The vocabulary problem in human-system communication

Abstract

In almost all computer applications, users must enter correct words for the desired objects or actions. For success without extensive training, or in first-tries for new targets, the system must recognize terms that will be chosen spontaneously. We studied spontaneous word choice for objects in five application-related domains, and found the variability to be surprisingly large. In every case two people favored the same term with probability <0.20. Simulations show how this fundamental property of language limits the success of various design methodologies for vocabulary-driven interaction. For example, the popular approach in which access is via one designer's favorite single word will result in 80-90 percent failure rates in many common situations. An optimal strategy, unlimited aliasing, is derived and shown to be capable of several-fold improvements.

Effective Use of Ontologies

“…Pot ser una etiqueta que evoqui el contingut de quelcom (per exemple, productes ecològics), una etiqueta que capti una impressió o acció que provoca quelcom a un usuari (per exemple: important o per llegir), o una combinació dels dos (per exemple, productes ecològics importants). La majoria de les etiquetes que trobem a Internet entren en la primera categoria, descrivint l'origen del element etiquetat (Furnas et al 1987). Per tant, proporcionen una manera ràpida d'identificar aspectes clau d'un domini, o dels elements que conté.…”

Towards a taxonomy of error-handling strategies in recognition-based multi-modal human–computer interfaces

“…Obviously, the more constraints are imposed on what users can say to an application in a restricted grammar, the less likely they are to encounter recognition errors, but this should not be obtained at the price of forcing users to memorise commands. A badly designed grammar or a grammar that imposes too many constraints on user speech can be the trigger for more errors, as user speech, especially beginners' speech, will often not conform to the predefined grammar (this is dubbed "the vocabulary problem" in [24]). The Wizard of Oz technique presents a useful methodology to elicit user vocabulary and natural expressions for building appropriate grammars and testing dialogue models [25].…”

“…Strategies for error correction include: (24) repeat, (25) spell out, (26) rephrase, (27) contradict, (28) modality switch, and (29) cross-modal correction. Finally, as recognition systems become more sophisticated, they may be able to interpret user natural (30) correction marks.…”

Metadata Extraction Using Case-basedReasoning for Heterogeneous Thai Documents

“…There are no changes in word form or word inflection as an expression of tense, case or gender; word ordering plays an important part in determining the syntactic role of word. The same form of words in different positions contains different syntactic properties and therefore conveys different meanings [8]. To express tense and case, additional words often are inserted to clarify the meaning.…”