User talk:Ychen129

Ye Chen Literature Management

Cognitive presence
coming soon

Latent semantic analysis
Thomas K Landauer (1999) Latent semantic analysis: A theory of the psychology of language and mind, Discourse Processes, 27:3, 303-310, DOI: 10.1080/01638539909545065

This article is a response to Charles Perfetti's doubt and critics toward LSA. "Charles Perfetti (1998), in a commentary, agreed that LSA is a useful research tool but argued that it should not be regarded as a plausible theory of mind because it is based on learning from co-occurrence data." (p.303) The value of LSA "LSA does not handle all aspects of language processing, but it offers a biologically and psychologically plausible mechanistic explanation of the acquisition, induction, and representation of verbal meaning." (p.303) "The remarkable feature of LSA is that it can accomplish so much with so little" (p.307)

"word meanings are of paramount importance in discourse" (p.309)

LSA is a model of mind "LSA provides a model of how a great deal (but not all) of the relatively stable, shared, central meanings of words derives from experience with words in discourse, and LSA provides a model of how a great deal (but not all) of the meaning of discourse derives from the meaning of the words of which it is composed". (p.308)

How LSA work "First, LSA applies a transformation that factors in rates of in- dependent occurrence to reflect information rather than raw contiguity frequency. That is, it determines the degree to which a word discriminates one context from another and the degree to which a context predicts what words are more likely to occur in it than elsewhere". (p.305) "LSA and HAL's extensive successes in simulating human meaning-basedjudgments and behavior, starting with word-context co-occurrence data, offers incontrovertible evidence that such data can provide what is needed to induce much of the conventional meaning of words and combinations of words" (p.306) "principal-component extraction is a provable property of artificial neural networks with architectures similar or mathematically equivalent to LSA" (p.306)

At this point, unfortunately we cannot simulate a sample of different humans due to computational and data limitations (p.307)

Too large a sample, if larger than that encountered by individuals, could actually distort the results (p.307)

LSA will be helped when real-world linguistic, perceptual, and pragmatic knowl- edge beyond text can be made available to it. (p.307)

Application of LSA the rate of growth of children's recognition vocabulary, on the basis of learning from the same data as humans pp.306

simulate vocabulary growth, mirror word-similarity sorting, pass courses, mimic students with respect to which kinds of questions are the most difficult, predict comprehensibility oftext, reflect enough of meaning to grade a substantive essay as reliably as a human expert (p.307)

Inspiring "any model that covers too much must necessarily be too vague to be valuable" p.306

"A complete theory would include the dynamic psychological processes of turning received discourse into meaning and of turning ideas into discourse."(p.309)

Others conditioning psychological reality

Interesting sentences they are the best candidates for the job currently running.p.306

it help account for, or at least be consistent with p.306