On the State Complexity of Semi-Quantum
Finite Automata
Shenggen Zheng1 (Timothy), Jozef Grusk๐1 , Daowen Qi๐ข2
1 Faculty of Informatics, Masaryk University, Czech Republic
2 Department of Computer Science, Sun Yat-sen University, China
arXiv:1307.2499
2014/3/17
27
1
1.Introduction
2014/3/17
27
2
1.Introduction
๏ Peter Shorโs algorithm (1994)
A polynomial time algorithm to factor
large numbers on a quantum
computer.
RSA ( public-key encryption)
2014/3/17
27
3
1.Introduction
๏ USA, China, Europe
๏ Google, D-Wave
2014/3/17
27
4
What is quantum computing
๏ Quantum mechanics + computing
2014/3/17
27
5
What is quantum computing
๏ QM=probability + minus signs
(by Scott Aaronson) Amplitude (complex number)
๐๐๐ ๐ข๐๐๐๐๐๐ก
ฮฑ
ฮฑ2
Probability: ๐1 + ๐2 > 0
Amplitude: ๐ผ1 + ๐ผ2 = 0
2014/3/17
27
6
What is quantum computing
If quantum computers are so great, how
come they have not been built yet?
๏ They have โ and they have proved that 15 = 3 × 5
(with high probability) .
๏ Perhaps 21 = 3 × 7 will be the next technological
milestone.
2014/3/17
27
7
Quantum finite automata
Quantum computer
Quantum finite
automata
2014/3/17
27
8
Semi-quantum finite automata
Read only tape head
Classical
Control
C
Unit
Quantum
Memory
C
0 1
1 0
1
0
=
0
1
1/2 1/3 1
1/2
=
1/2 2/3 0
1/2
cos ๐
sin ๐ 1
cos ๐
=
โ sin ๐ cos ๐ 0
โ sin ๐
Deterministic transition function: special stochastic matrix
Probabilistic transition function: stochastic matrix
Quantum transition function: unitary matrix
๏ Two-way finite automata with quantum and classical states
2QCFA (Ambainis, 02โTCS)
๏ 1QCFA ( Zheng, 2011)
2014/3/17
27
9
2.State succinctness for promise problem
Promise problem : A = ๐ด๐ฆ๐๐ , ๐ด๐๐ , A๐ฆ๐๐ โฉ ๐ด๐๐ = โ
Language: L = (๐ด๐ฆ๐๐ , ๐ด๐๐ ), where ๐ด๐ฆ๐๐ โช ๐ด๐๐ = ฮฃ โ
๐ฟ
Promise problem: ฮฃ = 0,1, # ,
๐ด๐ธ๐ (๐) = (๐ด๐ฆ๐๐ ๐ , ๐ด๐๐ (๐)), where
๐ด๐๐
๐ด๐ฆ๐๐
๐ด๐ฆ๐๐ = {๐ฅ#๐ฆ|๐ฅ = ๐ฆ, ๐ฅ, ๐ฆ โ {0,1}๐ }
๐ด๐๐ = {๐ฅ#๐ฆ|๐ฅ โ ๐ฆ, ๐ฅ, ๐ฆ โ {0,1}๐ , ๐ป ๐ฅ, ๐ฆ = ๐/2}
๐ป ๐ฅ, ๐ฆ is the Hamming distance between x and y.
๏ Deutsch-Jozsa promise problem in query complexity
๏ In communication complexity (H. Buhrman et al, STOCโ98)
Exact quantum communication complexity: ๐(log ๐)
Exact classical: ฮฉ(๐)
๏ Our result State complexity: ๐(๐) vs 2ฮฉ(๐)
2014/3/17
27
10
State succinctness for promise problem
๏Idea:
ฮฑ1 + ฮฑ2 + โฏ
2
=0
โฆ
2014/3/17
27
11
State succinctness for promise problem
An exact 1QCFA algorithm: ๐ฅ = ๐ฅ1 โฏ ๐ฅ๐ , ๐ฆ = ๐ฆ1 โฏ ๐ฆ๐
๐ฅ1 โฏ ๐ฅ๐ # ๐ฆ1 โฏ ๐ฆ๐
1
0
โฎ
0
๐๐
1
๐๐ ๐
2014/3/17
1
1 1
๐ โฎ
1
๐1,๐ฅ1
(โ1)
1
1
โฎ
๐
1
๐
๐ฅ๐ +๐ฆ๐
(โ1)
๐=1
โ
โฎ
โ
๐ฅ1
๐ฅ
โฏ
(โ1)๐ฅ1
1 (โ1)๐ฅ2
โฎ
๐
(โ1)๐ฅ๐
๐ฆ
โฏ
(โ1)๐ฅ1 +๐ฆ1
1 (โ1)๐ฅ2 +๐ฆ2
โฎ
๐
(โ1)๐ฅ๐ +๐ฆ๐
If w โ ๐ด๐ฆ๐๐ , ๐กโ๐๐ ๐ฅ = ๐ฆ,
2
๐
1
๐ฅ
+๐ฆ
โ1 ๐ ๐ = 1
๐
๐=1
If ๐ค โ ๐ด๐๐ , then H x, y = n/2
2
๐
1
โ1 ๐ฅ๐ +๐ฆ๐ = 0
๐
๐=1
27
12
State succinctness for promise problem
Theorem 1. The promise problem ๐ด๐ธ๐ can be solved
exactly by a 1QCFA with ๐ quantum basis states and ๐ ๐
classical states, whereas the sizes of the corresponding
1DFA are 2ฮฉ(๐) .
๏ The deterministic communication complexity for ๐ด๐ธ๐
is at least 0.007n (H. Buhrman et al, STOCโ98)
Therefore, the sizes of the corresponding 1DFA are
2ฮฉ(๐) (Kushilevitz, Advances in Computers, โ97).
2014/3/17
27
13
3.State succinctness for ๐ฟ ๐
Language ๐ฟ ๐ = ๐๐๐ ๐ โ โค+ }.
๏ 1DFA and 1NFA: ๐ states
๏ One-way quantum finite automata (1QFA): ๐๐๐๐ฆ
1
๐
log ๐
(Ambainis and Freivalds, FOCSโ98)
log ๐
๏1QFA: O( 3 ) (Bertoni et al., TCS โ06)
๐
log 2๐
๏1QFA: O(4
) (Ambainis,Nahimovs, TCS โ09)
๐
๐ผ
๐ผ
๐ผ
๏2DFA,2NFA:๐1 1 +๐2 2 +โฏ + ๐๐ ๐ ,where p =
๐ผ
๐ผ
๐ผ
๐1 1 ๐2 2 โฏ ๐๐ ๐
(prime factorization) (Mereghetti, Pighizzini, J.Autom. Lan. Com โ00)
๏One-way quantum finite automaton with restart (1QFAR):
1
๐
7 states, running time ๐( ๐ ๐๐โ2 ( )|๐ค|) (Yakaryilmaz, Cem Say,
DMTCS โ10)
2014/3/17
๐
๐
27
14
State succinctness for ๐ฟ ๐
2QCFA, two quantum states, several classical states,
Idea: wheel of fortune, ๐ squares,
|๐0 =
1
0
,|๐1 = , ๐๐ =
0
1
๐
๐
๐
sin
๐
cos
๐
๐
๐
๐
๐
cos
๐
โ sin
๏If ๐ค โ ๐ฟ ๐ , |๐ =|๐0
๏If ๐ค โ ๐ฟ ๐ , ๐ค = ๐
|๐ = (๐๐ ) ๐ |๐0 = cos
๐๐
|๐0
๐
+sin
๐๐
|๐1
๐
Probability to observe |๐1 is ๐๐ = ๐ ๐๐2
๐๐
๐
>
4
๐2
๏ Repeat for ๐(๐2 ) times ?
2014/3/17
27
15
State succinctness for ๐ฟ ๐
๏ Amplification of probabilities in a finite automaton
If ๐ค โ ๐ฟ: ๐๐ = 1 and ๐๐ = 0
If ๐ค โ ๐ฟ: ๐๐ = 1 โ ๐๐ and ๐๐ > ๐ก(๐)
If ๐ค โ ๐ฟ: ๐๐ = ๐ × ๐ก(๐) and ๐๐ = 0
If ๐ค โ ๐ฟ: ๐๐ = ๐ × ๐ก(๐) and ๐๐ > ๐ก(๐)
If ๐ค โ ๐ฟ: ๐๐๐๐ = 1
If ๐ค โ ๐ฟ: ๐๐๐๐ > 1 โ ๐
1
๐
Expected repetitions: ×
2014/3/17
๐ค
1
๐ก ๐
27
16
State succinctness for ๐ฟ ๐
Idea: 2QCFA, two quantum states, several classical states,
wheel of fortune, ๐ squares,
One iteration:
|๐0 =
1
0
,|๐1 = , ๐๐ =
0
1
๐
๐
๐
sin
๐
|๐ = (๐๐ ) ๐ |๐0 = cos
๐
๐
๐
cos
๐
cos
๏If ๐ค โ ๐ฟ ๐ , |๐ =|๐0
๏If ๐ค โ ๐ฟ ๐ , ๐ค = ๐
โ sin
๐๐ =
๐๐
|๐0
๐
๐
๐
4๐
๐2
๐๐
|๐1
๐
๐๐
๐ ๐๐2
๐
+sin
Probability to observe |๐1 is ๐๐ =
>
4
๐2
Use ๐๐,๐ to control the expected number of repetitions.
2014/3/17
27
17
State succinctness for ๐ฟ ๐
4๐
๐2
If ๐ค โ ๐ฟ ๐ , |๐ = |๐0 , ๐กโ๐๐ ๐๐ =
and ๐๐ = 0
๐ค
Repeated ad infinitum, the accepting probability
(1 โ ๐๐ )๐ 1 โ ๐๐ ๐ ๐๐ = 1
๏If ๐ค โ ๐ฟ ๐ , ๐๐ โฅ
๐โฅ0
4
and ๐๐
๐2
=
4๐
๐2
Repeated ad infinitum, the rejecting probability
(1 โ ๐๐ )๐ 1 โ ๐๐ ๐ ๐๐ > 1 โ ๐
๐โฅ0
1
๐
๏ Expected running time ๐( ๐2 |๐ค|)
2014/3/17
27
18
State succinctness for ๐ฟ ๐
๏ Using a lemma in (Dwork and Stockmeyer, SIAM J. Comput.
โ90), we can prove the following theorem.
For any integer ๐, any polynomial expected running
time 2PFA recognizing ๐ฟ ๐ with error probability
๐ < 1/2 has at least 3 (log ๐)/๐ states, where b is a
constant.
2014/3/17
27
19
4.State succinctness for C m
Language ๐ถ ๐ = {๐ค|๐ค โ ฮฃ ๐ }.
๏ DFA: ๐
๏ 1PFA: ๐(log 2 ๐) (Freivalds, Automatic Control and Computer
Sciences, โ82)
๏ 1QFA: ๐(log ๐) (Ambainis and Freivalds, FOCSโ98)
๏ 1QFAR: 7 quantum states, expected running time:
1 ๐
๐ 2 |๐ค| , exponential of ๐
๐
(Yakaryilmaz and Cem Say, DMTCSโ10)
Our result: 2QCFA, 2 quantum states, several classical
1
states, expected running time: ๐ ๐2 ๐ค 4 , polynomial of
๐
๐ and |๐ค|.
2014/3/17
27
20
State succinctness for C m
Idea: |๐0 =
1
0
,|๐1 = ,
0
1
cos ๐ 2๐ sin ๐ 2 ๐
โ sin ๐ 2 ๐ cos ๐ 2๐
(rotated by the angle โ ๐ 2 ๐)
๐¢ =
๐๐ = cos 2๐ โsin 2 ๐
sin 2 ๐ cos 2๐
(rotated by the angle 2๐)
๏ ๐ = |๐ค|,
2014/3/17
๐โ๐
2
27
21
State succinctness for C m
If ๐ค โ ๐ถ(๐), ๐ = ๐๐ผ
If w โ ๐ถ ๐ , ๐ค = ๐,
๐ = ๐๐ผ
๏ง
๐
๐
๐¢ ๐0 โช = ๐0 .
cos(๐ โ ๐) 2๐
๐¢ ๐0 โช=
โ sin(๐ โ ๐) 2 ๐
Pr =
๐
2๐2 ๐ + 1
2
sin(๐ โ ๐) 2 ๐
|๐0 โช
cos(๐ โ ๐) 2๐
๐ค
Probability to get |๐1 is
sin2 (๐
๐๐ =
1
โ ๐) 2 ๐ >
2 ๐โ๐
2
+1
(see Ambainis and Watrous, TCS โ02; Zheng et al. TCS โ13)
Use ๐๐,๐ and two random walks to control the expected number of
repetitions.
2014/3/17
27
22
State succinctness for C m
If ๐ค โ ๐ถ ๐ , |๐ = |๐0 , ๐กโ๐๐ ๐๐ =
๐
2๐2 ๐+1 2
and ๐๐ = 0
Repeated for ad infinitum, the accepting probability
(1 โ ๐๐ )๐ 1 โ ๐๐ ๐ ๐๐ = 1
๐โฅ0
๏If ๐ค โ ๐ถ ๐ , ๐๐ >
1
2 ๐โ๐ 2 +1
and ๐๐ <
๐
2๐2 ๐+1 2
Repeated for ad infinitum, the rejecting probability
(1 โ ๐๐ )๐ 1 โ ๐๐ ๐ ๐๐ > 1 โ ๐
๐โฅ0
๏ Expected running time ๐
2014/3/17
1 2
๐
๐
27
๐ค
4
.
23
State succinctness for C m
๏ Using a lemma in (Dwork and Stockmeyer, SIAMJ. Comput.
โ90), we can prove the following theorem.
For any integer ๐, any polynomial expected running
time 2PFA recognizing C ๐ with error probability
๐ < 1/2 has at least 3 (log ๐)/๐ states, where b is a
constant.
2014/3/17
27
24
5.A trade-off property of 1QCFA
๏ Quantum resources are expensive and hard to deal
with. One can expect to have only very limited
number of qubits in current quantum system.
๏ ๐ = ๐1 × ๐2 and
gcd ๐1 , ๐2 = 1.
๐ฟ ๐ = ๐ฟ ๐1 โฉ ๐ฟ(๐2 ) ,
๏ ๐ฟ ๐1 1QCFA ฮ(๐1 )
simulated by
๏ ๐ฟ ๐2 1DFA
1QCFA ฮ(๐2 )
๏ 1QCFA is closed under intersection
State complexity: ๐ log ๐1 + ๐(๐2 )
2014/3/17
27
25
6. Concluding remarks
๏ We have proved exact 1QCFA state succinctness for a
promise problem.
๏ We have proved 2QCFA state succinctness for two
languages
๏ A trade-off property of 1QCFA
๏ Future work
It has been proved (Klauck, STOC โ00) that exact 1QCFA have
not state complexity advantage over 1DFA in recognizing a
language. How about exact 2QCFA vs. 2DFA?
2014/3/17
27
26
2014/3/17
27
27
© Copyright 2026 Paperzz