Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Grover. Part 2 Components of Grover Loop • The Oracle -- O • The Hadamard Transforms -- H • The Zero State Phase Shift -- Z O is an H is Oracle Hadamards Grover Iterate Z is Zero State Phase Shift H is Hadamards Inputs oracle This is action of quantum oracle We need to initialize in a superposed state This is a typical way how oracle operates Encodes input combination with changed sign in a superposition of all This is a typical way how oracle operation is described Role of Oracle • We want to encode input combination with changed sign in a superposition of all states. • This is done by Oracle together with Hadamards. • We need a circuit to distinguish somehow globally good and bad states. Vector of Hadamards Notation Reminder a a Control with value a=1 Control with value a=0 a equivalent All information of oracle is in the phase but how to read it? This is just an example of a single minterm, but can be any function Zero State Phase Shift Circuit Flips the data phase This is value of oracle bit Flips the oracle bit when all bits are zero This is state of all zeros Rewriting matrix Z to Dirac notation, you can change phase globally 2 0 0 0 0 0 0 0 0 0 0 0 With accuracy to phase 0 0 0 0 -1 0 + 0 0 0 -1 0 0 0 0 -1 0 0 0 0 -1 = 1 0 0 0 0 -1 0 0 0 0 -1 0 0 0 0 -1 Here you have all components of Grover’s loop This is a global view of Grover. Repeatitions of G In each G Generality • Observe that a problem is described only by Oracle. • So by changing the Oracle you can have your own quantum algorithm. • You can still improve the Grover loop for particular special cases Here we explain in detail what happens inside G. This can be generalized to Glike circuits proof Grover iterate has two tasks: (1) invert the solution states and (2) invert all states about the mean Will be explained in next slide Explanation of the first part of Grover iterate formula Here we prove that |> < | used inside HZH calculates the mean a Vector of mean values From previous slide ( ) ( This proof is easy and it only uses formalisms that we already know. ) What does it mean invert all states about the mean? Positive or negative amplitudes in other explanations Amplitudes of bits after Hadamard For every bit All possible states Amplitudes of bits after one stage of G This value based on previous slide This slides explains the basic mechanism of the Grover-like algorithms Additional Exercise This is a lot calculations, requires matrix multiplication You can verify it also in simulation Here we calculate analytically when to stop For marked state For unmarked state The equations taken from the previous slides “Grover Iterate” recursion We want to find how many times to iterate We found k from these equations But you can do better if you have knowledge, for instance the upper bound of chromatic number in graph coloring Grover search example. • Here is an example of Grover search for n = 3 qubits, where N = 2n =8. – We omit reference to qubit n+1, which is in state 1 /√2 (|0>−|1>i) and does not change. • The dimension of the unitary operators for this example is thus 2n = 8 also.) oracle • (Remember that numbering starts with 0 and ends with 7, so that the -1 here is in the slot for |5>.) • This matrix reverses the sign on state |5>, and leaves the other states unchanged. •Suppose the unknown number is |a> = |5>. •The matrix or black box oracle Ufa is •The Walsh matrix W8 is Now we use normalization The matrix −Uf0 is This matrix changes the sign on all states except |0>. Finally, we have the repeated Grover algorithm: hadamards oracle shift step RsRa in the After second rotation we get Summary and our work When you know anything about the problem (symmetry, observation, bounds, function within some classification class) you can design a better Grover like algorithm but for your data only. This is enough in real life like CAD or Image Processing, since data are always specific, not the worst case data as in Mathematic proofs Problem for students • Build the Grover algorithm for ternary quantum logic. • First you need to generalize Hadamard transform to Chrestenson transform. • Next you need to have some kind of ternary reversible gates to build oracle. • The same gates will be used for Zero State Phase Shift circuit.