Learn More
A new method for automatic indexing and retrieval is described. The approach is to take advantage of implicit higher-order structure in the association of terms with documents (" semantic structure ") in order to improve the detection of relevant documents on the basis of terms found in queries. The particular technique used is singular-value decomposition,(More)
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(More)
In a new method for automatic indexing and retrieval, implicit higher-order structure in the association of terms with documents is modeled to improve estimates of term-document association, and therefore the detection of relevant documents on the basis of terms found in queries. Singular-value decomposition is used to decompose a large term by document(More)
SuperBook is a hypertext browsing system designed to improve the usability of conventional documents. Successive versions of SuperBook were evaluated in a series of behavioral studies. Students searched for information in a statistics text. presented either in conventional printed form or in SuperBook form. The best version of SuperBook enabled students to(More)
Singular value decomposition (SVD) can be viewed as a method for unsupervised training of a network that associates two classes of events reciprocally by linear connections through a single hidden layer. SVD was used to learn and represent relations among very large numbers of words (20k-60k) and very large numbers of natural text passages (lk-70k) in which(More)
For 11 studies, we find that the detection of usability problems as a function of number of users tested or heuristic evaluators employed is well modeled as a Poisson process. The model can be used to plan the amount of evaluation required to achieve desired levels of thoroughness or benefits. Results of early tests can provide estimates of the number of(More)
This paper describes cross-language information-retrieval experiments carried out for TREC-6. Our retrieval method, cross-language latent semantic indexing (CL-LSI), is completely automatic and we were able to use it to create a 3-way English-French-German IR system. This study extends our previous work in terms of the large size of training and testing(More)
The following seven articles describe new educational tools that rely on a unique capability of computer programs to abstract knowledge relationships from vast quantities of text, and using this, to determine the similarity of knowledge expressed in two or more texts. The tools can make comparisons among instructional sources and expository student writing,(More)
How much information from experi ence does a normal adult remember? The " functional informotion content " of humon memory wos estimated i n several ways. The methods depend on measured rotes of input and loss from very long-term memory and on onolyses of the informotionol demands of human memoty-bosed performance. Estimates ranged around 10' bits. It is(More)