Exact quantum algorithms Andris Ambainis University of Latvia Types of quantum algorithms Bounded-error: correct answer with probability at least 2/3. Exact: correct answer with certainty (probability 1). Grover's search 0 1 0 ... 0 x1 x2 x3 xN Is there i:xi=1? Classically, N queries required. Quantum: O(N) queries [Grover, 96]. Quantum, exact: N queries. Model Query model f(x1, ..., xN), xi{0,1}. xi given by a black box: Function i Complexity = number of queries xi Queries in the quantum world states: |1,1, |1, 2, …, |N, M. Query: Basis |i, j |i, j, if xi=0; |i, j -|i, j, if xi=1; Example x1 x2 x3 0 1 0 1,1|1, 1+1,2|1, 2+2,1|2, 1+3,1|3,1 Query 1,1|1, 1+1,2|1, 2- 2,1|2, 1+3,1|3,1 Quantum query model U0 Fixed Q U1 … Q UT starting state. U0, U1, …, UT – independent of x1, …, xN. Q – queries. Measuring final state gives the result. Known exact algorithms Deutsch’s problem 0 1 x1 x2 Determine x1x2, with query access to xi. [Cleve et al., 1998]: 1 quantum query, always the correct answer. Dutsch-Jozsa 0 1 0 ... 0 x1 x2 x3 xN Distinguish whether: x1 = x2 = ... = xN or xi=0 (xi=1) for exactly ½ of i{1, 2, ..., N}. Deterministic: N/2+1 queries. Quantum: 1 query. Grover's search 0 1 0 ... 0 x1 x2 x3 xN Is there i:xi=1? Promise: there is 0 or 1 i: xi=1. Classically: N queries. Quantum, exact: O(N) queries. Exact algorithms for total functions? Deutsch’s problem 0 1 x1 x2 Determine x1x2, with query access to xi. [Cleve et al., 1998]: 1 quantum query, always the correct answer. x1x2...xN can be computed with N/2 queries Montanaro et al., 2011. EXACT24(x1, x2, x3, x4)=1 if there are exactly 2 i:xi=1. Classical: 4 queries. Quantum: 2 queries, exact. Is there a total function f(x1, ..., xN) for which QE(f) < D(f)/2? quantum exact deterministic Our results Superlinear separation Theorem There is f(x1, ..., xN) such that D(f)=N; QE(f)=O(N0.86...). What should f be? Polynomial degree lower bound – degree of f(x1, ..., xN) as a multilinear polynomial. [Nisan, Szegedy, 92, Beals et al., 98] deg(f) Basis function D(f)=3, deg(f)=2 Iterated NE NE NE NE x1 x2 x3 x4 x5 NE x6 x7 x8 x9 d levels D(f)=3d, deg(f)=2d Our result NE NE NE x1 x2 Theorem x3 x4 x5 NE x6 x7 x8 x9 For d levels, QE(f)=O(2.593...d). Step 1 Algorithm for NE(x1, x2, x3). Starting state: Result: Step 2 p-algorithm: |start |start if f=0; |start p|start + | with ||start, if f=1. p=0 exact quantum algorithm Step 3 p-algorithm: |start |start if f=0; |start p|start + | with ||start, if f=1. NE(x1, x2, x3) – 2 queries, p = -7/9 NE f p-algo, k queries f f f p’-algo, 2k queries Step 3: result NE NE NE x1 x2 x3 x4 x5 NE x6 d levels, 3d variables; Badd p! p-algorithm with 2 queries. x7 x8 x9 Step 4 Amplification f p-algo, k queries f 2k queries, smaller p Form of amplitude amplification [Brassard et al., 2000] Final algorithm 1 level, 3 variables, 2 queries Iterate 2 levels, 9 variables, 4 queries Iterate 3 levels, 27 variables, 8 queries Amplify 3 levels, 27 variables, 16 queries ... Final result 211 queries for each 8 levels. N=38 variables, 211 queries. N=38k variables, 211k queries. QE(f)=N0.86... Other exact quantum algorithms EXACT 0 1 0 ... 0 x1 x2 x3 xN Determine whether xi=1 for exactly k of N variables. Montanaro et al., 2011: Algorithm: 2 out of 4, 2 queries; Computer optimization: 3 out of 6, 3 queries; Conjecture: N/2 out of N, N/2 queries. A, Iraids, Smotrovs Exact algorithms for determining: if xi=1 for exactly N/2 i, N/2 queries; if xi=1 for exactly k i, max(k, N-k) queries; Provably optimal. Natural computational problems; Simple algorithms. Algorithm: summary 1 query 1 query ... ... Threshold functions 0 1 0 ... 0 x1 x2 x3 xN it true that xi=1 for k of N variables? Exact algorithm, max(k, N-k+1) queries. Easiest: k=N/2, N/2+1 queries. Hardest: k=0 or k=N, N queries. Is Summary A function that requires N queries classically, O(N0.86...) queries for exact quantum algorithms. First separation by more than a factor of 2. Several other exact quantum algorithms. Advantages for exact quantum algorithms are more common that I thought Open problems d-level NE function (with 3d variables): 1. 2. 3. 4. O(2.593...d) query exact algorithm; Lower bound: (2.11...d). Other iterated functions? Other symmetric functions? More exact algorithms? Open problems Lower bound methods for exact quantum algorithms? Currently known: 5. Bounded-error quantum lower bounds; QE(f) deg(f)/2; For NEd, both of them fail. More information A. Ambainis. Superlinear advantage for exact quantum algorithms, arxiv:1211.0721. A. Ambainis, J. Iraids, J. Smotrovs. Exact quantum query complexity of EXACT and THRESHOLD, arxiv:1302.1235.