1: %------------------------------------------------------------------------
2: % Required conference format WinWord6, file: scm00vlx.doc (23/12/99)
3: % Rewritten for LaTeXe format, file: scm00vlx.tex (31/12/99 - 1/1/00)
4: %------------------------------------------------------------------------
5: \documentclass[12pt,a4paper]{article}
6: \newcommand{\idx}[1]{\mbox{\scriptsize [}#1\mbox{\scriptsize ]}}
7: \newcommand{\Idx}[1]{\mbox{\small [}#1\mbox{\small ]}}
8: \newcommand{\hc}{{\scriptscriptstyle{+}}}
9: \newcommand{\pht}[1]{\mbox{\phantom{#1}}}
10: \newcommand{\pow}{\raisebox{0.4ex}{\small$\,\uparrow\,$}}
11:
12: \setlength{\topmargin}{-1.2cm}
13: \setlength{\oddsidemargin}{-1cm}
14: \setlength{\evensidemargin}{-1cm}
15: \setlength{\textwidth}{18cm}
16: \setlength{\textheight}{24.5cm}
17:
18: %BinPic version 1.5-G 8/12/1999
19: \newdimen{\pxh}
20: \newdimen{\pxv}
21: \newcount{\BinPicN}
22: \pxh=0.48bp
23: \pxv=0.48bp
24: \def\binpic#1{{%
25: \def\h{1}%
26: \def\L##1{\hbox{{\R0}##1}\nointerlineskip}%
27: \def\V##1{\vskip##1\pxh\relax}%
28: \def\R##1{\vrule width##1\pxh height\h\pxv depth0pt}%
29: \def\S##1{\hskip##1\pxh\relax}%
30: \def\H##1##2{{\def\h{##1}{##2}}}%
31: \def\N##1##2{{\BinPicN=\number##1\loop\ifnum\BinPicN>0{##2}\advance\BinPicN by-1\repeat}}%
32: \def\0{\S4}\def\1{\S3\R1}\def\2{\S2\R1\S1}\def\3{\S2\R2}%
33: \def\4{\S1\R1\S2}\def\5{\S1\R1\S1\R1}\def\6{\S1\R2\S1}\def\7{\S1\R3}%
34: \def\8{\R1\S3}\def\9{\R1\S2\R1}\def\A{\R1\S1\R1\S1}\def\B{\R1\S1\R2}%
35: \def\C{\R2\S2}\def\D{\R2\S1\R1}\def\E{\R3\S1}\def\F{\R4}%
36: \hbox{\vbox{#1}}%
37: }}%
38: \def\picscale#1#2#3{{%
39: \ifx#1?\relax\else\pxh=#1\fi%
40: \ifx#2?\relax\else\ifx#2*\pxv=\pxh\else\pxv=#2\fi\fi%
41: #3%
42: }}%
43:
44:
45: \begin{document}
46: \pagestyle{empty}
47: \title{\bf\Large Von Neumann Quantum Logic {\em vs.}
48: Classical von Neumann Architecture?}
49: \author{{\sc A. Yu. Vlasov}\\
50: Federal Radiological Center, IRH\\
51: Mira Street 8, St.-Petersburg, Russia}
52: \date{}
53: \twocolumn[
54: \smash{$\mbox{\vtop{\vskip 4cm\relax{\Large\em Conference}\\
55: \mbox{\hskip 1pt\picscale{0.24bp}*{%
56: % 13 90
57: % prn2texp v3, BinPic v1.3/1.4 compatible output
58: \binpic{%
59: \H3{\L{\0}}%
60: \L{\S{60}\1\4}%
61: \L{\S{28}\2\8\S{24}\2\A}%
62: \L{\S{28}\5\4\S{24}\5\5}%
63: \L{\S{28}\A\A\S{24}\A\A\8}%
64: \L{\S{24}\1\5\5\S{24}\5\5}%
65: \L{\S{24}\2\A\A\S{24}\A\A\8}%
66: \L{\S{24}\1\5\5\S{20}\1\5\5\4}%
67: \L{\S{24}\2\A\A\S{24}\A\A\8}%
68: \L{\S{24}\5\5\5\S{20}\1\5\5\4}%
69: \L{\S8\N2{\S{16}\2\A\A\8}}%
70: \L{\S{24}\5\5\5\S{20}\1\5\5\4}%
71: \L{\S{24}\2\A\A\8\S{16}\2\A\A\A}%
72: \L{\S{24}\5\5\5\4\S{16}\1\5\5\4}%
73: \L{\S{24}\A\A\A\8\S{16}\2\A\A\A}%
74: \L{\S{24}\5\5\5\4\S{16}\5\5\5\4}%
75: \L{\S{24}\A\A\A\8\S{16}\2\A\A\A}%
76: \L{\S{20}\1\5\5\5\4\S{16}\N4{\5}}%
77: \L{\S{24}\N4{\A}\S{16}\2\A\A\A}%
78: \L{\S{20}\1\5\5\5\4\S{16}\N4{\5}}%
79: \L{\S{24}\N4{\A}\S{16}\N4{\A}}%
80: \L{\S{20}\1\5\5\5\4\S{16}\N4{\5}}%
81: \L{\S{24}\N4{\A}\S{16}\N4{\A}}%
82: \L{\S{20}\1\5\5\5\4\S8\0\1\N4{\5}}%
83: \L{\S{20}\2\N4{\A}\S{16}\N4{\A}\8}%
84: \L{\S{20}\1\5\4\5\5\S8\0\1\5\4\5\5}%
85: \L{\S{20}\2\A\8\A\A\S{16}\A\A\2\A\8}%
86: \L{\S8\N2{\S{12}\1\5\4\5\5}}%
87: \L{\S{20}\2\A\8\A\A\S8\0\2\A\8\2\A\8}%
88: \L{\S{20}\5\5\4\5\5\S8\0\1\5\4\5\5\4}%
89: \L{\S{20}\N2{\2\A\8}\S8\N2{\2\A\8}}%
90: \L{\S{16}\7\F\D\0\5\5\S8\0\5\5\0\1\5\7\R8}%
91: \L{\S8\0\3\R8\F\C\3\E\8\S8\2\A\8\2\R{20}\0\7}%
92: \L{\S8\0\R{20}\7\D\4\S8\5\5\0\1\R{20}\C\F}%
93: \L{\S8\1\F\E\A\R8\B\E\8\S8\A\A\8\7\F\C\0\1\F\E\F}%
94: \L{\S8\3\F\0\5\5\R8\D\4\0\1\5\5\0\R8\4\S8\3\R8}%
95: \L{\S8\7\E\0\A\A\7\F\E\8\S8\A\A\3\F\E\8\S8\1\R8}%
96: \L{\S8\F\C\1\5\5\3\F\D\4\0\1\5\4\7\F\D\4\0\S8\R8}%
97: \L{\0\1\F\C\0\A\A\3\F\E\8\S8\A\A\7\F\A\A\0\S8\7\F}%
98: \L{\0\1\F\8\1\5\4\0\F\D\4\0\1\5\4\R8\5\4\0\S8\7\F}%
99: \L{\0\3\F\8\2\A\A\0\F\E\A\0\2\A\9\F\C\A\A\0\S8\3\F}%
100: \L{\0\3\F\8\1\5\4\0\7\D\4\0\1\5\7\F\D\5\5\0\S8\3\F}%
101: \L{\0\3\F\8\2\A\8\0\7\E\A\0\2\A\B\F\8\A\A\0\S8\1\F}%
102: \L{\0\7\F\8\5\5\4\0\7\D\5\0\5\5\7\F\8\5\5\S{16}\F}%
103: \L{\0\7\F\8\2\A\8\0\3\E\A\0\A\A\R8\0\A\A\S{16}\F\8}%
104: \L{\0\7\F\8\5\5\4\0\3\D\5\0\5\5\R8\0\5\5\S{16}\F\8}%
105: \L{\0\7\F\C\2\A\8\0\1\E\A\8\A\B\R8\0\A\A\8\S8\0\F\8}%
106: \L{\0\7\F\C\5\5\4\0\1\F\5\0\5\5\F\E\0\5\5\S{16}\7\8}%
107: \L{\0\3\F\C\A\A\8\0\1\E\A\8\A\B\F\E\0\2\A\8\S8\0\7\8}%
108: \L{\0\3\F\E\5\5\S8\0\F\5\1\5\5\F\E\0\5\5\S{16}\7\8}%
109: \L{\0\3\R8\A\A\8\S8\2\A\A\A\B\F\E\0\2\A\8\S8\0\3\8}%
110: \L{\0\1\R8\D\5\S8\0\N4{\5}\7\F\C\0\5\5\4}%
111: \L{\0\1\R8\F\A\S8\0\2\A\A\A\B\F\C\0\2\A\8}%
112: \L{\0\1\R{16}\S8\0\1\5\5\5\7\F\C\0\1\5\4}%
113: \L{\S8\R{16}\8\S8\2\A\A\A\R8\C\0\2\A\A}%
114: \L{\S8\7\F\R8\E\S8\1\5\5\5\7\F\C\0\1\5\4}%
115: \L{\S8\3\F\R8\F\8\S8\A\A\A\R8\C\0\2\A\A}%
116: \L{\S8\1\R{20}\0\1\5\5\5\7\F\C\0\1\5\5}%
117: \L{\S8\0\7\R{16}\8\0\A\A\A\R8\C\S8\A\A}%
118: \L{\S8\0\3\R{16}\E\1\5\5\5\7\F\C\0\1\5\5}%
119: \L{\S8\0\2\R{20}\0\A\A\A\7\F\C\S8\A\A}%
120: \L{\S8\0\1\5\F\R8\F\8\5\5\4\7\F\C\0\1\5\5}%
121: \L{\S8\0\2\A\B\R8\F\C\A\A\A\7\F\C\S8\A\A\8}%
122: \L{\S8\0\1\5\4\R8\F\E\5\5\4\7\F\C\S8\5\5}%
123: \L{\S8\0\2\A\8\1\R8\E\2\A\8\7\F\C\S8\A\A\8}%
124: \L{\S8\0\1\5\4\0\F\R8\1\5\4\7\F\C\S8\5\5\4}%
125: \L{\S8\0\2\A\8\0\7\R8\2\A\8\7\F\C\S8\2\A\8}%
126: \L{\S8\0\1\5\S8\3\R8\1\5\0\3\F\E\S8\1\5}%
127: \L{\0\E\S8\A\0\S8\R8\S8\0\3\F\E\S8\2\A}%
128: \L{\0\E\S{24}\R8\8\S8\3\F\E\S8\0\4}%
129: \H2{\L{\0\E\S{24}\7\F\8\S8\1\R8}}%
130: \L{\0\F\S{24}\7\F\8\S8\1\R8\S{32}\C}%
131: \L{\0\F\S{24}\7\F\8\0\S8\R8\8\S{24}\1\E}%
132: \L{\0\F\8\S{20}\7\F\S{16}\R8\8\S{24}\1\C}%
133: \L{\0\F\C\S{20}\7\F\S{16}\7\F\C\S{24}\3\C}%
134: \L{\0\F\C\S{20}\7\F\S{16}\7\F\E\S{24}\7\8}%
135: \L{\0\F\E\S{20}\7\E\S{16}\3\F\E\S{24}\7\8}%
136: \L{\0\F\E\S{20}\7\E\S{16}\1\R8\S{24}\F}%
137: \L{\0\R8\S{20}\7\E\S{16}\1\R8\8\S{16}\1\F}%
138: \L{\0\R8\8\S{16}\F\C\S{20}\R8\C\S{16}\3\E}%
139: \L{\0\R8\C\S{16}\F\8\S{20}\7\R8\S{16}\7\C}%
140: \L{\0\F\R8\S8\0\3\F\S{24}\3\R8\C\S8\3\F}%
141: \L{\0\F\R8\8\S8\7\E\S{28}\R8\F\8\1\R8}%
142: \L{\0\F\3\R8\0\3\F\C\S{28}\7\R{20}\C}%
143: \L{\0\F\1\R{20}\8\S{28}\1\R{20}\8}%
144: \L{\0\E\0\3\R8\F\C\S{32}\0\7\R8\F\C}%
145: \L{\S{16}\7\F\E\0\S{40}\R8\E}%
146: }%
147: }}\mbox{\huge$\!${\em '2000}} }}$}
148: \maketitle
149: \thispagestyle{empty}
150: {\small
151: %\begin{abstract}
152: \begin{quote}
153: {\em Abstract. }
154: The name of John von Neumann is common both in quantum mechanics and
155: computer science. Are they really two absolutely unconnected areas?
156: Many works devoted to quantum computations and communications are
157: serious argument to suggest about existence of such a relation, but it
158: is impossible to touch the new and active theme in a short review.
159: In the paper are described the structures and models of linear algebra
160: and just due to their generality it is possible to use universal description
161: of very different areas as quantum mechanics and theory of Bayesian
162: image analysis, associative memory, neural networks, fuzzy logic.
163: \end{quote}
164: %\end{abstract}
165: }
166: \medskip
167: ]
168: \sloppy
169:
170: J. von Neumann is considered as one of ``fathers'' of modern
171: computers due to his theoretical works and elaboration of computer
172: EDVAC at 1945--50, but already 20 year before it, at \mbox{1926--31} he
173: participated in drawing yet another kind of logic --- logic of quantum
174: mechanics and only now the different areas of research of one
175: scientist seem going to meet and born new family of ultrascale
176: cybernetic devices --- {\em quantum computers}.
177:
178: Currently the {\em quantum information science} exists mainly as
179: theoretical area of research and size of quantum registers does not
180: exceed of 3--5 quantum bits ({\em qubits}), but in the paper is
181: considered question: does the quantum logic has some useful
182: application as {\em an abstract mathematical model} in computer
183: science? Such applications of physical models nowadays are not
184: unusual, for example Boltzmann machines and other methods came
185: to area of artificial neural networks from statistical physics [1, 2].
186:
187: It is also useful to draw some analogy with fuzzy sets and logic [3].
188: Here is used discrete representation of 2D sets on some lattice due to
189: understanding analogy with bit-map pictures.
190:
191: Let us introduce few data types in some ``Pascal-like'' notation to
192: come from fuzzy sets to quantum registers. Here $n$ and $N = 2^n$
193: are \mbox{\bf integer} numbers.
194:
195: 1) $x$ : {\bf dot} = [1 .. N, 1 .. N];
196:
197: Here $x$ is represented as point on 2D space for simplicity,
198: because of digitization we can consider space of any dimension as
199: interval of natural numbers $x$~:~1~..~$N^2$. In this
200: example $x$ is visual model for $n + n = 2n$-bits register.
201:
202: 2) $I$ : {\bf set\_2D} = {\bf set of dot};
203:
204: or
205:
206: \pht{2)} $I$ : {\bf set\_2D} = {\bf array} [{\bf dot}] {\bf of Boolean};
207:
208: The set, ``image'' can be considered as black and white picture --- black
209: set on white sheet. It is necessary $2^{N\cdot N}$ {\em i.e.}
210: $2\pow(2\pow(2n))$ bits for the set.
211:
212: 3) $F$ : {\bf fuzzy$\!$\_set} = {\bf array} [{\bf dot}] {\bf of real};
213:
214: The {\bf real} is interval [0.0 .. 1.0] of real numbers represented with
215: some finite precision $d$. It can be considered as an analogue of
216: gray-scaled picture. The fuzzy set also can be {\em standardized} by
217: condition $\sum_x{F\idx{x}} = 1$ where $\sum_x$ in example under consideration
218: is double sum:
219: \begin{equation}
220: \sum_{i,j=1}^N{F_{ij}} = 1
221: \end{equation}
222: % (1)
223: (in continuous case an integration used instead of the summation and
224: distributions $F(x)$ are used instead of arrays).
225:
226: The condition (1) is used in statistical interpretation of fuzzy set,
227: when it is suggested to choose some point of set with probability
228: proportional of {\em adequacy function} $F\idx{x}$, and sum
229: of all the probabilities is unit. Let us use notation $p\idx{x}$
230: for the probabilities.
231:
232: \medskip
233:
234: Now let us introduce notion of {\em quantum set }(it is not standard
235: term, mathematical object discussed further corresponds to
236: $2n$-{\em qubit register} in quantum information science [6] or quantum
237: mechanical system with $N^2$ states [7]):
238:
239: 4) $Q$ : {\bf qu\_set} = {\bf array} [{\bf dot}] {\bf of complex};
240:
241: Where {\bf complex} is $q = u + i\,w$, $|q| \le 1$.
242: Instead of standardization condition here is used {\em normalization}:
243: $\sum_x{|Q\idx{x}|^2} = 1$ {\em i.e.}:
244: \begin{equation}
245: \sum_{i,j=1}^N{\left|Q_{ij}\right|^2} = 1
246: \end{equation}
247: % (2)
248: for example under consideration.
249:
250: In quantum mechanics the complex numbers $q\idx{x}$ is
251: called {\em amplitudes}, related with classical probabilities as
252: $p= |q|^2 = u^2 + w^2$, {\em i.e.} there
253: is {\em standardized} fuzzy set related with given {\em normalized}
254: quantum set via formula: $p\idx{x} = |q\idx{x}|^2$.
255:
256: It is possible for simplification to consider case with real amplitudes
257: ($w~=~0$), and let us explain why an ``auxiliary'' fuzzy set with
258: $q\idx{x}=\sqrt{p\idx{x}}$ has some independent useful application.
259:
260: Let us consider two sets $p_1\idx{x}$, $p_2\idx{x}$ and look for some
261: {\em likelihood } function $H(p_1, p_2)$ with properties:
262: $H(p_1,p_2) < 1$ for $p_1\ne p_2$, $H(p,p) = 1$ [1, 3]. For standardized
263: sets such function can be
264: chosen as $H(p_1,p_2) = \sum_x{(p_1\idx{x} p_2\idx{x})}^{1/2}$. If
265: we use quantum sets $q_1$, $q_2$ ($p_1=q_1^2$, $p_2=q_2^2$),
266: the formula is $H(q_1, q_2) = \sum_x{q_1\idx{x} q_2\idx{x}}$.
267:
268: In our example $x$ was chosen as multi-index of 2D array mostly
269: for simple visualization and any array can be described as one-dimensional.
270: Here $x = (i, j)$ can be substituted by
271: index of the 1D array like $K~=~(i-1)N+j$,
272: $K \in [1\ ..\ N^2]$, then formula can be written
273: as:
274: \begin{equation}
275: H(q,q')=\sum_{i,j=1}^N{q_{ij} q'_{ij}} \equiv \sum_{K=1}^{N^2}{q^{\,}_K q'_K}
276: \end{equation}
277: % (3)
278: and it shows, that $H(q,q')$ is simply scalar
279: product of two vectors with $N^2$ elements and unit lengths
280: (due to normalization condition) and so the function really can be
281: unit only for equivalent vectors.
282:
283: It should be mentioned also, that the formula (3) is given for {\em
284: real} $q^{\,}_K$ for simplicity and in {\em complex} case it contains
285: terms with complex conjugation: $q^{\,}_K\bar{q}'_{K}$. It is {\em Hermitian}
286: norm [7]. Such case also has applications, for example then we work with
287: {\em complex Fourier} {\em transform} of some image.
288:
289: The discussed property of quantum set as ``square root of fuzzy set''
290: makes clear, why it can be useful in such abstract area, as image
291: recognition [4, 5], quite far from initial appearance in quantum
292: mechanics.
293:
294: Let us discuss now some question related with ``hardware''. We had
295: few data structures: \\
296: 1) $x$ : {\bf dot}, 2) $I\idx{x}$ : {\bf set\_2D},
297: 3) $p\idx{x}$ : {\bf fuzzy$\!$\_set}\\
298: It is sequence with more and more complicated structure with
299: occupation of more and more computer memory. But let us suggest,
300: that register $x$ is permanently changing its value by such a law,
301: that after enough period of time $T$ it can be found, the $x$
302: had value $v$ during time $t\idx{v}$ and
303: $\lim\limits_{T\to\infty}t\idx{v}/T=p\idx{v}$.
304:
305: The algorithm can be implemented by software, but it also can be
306: considered as some hardware register (like implementation of random
307: number generator in some computers with main difference, that
308: $p\idx{v}$ depends on $v$; or input port of some analog-to-digital
309: converter is scanning some ``physical model'' of fuzzy set $p$).
310:
311: Then each access to the register produces some value of $x$ with
312: probability $p\idx{x}$, so one $2n$-bits stochastic register
313: is enough to implement statistical model of standardized fuzzy set
314: discussed earlier.
315:
316: \medskip
317:
318: Now let us come to {\bf qu\_set} data type. Why it can be modeled by
319: one quantum register? The example with stochastic register as model
320: of fuzzy set is some analogy (see also [8]). But procedure of access to
321: such register due to laws of quantum mechanics has some differences
322: with classical statistical register discussed below.
323:
324: Let us consider some value $q \in$ {\bf qu\_set} of the register, it is
325: array of numbers $q\idx{x}$, we may not read all the numbers,
326: but if we access to the register, we read number $x$ with
327: probability $p\idx{x} = |q\idx{x}|^2$ and it
328: coincides with functionality of stochastic register described above.
329:
330: It should be mentioned only, that any access to $q$ destroys the
331: quantum register by substitution instead of $q$ new array with 1
332: in element with index $x$ and with all other is 0, and so the
333: register should be reset ({\em preparation} in terminology of
334: quantum mechanics) in $q$ after each access (quantum {\em
335: measurement}).
336:
337: But the quantum register has other useful property, it is possible
338: instead of simple access described above to perform another
339: operation, we prepare some given $q' \in$ {\bf qu\_set} and read
340: the register $q$ with using $q'$ as some ``quantum bit-mask'',
341: then with probability $|H(q,q')|^2$ (see
342: Eq. 3) operation is successful and so by repeating it more times we
343: may found $|H|$ with more precision.
344:
345: Because the $|H|$ has useful application as likelihood function,
346: the quantum register can be used as some hardware accelerator for
347: image analysis. Currently such hardware is not accessible and so it
348: was interesting to research advantages and disadvantages of the
349: particular function $|H|$ in usual software applications.
350:
351: It is promising not only because such kind of software would suffer
352: giant speed-up after creation specific quantum hardware, but also
353: because the used mathematical constructions and methods of linear
354: algebra are quite convenient and powerful.
355:
356: It should be mentioned, that similar mathematical methods already
357: was used in models of associative memory [9] and formal neural
358: network [10] without any relation with quantum mechanical models.
359: Only noticeable difference was using {\em real} linear spaces instead
360: of {\em complex} and Euclidean norm (Eq. 3) instead of Hermitian
361: (with complex conjugation).
362:
363: \medskip
364:
365: Let us now consider some operations with fuzzy and quantum sets,
366: discuss fuzzy and quantum {\em logic}.
367:
368: For usual sets we have basic operations for $A$,~$B$: {\bf set\_2D},
369: --- {\em intersection}: $A\cap B$, {\em union}: $A\cup B$, {\em
370: complementation}: $A^c$. With using presentation of set as
371: Boolean array the operations in components can be written:
372: $(a^c)\idx{x}~=~\mbox{\bf not}~a\idx{x}$, $(a\cap b)\idx{x} = a\idx{x}\land b\idx{x}$,
373: $(a\cup b)\idx{x} = a\idx{x} \lor b\idx{x}$.
374:
375: Similarly it is possible to define operations with fuzzy sets by
376: definition of {\em real }analogs of Boolean operations {\bf not }($\lnot$),
377: {\bf and} ($\land$), {\bf or} ($\lor$).
378:
379: For example $\lnot a \mapsto 1-a$, $a \land b \mapsto \min(a,b)$,
380: $a \lor b \mapsto \max(a,b)$ is a good choice, but here is used a second one,
381: more algebraic $a\land b \mapsto a\cdot b$,
382: $a \lor b \mapsto a+b-ab = \lnot ((\lnot a) \land (\lnot b))$.
383:
384: But {\bf qu\_set} introduced above is not directly used in quantum
385: logic --- the linear operators are used here instead of vectors:
386:
387: $L$ : {\bf qu\_map} = {\bf array} [{\bf dot},{\bf dot}] {\bf of complex};\\
388: It is matrix for linear map: {\bf qu\_set} $\to$ {\bf qu\_set}:
389: \begin{equation}
390: q'_{ij}= \sum_{k,l=1}^N{L^{\,}_{ij,kl} q'_{kl}} \quad \mbox{or} \quad
391: q'_I = \sum_{K=1}^{N^2}{L^{\,}_{IK} q'_K}
392: \end{equation}
393: % (4)
394: where indexes $I$, $K$ are used instead of multi-indexes
395: \idx{$i,j$} and \idx{$k,l$} (let us for simplicity use
396: further the indexes like $I, K : [1\ ..\ M]$, $M=N^2$).
397:
398: The operators $A, B,\ldots \in$ {\bf qu\_map} form an
399: algebra with usual matrix multiplication {\em C=AB}:
400:
401: \begin{equation}
402: C_{KJ}= \sum_{I=1}^M{A_{KI} B_{IJ}}
403: \end{equation}
404: % (5)
405:
406: A special kind of operators, {\em projectors}, make possible
407: comparison of the algebra with logic, {\em i.e.} Boolean algebra. The
408: projector is operator with property $P^2 = P$. Let us
409: consider set of orthogonal projectors, {\em i.e.}
410: $P_i P_j = 0$, $i\ne j$, then the operators produce
411: Boolean algebra in respect of operations:
412: \begin{equation}
413: \lnot P{\equiv}1{-}P,\, P{\land}R{\equiv}PR,\, P{\lor}R{\equiv}P{+}R{-}PR %(6)
414: \end{equation}
415:
416: Elements of the algebra have form $P_{(S)}$ there $S$~:~{\bf set of}~1~..~$M$:
417: \begin{equation}
418: P_{(S)}= \sum_{I \in (S)} {P_I}
419: \end{equation}
420: % (7)
421:
422: There is relation between $q\in$ {\bf qu\_set} and some projector
423: $P_q{\in}${\bf qu\_map}: $P_q\Idx{I,J}=q\Idx{I}q\Idx{J}$.
424: To describe properties of the
425: projector it is convenient together with $q$ considered as {\em
426: row} with $M$ elements to consider transposed {\em column}
427: $q^\hc$ (conjugated for complex case).
428:
429: Then $P_q = q^\hc \cdot q$ and $q\cdot q^\hc = |H(q,q)|^2~=~1$ (the row
430: $q$ can be considered as $1{\times}M$ matrix, column $q^\hc$ as
431: $M{\times}1$ matrix and due to law of
432: multiplication the $q\cdot q^\hc$ is $1{\times1}$ matrix {\em
433: i.e.} number and $q^\hc\cdot q$ is $M{\times}M$ matrix) and
434: \mbox{$P_q P_q = q^\hc\cdot q\cdot q^\hc\cdot q = q^\hc\cdot1\cdot q =
435: q^\hc\cdot q = P_q$}.
436:
437: If we consider family of nonintersected sets $q_1$, $q_2, \ldots, q_k$
438: then $P_i = q_i^\hc\cdot q_i$ are orthogonal projectors and so
439: quantum sets in such representations have rather relations with usual
440: logic than with fuzzy one.
441:
442: For more clear explanation of properties of $P_i$ it is
443: possible to use existence of some orthogonal (unitary for complex
444: case) matrix $U$ same for all $P_i$ such, that all
445: $P'_i = U P_i U^{-1}$ are diagonal and have
446: very simple form: $P'_1= \mbox{diag}(1,0,$\ldots$,0)$,
447: $P'_2 = \mbox{diag}(0,1,\ldots,0)$, {\em etc.}.
448:
449: The classical Boolean structure of the operators $P_i$ and
450: their sums $P_{(S)}$ (see Eq. 7) is because of all the
451: operators commute [11]. If to choose projectors $P$, $R$:
452: $PR\ne RP$ the Eq. 6 do not produce Boolean algebra, but it
453: is other kind of non-Boolean logic, than fuzzy one.
454:
455: It should be mentioned, that more direct relation with fuzzy set have
456: so-called {\em mixed} quantum states $R = \sum_i{w_i P_i}$,
457: $\sum_i{w_i} = 1$ where $w_i$ have statistical nature and so here is
458: written analog of standardized fuzzy set.
459:
460: A representation of some kind of fuzzy operations is example with
461: family of commuting operators [12], but not projectors. They are
462: described by diagonal matrices: $D_q\Idx{I,I} = q\Idx{I}$,
463: $D_q\Idx{I,J} = 0$, $I\ne J$ and already shown Eq. 6.
464:
465: \smallskip
466:
467: {\em Bibliographical notes}: In addition to references [7, 11] some
468: general handbook on quantum mechanics like [13] is appropriate for
469: most text of the paper. New area of quantum computation is
470: presented in [6, 14--17]. Two works
471: of J. von Neumann devoted to quantum logic [18] and electronic
472: computers [19] are included for completeness.
473:
474: {\footnotesize
475: \renewcommand{\refname}{{\hfil\bf\normalsize References\hfil}}
476: \begin{thebibliography}{99}
477: \setlength{\parsep}{0pt}
478: \setlength{\itemsep}{2pt}
479: \bibitem[1]{1}G. Winkler, {\em Image Analysis, Random Fields and Dynamic
480: Monte Carlo Methods}, Springer, 1995
481: %
482: \bibitem[2]{2}B. Kosko, {\em Neural Networks and Fuzzy Systems}, Prentice-Hall
483: Int., 1992
484: %
485: \bibitem[3]{3}T. Terno, {\em etc.} (editors), {\em Applied Fuzzy Systems}, Mir,
486: M. 1993 [Russ.], Omsa, Tokyo, 1989 [Jap.]
487: %
488: \bibitem[4]{4}A. Yu. Vlasov, ``Quantum Computations and Images Recognition,''
489: {\em Conference QCM'96}, full paper: quant-ph/9703010
490: %
491: \bibitem[5]{5}A. Yu. Vlasov, ``Analogue Quantum Computers for Data Analysis,''
492: quant-ph/9802028
493: %
494: \bibitem[6]{6}C. H. Bennett, ``Quantum Information and Computation,''
495: {\em Physics Today} {\bf 48} (1995) 24
496: %
497: \bibitem[7]{7}A. I. Kostrikin, Yu. I. Manin,{\em Linear Algebra and Geometry},
498: Nauka, M. 1986 [Russ.]
499: %
500: \bibitem[8]{8}R. R. Zapatrin, ``Logic Programming as Quantum Measurement'',
501: {\em Int. J. Theor. Phys.} {\bf 34} (1995) 1813
502: %
503: \bibitem[9]{9}T. Kohonen, {\em Associative Memory: A System- Theoretical
504: Approach}, Springer 1978 [Mir, M. 1980]
505: %
506: \bibitem[10]{10}E. N. Sokolov, G. G. Vatkyavichus, {\em The neuro-intelligence:
507: from neuron -- towards neurocomputer}, Nauka, M. 1988 [Russ.]
508: %
509: \bibitem[11]{11}A. A. Grib, {\em Violation of Bell's Inequalities and Problem of
510: Measurements in Quantum Theory}, JINR P2--92--211, Dubna 1992 [Russ.]
511: %
512: \bibitem[12]{12}A. Yu. Vlasov, ``Representation and Processing of Information in
513: Quantum Computers,'' {\em Tech. Phys. Lett.} {\bf 20} (1994) 992
514: \newblock [{\em Pis'ma Zh. Tekh. Fiz.} {\bf 20} (1994) 45]
515: %
516: \bibitem[13]{13}Landau and Lifshitz, Course of Theoretical Physics, III (Quantum
517: Mechanics) Nauka, M. 1989 [Russ.]
518: %
519: \bibitem[14]{14}R. P. Feynman, ``Quantum-Mechanical Computers,''
520: {\em Found. Phys. } {\bf 16} (1986) 507
521: %
522: \bibitem[15]{15}D. Deutsch, ``Quantum Theory, the Church-Turing Principle and the
523: Universal Quantum Computer,'' {\em Proc. R. Soc. London }{\bf A 400}
524: (1985), 97
525: %
526: \bibitem[16]{16}D. Deutsch, ``Quantum Computational Networks,''
527: {\em Proc. R. Soc. London }{\bf A 425} (1989), 73
528: %
529: \bibitem[17]{17}D. Deutsch, A. Ekert, R. Lupacchini, ``Machines, Logic and
530: Quantum Physics,'' math/9911150
531: %
532: \bibitem[18]{18}G. Birkhoff, J. v. Neumann, ``The logic of quantum mechanics,''
533: {\em Ann. Math.} {\bf 37} (1936) 823
534: %
535: \bibitem[19]{19}A. Burks, H. Goldstine, J. v. Neumann, {\em Preliminary discussion
536: of the logical design of an electronic computing instrument},
537: Princeton 1946
538: \end{thebibliography}
539: }
540: \end{document}
541:
542:
543: