## Weitere Artikel dieser Ausgabe durch Wischen aufrufen

Erschienen in:

01.04.2022 | INTELLIGENT SYSTEMS

# JSM Reasoning and Knowledge Discovery: Ampliative Reasoning, Causality Recognition, and Three Kinds of Completeness#

verfasst von: V. K. Finn

Einloggen, um Zugang zu erhalten

### Abstract

Inductive inferences and inferences, by analogy with JSM reasoning, are characterized as ampliative inferences generating new knowledge. New predicates for inductive inference rules and their ordering are considered. The case of a single-element effect for the predicate “X has effect Y” is also investigated.
Anhänge
Nur mit Berechtigung zugänglich
Fußnoten
1
The fifth level of acceptance of the generated hypotheses is an M-sequence of modal operators of rank r, where r > 1, representing empirical regularities that are obtained by applying JSM reasoning r-fold to sequences of extensible fact bases [1, p. 22].

2
Thus, the empirical patterns are a formalization of the idea of the knowledge-discovery process formulated in [2].

3
To understand the definition of knowledge in a computer system, one should become familiarized with its definition in [5, p. 31].

4
FB(0), FB(1), …, FB(s) is the sequence of expandable fact bases: FB(0) ⊂ FB(1) ⊂ … ⊂ FB(s).

5
Note that the degree of plausibility of a fact is 0 and that of a hypothesis is greater than 0, since it represents the number of steps of a plausible inference.

6
Since the truth values of JL formulas have the form $$\bar {v} = \left\langle {v,n} \right\rangle$$, where $$v \in \left\{ {1, - 1,0,\tau } \right\}$$, and $$n \in N$$, the logic is infinite-valued.

7
In the case |Y| =1, $${{J}_{{\left\langle {1,1} \right\rangle }}}\left( {{{V}_{i}} \Rightarrow _{2}^{{(P)}}{{Y}_{i}}} \right)$$ and $${{J}_{{\left\langle {1,0} \right\rangle }}}\left( {X \Rightarrow _{1}^{{(P)}}Y} \right)$$ are used.

8
That is, similarities for (+)-examples and (–)-examples.

9
For convenience of notation, we will represent finite sets {A, B, ..., P} as words AB...P; for example, we will represent {A, B, C} as АВС.

10
Therefore, the relation $$\Rightarrow _{2}^{*}$$ is not functional.

11
The premises of p.i.r.-1 are the corresponding elements of the diagram for $${{\Re }_{1}}$$, and the premises of p.i.r.-2 are the corresponding elements of the diagram for $${{\bar {\Re }}_{1}}$$.

12
In IS–JSM intelligent systems that implement the ARS JSM method [4, 19], D1,0(p) form a fact base.

13
If CCA(σ) are true, then the admissible JSM reasoning is strong [5].

14
This condition can be generalized to all Strx, y from [7]. The same holds true for the sufficient condition (**).

15
Obviously, for strong admissible JSM reasoning ρσ(s) = 1, where σ = +,–.

16
In Section 4 of this paper, possible strengthenings will be considered.

17
Since $$X{{ \Rightarrow }_{1}}Y$$ and $$V{{ \Rightarrow }_{2}}Y$$ are defined with respect to relational systems and their corresponding D0,1(P), these primitive predicates depend on the parameter Р.

18
In [8], J.S. Mill used the term agreement–difference. In the ARS JSM, the term similarity–difference is preferred.

19
The elements of M have a superscript σ, where σ = +, –, which, for the sake of notation, will sometimes be omitted.

20
Definitions of operations and are contained in the Appendix.

21
φ is obtained in the Appendix.

22
Indices (P) and (x, y) will sometimes be omitted for convenience of notation.

23
In the Appendix, the relational system $${{\bar {\Re }}_{f}}$$ is considered such that the predicates $$M_{{{{a}_{{12}}}fg,0}}^{\sigma }\left( {V,Y} \right)$$ are satisfiable in it.

24
Part II of this article will deal with the cases when |Y| > 1 and there is ¬α.

25
The content of this section assumes familiarity with papers [1, 5].

26
They are empirical nomological statements [1, 26].

27
This means strengthened realization of the scientific research demarcation criterion [27].

28
This definition is formulated for computer systems that implement the ARS JSM method.

29
We intend to consider the Boolean function F2() for the case |Y | > 1 in the second part of this article.

30
The sets $$\overline {Str} \times A_{E}^{\sigma }$$ are partially ordered with the largest and smallest element for σ = +, –, respectively.

31
Recall that $$A_{\chi }^{\sigma }$$ is an element of the intension of the concept of empirical regularity and $$A_{\chi }^{\sigma }(C{\kern 1pt} ',\,\,Q)$$ is an extension element of this concept.

32
Situational JSM reasoning is obviously in demand for the analysis of sociological data. Note also that there are cases when V = Ø or S = Ø.

33
We intend to consider this problem in the second part of this article.

Literatur
1.
Finn, V.K., Exact epistemology and artificial intelligence, Autom. Docum. Math. Linguist., 2020, vol. 54, pp. 140–173.   https://​doi.​org/​10.​3103/​S000510552003007​3 CrossRef
2.
Fayyad, U.M., Piatetsky-Shapiro, G., Smyth, P., and Uthurusamy, R., Advances in Knowledge Discovery and Data Mining, Cambridge: The AAAI Press, 1996.
3.
Finn, V.K., On the intellectual data analysis, Novosti Iskusstv. Intell., 2004, no. 3, pp. 3–17.
4.
Finn, V.K. and Shesternikova, O.P., The heuristics of detection of empirical regularities by JSM reasoning, Autom. Docum. Math. Linguist., 2018, vol. 52, pp. 215–247.   https://​doi.​org/​10.​3103/​S000510551805002​3 CrossRef
5.
Finn, V.K., On the heuristics of JSM research (additions to articles), Autom. Docum. Math. Linguist., 2019, vol. 53, pp. 250–282.   https://​doi.​org/​10.​3103/​S000510551905007​8 CrossRef
6.
Rosser, J.B. and Turquette, A.R., Many-Valued Logics, Amsterdam: North-Holland, 1958. MATH
7.
Finn, V.K., Distributive lattices of inductive JSM procedures, Autom. Docum. Math. Linguist., 2014, vol. 48, pp. 265–295.   https://​doi.​org/​10.​3103/​S000510551406002​8 CrossRef
8.
Mill, J.S., A System of Logic Ratiocinative and Inductive, Being a Connected View of Principles of Evidence and the Methods of Scientific Investigation, London: Parker, Son and Bowin, 1843. CrossRef
9.
Fann, K.T., Peirce’s Theory of Abduction, The Hague: Martinus Nijhoff Publishers, 1970. CrossRef
10.
Skvortsov, D.P., On some ways of constructing logical languages with quantors by tuples, Semiotika Inf., 1983, no. 20, pp. 102–126.
11.
Handbook of Mathematical Logic, Barvais, D., Ed., Amsterdam: North-Holland, 1977.
12.
Anshakov, O.M., Finn, V.K., and Skvortsov, D.P., On axiomatization of many-valued logics associated with formalization of plausible reasoning, Stud. Logica, 1989, vol. 48, pp. 423–447.   https://​doi.​org/​10.​1007/​BF00370198
13.
Finn, V.K., A form of argumentation logic, Autom. Docum. Math. Linguist., 1996, vol. 3, no. 3, pp. 3–27.
14.
Finn, V.K., On the class of JSM reasoning that uses the isomorphism of inductive inference rules, Sci. Tech. Inf. Process., 2017, vol. 44, pp. 387–396.   https://​doi.​org/​10.​3103/​S014768821706004​1 CrossRef
15.
Reichenbach, H., Elements of Symbolic Logic, New York: Macmillan, 1947. MATH
16.
Mal’tsev, A.I., Algebraicheskie sistemy (Algebraic Systems), Moscow: Nauka, 1970.
17.
Weingartner, P., Basic Question of Truth, Dordrecht: Kluver Academic Publishers, 2000. CrossRef
18.
Rescher, N., The Coherence Theory of Truth, Oxford: The Clarendon Press, 1973.
19.
Finn, V.K., Iskusstvennyi intellekt (metodologiya, primeneniya, filosofiya) (Artificial Intelligence: Methodology, Applications, and Philosophy), Moscow: LENAND, 2021, 2nd ed.
20.
Grätzer, G., General Lattice Theory, Berlin: Academic-Verlag, 1978. CrossRef
21.
Smullyan, R.M., First-Order Logic, New York: Springer-Verlag, 1968. CrossRef
22.
DSM-metod avtomaticheskogo porozhdeniya gipotez: logicheskie i epistemologicheskie osnovaniya (JSM Method of Automatic Hypotheses Generation: Logical and Epistemological Foundations), Moscow: Librokom, 2009.
23.
Frege, G., On sense and meaning, Logika i logicheskaya semantika (Logic and Logical Semantic), Moscow: Aspekt Press, 2000.
24.
Finn, V.K., On the non-Aristotelian concept structure, Logicheskie Issled., 2015, vol. 21, no. 1, pp. 9–48.   https://​doi.​org/​10.​21146/​2074-1472-2015-21-1-9-48
25.
Gillies, D., Artificial Intelligence and Scientific Method, New York: Oxford Univ. Press, 1996. MATH
26.
Reichenbach, H., Nomological Statement and Admissible Operations, Amsterdam: North-Holland, 1954. MATH
27.
Popper, K.R., Objective Knowledge: An Evolutionary Approach, Oxford: Clarendon Press, 1979.
28.
McCarthy, D. and Hayes, P.J., Some philosophical problems from the standpoint of artificial intelligence, Mach. Intell., 1979, no. 4, pp. 463–502.
29.
Bridgman, P.W., The nature of some of our physical concepts: I, Br. J. Philos. Sci., 1951, vol. 1, no. 4, pp. 257–272.   https://​doi.​org/​10.​1093/​bjps/​I.​4.​257 CrossRef
30.
Kant, I., Kritik der reinen Vernunft, Riga: Johann Friedrich Hartknoch, 1781.
31.
Finn, V.K., J.S. Mill’s inductive methods in artificial intelligence systems. Part I, Sci. Tech. Inf. Process., 2011, vol. 37, no. 6, pp. 385–402.   https://​doi.​org/​10.​3103/​S014768821106003​7 CrossRef
32.
Finn, V.K., J.S. Mill’s inductive methods in artificial intelligence systems. Part II, Sci. Tech. Inf. Process., 2012, vol. 39, no. 5, pp. 241–260.   https://​doi.​org/​10.​3103/​S014768821205003​6 CrossRef
33.
Finn, V.K. and Mikheyenkova, M.A., Situation extension of the JSM automatic hypotheses generation method, Autom. Docum. Math. Linguist., 2000, vol. 34, no. 6, pp. 11–26.
34.
Finn, V.K. and Shesternikova, O.P., A new variant of the generalized JSM-method for automatic support of scientific research, Sci. Tech. Inf. Process., 2017, vol. 44, no. 5, pp. 338–344.   https://​doi.​org/​10.​3103/​S014768821705004​5 CrossRef
35.
Anshakov, O.M., Skvortsov, D.P., and Finn, V.K., On deductive imitation of some variants of the JSM method of automatic hypotheses generation, Semiotika Inf., 1988, no. 33, pp. 164–233.
36.
Finn, V.K., Foreword to the second edition, Iskusstvennyi intellekt (Artificial Intelligence), Moscow; LENAND, 2021, 2nd ed., pp. 14–31.
Titel
JSM Reasoning and Knowledge Discovery: Ampliative Reasoning, Causality Recognition, and Three Kinds of Completeness#
verfasst von
V. K. Finn
Publikationsdatum
01.04.2022
Verlag