Quantum algorithms vs. polynomials and the maximum quantum-classical gap in the query model
Query model Function f(x 1,..., x N ), x i {0,1}. x i given by a black box: i xixi Complexity = number of queries
Quantum query model Fixed starting state. U 0, U 1, …, U T – independent of x 1, …, x N. Q – queries: U0U0 QQ U1U1 UTUT …
Reasons to study query model Encompasses many quantum algorithms (Grover’s search, quantum part of factoring, etc.). Provable quantum-vs-classical gaps.
Quantum vs. classical 1 query quantumly How many queries classically?
Period finding x 1, x 2,..., x N - periodic i xixi Find period r 1 query quantumly
Period-finding Quantum algorithm works if N r 2. T classical queries – can test T 2 possible periods. i xixi queries classically
Our result [Aaronson, A] Task that requires 1 query quantumly, ( N) classically. 1 query quantum algorithms can be simulated by O( N) query probabilistic algorithms.
Fourier checking/Forrelation
Forrelation Input: (x 1,..., x N, y 1,..., y N ) {-1, 1} 2N. Are vectors highly correlated? F N – Fourier transform over Z N. F N – Fourier transform over Z N.
More precisely... Is the inner product 3/5 or 1/100?
Quantum algorithm 1. Generate states in parallel (1 query). 2. Apply F N to 2 nd state. 3. Test if states equal (SWAP test).
Classical lower bound Theorem Any classical algorithm for FORRELATION uses queries.
REAL FORRELATION Distinguish between random (x i ’s - Gaussian); random,. Real-valued vectors
Lower bound Claim REAL FORRELATION requires queries. Intuition: if, each variable – Gaussian, correlations between x i ’s and y j ’s - weak. o( N) values x i and y j uncorrelated random variables.
Reduction Proof idea: Replace x i sgn(x i ) to achieve x i {-1, 1}. T query algorithm for FORRELATION T query algorithm for REAL FORRELATION
Simulating 1 query quantum algorithms
Simulation Theorem Any 1 query quantum algorithm can be simulated probabilistically using O( N) queries.
Analyzing query algorithms QQ QUTUT … U1U1 1,1 |1,1 + 1,2 |1, 2 + … + N, M |N, M 1,1 is actually 1,1 (x 1,..., x N )
Polynomials method Lemma [Beals et al., 1998] After k queries, the amplitudes are polynomials in x 1,..., x N of degree k. Measurement: Polynomial of degree 2k
Our task Pr[A outputs 1] = p(x 1,..., x N ), deg p =2. 0 p(x 1,..., x N ) 1. Task: estimate p(x 1,..., x N ) with precision . Solution: random sampling.
Pre-processing Problem: large error if sampling omits x i with large influence in p(x 1,..., x N ). Solution: replace influential x i ’s by several variables with smaller influence.
Sampling 1 ample N of N 2 terms independently. Good if we sample N of N 2 terms independently. Estimator: Requires sampling N variables x i !
Sampling 2 Sampling N terms a i,j x i x j Sampling N variables x i
Extension to k queries Theorem k query quantum algorithms can be simulated probabilistically with O(N 1-1/2k ) queries. Proof: Algorithm polynomial of degree 2k; Random sampling. Question: Is this optimal?
K-fold forrelation
Forrelation: given black box functions f(x) and g(y), estimate K-fold forrelation: given f 1 (x),..., f k (x), estimate
Results Theorem k-fold forrelation can be solved with k/2 quantum queries. Conjecture k-fold forrelation requires (N 1-1/k ) queries classically.
From polynomials to quantum algorithms (with Scott Aaronson, Jānis Iraids, Mārtiņš Kokainis, Juris Smotrovs)
Quantum algorithm with t queries Polynomials of degree 2t ??
Quantum algorithm with 1 query Polynomials of degree 2 Our result
More precisely... Polynomial p represents f with error if: f = 0 p [0, ]; f = 1 p [1- , 1]; f – undefined p [0, 1]. Theorem Q (f)=1 for some <1/2 iff f can be represented by p: deg p=2 with error <1/2.
Standard polynomial representation Block-multilinear representation Step 1
Requirements q(x 1,..., x N, x 1,..., x N ) f(x 1,..., x N ); q(x 1,..., x N, y 1,..., y N ) [-1, 1] for all x i, y j {0, 1}.
Step 2: evaluating q U = (N a i,j ) – unitary. SWAP test on | x and U| y : Still works if ||U|| C!
Two norms Have: |q| 1 Need:
Step 3: variable splitting Replace x i by, - new variables. 1. |q| 1 preserved; 2. Influential variables - eliminated.
Result Variable-splitting K – Groethendieck’s constant
Summary 1 quantum query = ( N) classical queries. k quantum queries can be simulated with O(N 1-1/2k ) classical queries. 1 quantum query = polynomials of degree 2.
Open problem 1 Does k-fold FORRELATION require (N 1-1/2k ) queries classically? Plausible but looks quite difficult matematically.
Open problem 2 Best quantum-classical gaps: 1 quantum query - ( N) classical queries; 2 quantum queries - ( N) classical queries;... log N quantum queries - classical queries. Any problem that requires O(log N) queries quantumly, (N c ), c>1/2 classically?
Open problem 3 Characterize quantum algorithms with 2, 3,..., queries? 2 queries polynomials of degree 4? Polynomials of degree 3 2 query algorithms?