Skip to content

Instantly share code, notes, and snippets.

Show Gist options
  • Save cgranade/d20fcf3e9907eb26c7fc2f6a8d855918 to your computer and use it in GitHub Desktop.
Save cgranade/d20fcf3e9907eb26c7fc2f6a8d855918 to your computer and use it in GitHub Desktop.
Display the source blob
Display the rendered blob
Raw
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Proposed Semantics for Qobj.\\_\\_call\\_\\_ #"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"This Notebook describes proposed semantics for representing the actions of superoperators and operators on operators and states as *functions*. In particular, this will greatly simplify the syntax used to represent evolving a density operator under a quantum channel stored as a ``type=\"super\"`` Qobj."
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"import qutip as qt"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"A operator $X$ acting on a Hilbert space $\\mathcal{H}$ is a representation of a function $f_X(\\left|\\psi\\right\\rangle) = X\\left|\\psi\\right\\rangle$ such that $f_X(\\left|\\psi\\right\\rangle + \\left|\\phi\\right\\rangle) = f_X(\\left|\\psi\\right\\rangle) + f_X(\\left|\\phi\\right\\rangle)$ and such that for any scalar $\\alpha$, $f_X(\\alpha \\left|\\psi\\right\\rangle) = \\alpha f_X(\\left|\\psi\\right\\rangle)$. Similarly, given a function $g_Y(\\rho)$ such that $g_Y(\\rho + \\sigma) = g_Y(\\rho) + g_Y(\\sigma)$ and $g_Y(\\alpha \\rho) = \\alpha g_Y(\\rho)$, we can represent the action of $g_Y$ as a superoperator $Y$.\n",
"\n",
"In QuTiP, superoperators $Y$ are represented by ``type=\"super\"`` Qobj instances with ``superrep in ['super', 'choi', 'chi']``. For example, a map $g(\\rho) = (\\sigma_x \\rho \\sigma_x + \\rho) / 2$ representing dephasing in the $x$-direction corresponds to the following Qobj:"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/latex": [
"Quantum object: dims = [[[2], [2]], [[2], [2]]], shape = [4, 4], type = super, isherm = True\\begin{equation*}\\left(\\begin{array}{*{11}c}0.500 & 0.0 & 0.0 & 0.500\\\\0.0 & 0.500 & 0.500 & 0.0\\\\0.0 & 0.500 & 0.500 & 0.0\\\\0.500 & 0.0 & 0.0 & 0.500\\\\\\end{array}\\right)\\end{equation*}"
],
"text/plain": [
"Quantum object: dims = [[[2], [2]], [[2], [2]]], shape = [4, 4], type = super, isherm = True\n",
"Qobj data =\n",
"[[ 0.5 0. 0. 0.5]\n",
" [ 0. 0.5 0.5 0. ]\n",
" [ 0. 0.5 0.5 0. ]\n",
" [ 0.5 0. 0. 0.5]]"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"dephasing_superop = (qt.to_super(qt.sigmax()) + qt.to_super(qt.qeye(2))) / 2\n",
"dephasing_superop"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"This Qobj then acts on vectorized operators, $|\\rho\\rangle\\!\\rangle$, as given by the ``operator_to_vector`` function. Thus, to apply this superoperator to a density operator requires calling ``operator_to_vector`` and the corresponding ``vector_to_operator``:"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/latex": [
"Quantum object: dims = [[2], [2]], shape = [2, 2], type = oper, isherm = True\\begin{equation*}\\left(\\begin{array}{*{11}c}0.500 & 0.187\\\\0.187 & 0.500\\\\\\end{array}\\right)\\end{equation*}"
],
"text/plain": [
"Quantum object: dims = [[2], [2]], shape = [2, 2], type = oper, isherm = True\n",
"Qobj data =\n",
"[[ 0.5 0.18655508]\n",
" [ 0.18655508 0.5 ]]"
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"rho = qt.rand_dm_ginibre(2)\n",
"rho_out = qt.vector_to_operator(dephasing_superop * qt.operator_to_vector(rho))\n",
"rho_out"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"This is confusing to newcomers to the library, and while it accurately models the Liouvillian representation (that is, ``superrep=\"super\"``), does not model the more general view of a channel as an abstract object. For instance, $\\mathcal{U}(\\rho) = U \\rho U^\\dagger$ (as returned by ``qt.to_super(U)``) is already a quantum channel representing conjugation by a unitary $U$, and need not be written as a superoperator. Using ``__call__``, the group action of superoperators and operators on operators can be modeled in a uniform manner."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Here, we prototype by introducing a new method and monkey-patching it into ``Qobj``."
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"def qobj_call(self, other):\n",
" if not isinstance(other, qt.Qobj):\n",
" raise TypeError(\"Only defined for quantum objects.\")\n",
" \n",
" if self.type == \"super\":\n",
" if other.type == \"ket\":\n",
" other = qt.ket2dm(other)\n",
" elif other.type == \"oper\":\n",
" return qt.vector_to_operator(self * qt.operator_to_vector(other))\n",
" else:\n",
" raise TypeError(\"Can only act super on oper or ket.\")\n",
" \n",
" elif self.type == \"oper\":\n",
" if other.type == \"ket\":\n",
" return self * other\n",
" elif other.type == \"oper\":\n",
" return self * other * self.dag()\n",
" else:\n",
" raise TypeError(\"Can only act oper on oper or ket.\")\n",
"\n",
"qt.Qobj.__call__ = qobj_call"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"psi = qt.rand_ket_haar(2)\n",
"rho = qt.rand_dm_ginibre(2)\n",
"X = qt.sigmax()\n",
"I = qt.qeye(2)\n",
"S = (qt.to_super(I) + qt.to_super(X)) / 2"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We finally assert a few usecases of these new semantics."
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"assert X(psi) == X * psi\n",
"assert X(rho) == X * rho * X.dag()\n",
"assert S(rho) == qt.vector_to_operator(S * qt.operator_to_vector(rho))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 2",
"language": "python",
"name": "python2"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 2
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython2",
"version": "2.7.11"
}
},
"nbformat": 4,
"nbformat_minor": 0
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment