (within a broad class he calls command systems.) It follows that the constraints on the PUPS-style method lie not in any structural condition on a to-be-learned system but rather in obtaining the knowledge needed to impose analogical structure.
TL;DR: Blissymbols were learned significantly faster than manual sign particularly in early learning trials, and the memory requirements of the two systems and their relationship with sign/symbols learning were discussed.
Book•
1 Jan 1988
TL;DR: Learnability issues in the acquisition of the dative alternation in English are addressed.
Abstract: Learnability issues in the acquisition of the dative alternation in English، للحصول على النص الكامل يرجى زيارة مكتبة الحسين بن طلال في جامعة اليرموك او زيارة موقعها الالكتروني
31 Dec 1988