Adaptive Parsing: Self-Extending Natural Language Interfaces by Jill Fain Lehman

By Jill Fain Lehman

As the pc steadily automates human-oriented initiatives in a number of environ­ ments, the interface among pcs and the ever-wider inhabitants of human clients assumes steadily expanding significance. within the place of work setting, for example, clerical initiatives equivalent to record submitting and retrieval, and higher-level projects akin to scheduling conferences, making plans journey itineraries, and generating records for book, are being partly or completely automatic. the variety of clients for workplace­ orientated software program contains clerks, secretaries, and businesspersons, none of whom are predominantly desktop literate. an analogous phenomenon is echoed within the manufacturing facility creation line, within the securities buying and selling flooring, in govt enterprises, in educa­ tional associations, or even in the house. The arcane command languages of definite­ teryear have confirmed too excessive a barrier for delicate reputation of automated func­ tions into the place of work, regardless of how helpful those features will be. computing device­ naive clients easily don't take some time to profit intimidating and complicated machine interfaces. to be able to position the performance of modem pcs on the disposition of numerous consumer populations, a few varied methods were attempted, many assembly with an important degree of good fortune, to wit: targeted classes to coach clients within the easier command languages (such as MS-DOS), designing point-and-click menu/graphics interfaces that require less person familiarization (illustrated so much in actual fact within the Apple Macintosh), and interacting with the consumer in his or her language of choice.

Show description

Read Online or Download Adaptive Parsing: Self-Extending Natural Language Interfaces PDF

Similar intelligence & semantics books

Natural language parsing systems

During the past there was no medical booklet on common language examine that provides a extensive and intricate description of the present difficulties of parsing within the context of man-made Intelligence. although, there are lots of fascinating effects from this area showing as a rule in several articles released in expert journals.

Partial Covers, Reducts and Decision Rules in Rough Sets: Theory and Applications

This monograph is dedicated to theoretical and experimental examine of partial reducts and partial determination ideas at the foundation of the examine of partial covers. using partial (approximate) reducts and selection principles rather than precise ones permits us to procure extra compact description of information contained in selection tables, and to layout extra particular classifiers.

Hidden Semi-Markov Models: Theory, Algorithms and Applications

Hidden semi-Markov types (HSMMs) are one of the most vital versions within the zone of synthetic intelligence / computer studying. because the first HSMM was once brought in 1980 for laptop popularity of speech, 3 different HSMMs were proposed, with quite a few definitions of length and commentary distributions.

Extra resources for Adaptive Parsing: Self-Extending Natural Language Interfaces

Example text

The low values were due largely to the fact that she never managed to find a parsable form for an entire class of subtasks (those involving the change action). Although User 8 did rely on the few forms that 12specifically, she was told to try to type simple, but fully grammatical sentences and to think of the system as someone to whom she was writing instructions. She was the only user who was given this aid. " indicates approximate value due to damage to log ftles "+" indicates instructions to work quickly given this session).

With the exception of extendable classes, our model proceeds from the opposite assumption: an unparsable sentence indicates a new form of reference, not a new referent. Zemik's assumption provides his system with a strong mechanism for limiting search, just as the Fixed Domain Assumption provides a different but equally powerful method for search reduction in our model. The empirical studies discussed in the next chapter demonstrate that a fixed domain, variable language assumption is the more realistic one for taskoriented natural language interfaces.

If all users employ the same restrictions on form and reference then a methodology such as that of Kelley [39,40,41] or Good et al. [24] is preferable; through a generate-and-test cycle one builds a single, monolithic system incorporating the complete common subset. 8 The fInal condition requires limited adaptability in the user. A non-adaptive approach to interface design assumes that the cognitive burden of learning the interface's sublanguage is a minor one. If the user is able to adapt quickly and with little effort to a fixed system, adaptation on the part of the system is no longer a clear advantage.

Download PDF sample

Rated 4.80 of 5 – based on 9 votes