0:00:06 | our main result is an efficient algorithm the design efficient content circuits assume that condom |
---|
0:00:12 | anybody estonian |
---|
0:00:15 | this algorithm is optimal number of you bit as well that in a scaling up |
---|
0:00:19 | number of gates |
---|
0:00:21 | another advantage that we believe essential for content assimilation is that the user can specify |
---|
0:00:27 | the error of simulation |
---|
0:00:29 | we also introduce a method to reduce the selected taps this is essential for parallel |
---|
0:00:34 | quantum computation |
---|
0:00:37 | then we present numerical evidence is which indicate that and upper bounds for quite some |
---|
0:00:42 | simulation overestimate assimilation error by several orders of magnitude |
---|
0:00:49 | the goal and hamiltonian simulation our corpus is the and the action of dynamic shared |
---|
0:00:54 | by hamiltonian an initial state |
---|
0:00:56 | such simulations are important because the only known methods for efficiently simulating broad classes physically |
---|
0:01:03 | relevant in quantum systems |
---|
0:01:05 | the first step in these algorithms to find a mapping between states in the original |
---|
0:01:09 | hilbert space and the sub-spaces simulator so space this mapping allows one standard error in |
---|
0:01:15 | the simulator that is logically equivalent to the initial state stimuli |
---|
0:01:20 | simulated data for sequence one or more operations to what state transforming tuesday that is |
---|
0:01:25 | logically equivalent one is close to the involves a simulated system |
---|
0:01:30 | this is in state one of the one algorithm of the resultant state you know |
---|
0:01:35 | generally learned efficiently owen quantities such as expectation values local of their most can be |
---|
0:01:41 | found efficiently by measuring state you but |
---|
0:01:46 | with a simulated to be one here in vector form at a more gain i |
---|
0:01:51 | k and control not okay our objective is to construct a classical algorithm that you |
---|
0:01:56 | know simulation circuits that uses fewer these cases possible |
---|
0:02:02 | the in this algorithm are the biggest thing is in downtown the error tolerance for |
---|
0:02:07 | assimilation and evolution and then you know with the basis during representing the final circuit |
---|
0:02:13 | that's image the evolution under the input hamiltonian altogether |
---|
0:02:19 | the hamiltonian is probably to be available i mean and forty one time simulator using |
---|
0:02:26 | quantum computer engineering university of them |
---|
0:02:33 | this process is performed in three steps |
---|
0:02:35 | first the hamiltonian sorted into a sum mutual each meeting groups of operators to reduce |
---|
0:02:40 | the circuit |
---|
0:02:41 | next believe further suzuki algorithm is applied to decompose the evolution into a sequence of |
---|
0:02:46 | exponential ali operators |
---|
0:02:49 | the final step involves designing circuits implement each the exponentials |
---|
0:02:53 | circuit implementation of the middle switching to the eigen basis the exponential implementing evolution in |
---|
0:02:58 | the eigen faces and then transforming back to the computational bases resulting circuits are session |
---|
0:03:04 | because only a raiders communication you guys a lot |
---|
0:03:10 | the using this is sort operation such that those that you can be done it |
---|
0:03:16 | our algorithm on all you operations and group |
---|
0:03:21 | as example consider the following operation one or you |
---|
0:03:27 | so algorithm so the operation and one or what gender issue using c |
---|
0:03:37 | conclusions of our work provides an efficient algorithm for designing quantum simulation circuits from anybody |
---|
0:03:42 | have also it's we introduce grouping techniques of optimize the data the resulting simulations are |
---|
0:03:48 | it's |
---|
0:03:48 | we also provide numerical estimates of the typical errors that occur from using shorter suzuki |
---|
0:03:53 | formulas and show that existing estimates can be part two loops |
---|