[Cover] [Contents] [Index] Previous page Next Section

Page 244
presence of Uap-type positive additional information of density d1, even if the scientist is also given the best possible negative additional information (Uacp-type density 1) and is only required to output a grammar for a finite variant of the language.
10.49 Proposition Suppose 0244-001.gif. Then, 0244-002.gif
The proof of the proposition can be found in Jain and Sharma [85].
We next present a negative additional information counterpart of the previous proposition, and illustrate the power of Acp-type negative additional information over lower density Uacp-type negative additional information. Again, the reader is referred to Jain and Sharma [85] for a proof.
10.50 Proposition Suppose 0244-003.gif. Then, 0244-004.gif.
The following corollary summarizes the complete relationship between all the paradigms introduced in this subsection. It follows from the results presented above, as the reader may check.
10.51 Corollary Let a 0244-005.gif and d, d1, d2, d3, 0244-006.gif, 0244-007.gif.
(a) 0244-008.gif
(b) 0244-009.gif
(c) 0244-010.gif
(d) 0244-011.gif
(e) 0244-012.gif
(f) 0244-013.gif
§10.4 Bibliographic Notes
The subject of upper bound on program size as additional information for function identification was first considered by Freivalds and Wiehagen [67] and generalized by Jain and Sharma [86]. The latter paper also investigated analogs of these paradigms for Bc, TxtEx, TxtFex, and TxtBc paradigms.

 
[Cover] [Contents] [Index] Previous page Next Section