Totally corrective boosting with cardinality penalization
Methods, systems, and apparatus, for totally corrective boosting with cardinality penalization are described. One of the methods includes obtaining initialization data identifying training examples, a dictionary of weak classifiers, and an active weak classifier matrix. Iterations of a totally corrective boosting with cardinality penalization process are performed, wherein each iteration performs operations comprising selecting a weak classifier from the dictionary of weak classifiers that most violates a constraint of a dual of the primal problem. The selected weak classifier is included in the active weak classifier matrix. The primal problem is optimized, and a discrete weight vector is determined. Weak classifiers are identified from the active weak classifier matrix with respective discrete weights greater than a threshold. The regularized risk is optimized, and a continuous weight vector is determined. The classifier is determined as an ensemble identified by the weak classifiers and the continuous weight vector.
1. A system comprising:
a programmable quantum annealing chip and one or more computers, the programmable quantum annealing chip comprising a plurality of qubits arranged in a plurality of unit cells, wherein the plurality of qubits are connected by programmable couplers, and one or more storage devices storing instructions that are operable, when executed by the programmable quantum annealing chip, to cause the programmable quantum annealing chip to perform operations for generating a classifier by optimizing a primal problem, wherein the operations comprise:
obtaining initialization data identifying a plurality of training examples, a dictionary of weak classifiers that includes a plurality of weak classifiers, and an active weak classifier matrix;
performing one or more iterations of a totally corrective boosting with cardinality penalization process until a termination condition is satisfied, wherein each iteration performs operations comprising:
selecting a weak classifier from the dictionary of weak classifiers that most violates a constraint of a dual of the primal problem,
including the selected weak classifier in the active weak classifier matrix,
optimizing the primal problem using the training examples and the active weak classifier matrix including determining a discrete weight vector, wherein each discrete weight in the discrete weight vector identifies a respective weight of a weak classifier in the active weak classifier matrix, wherein a plurality of discrete weights are mapped to the plurality of qubits, and wherein determining the discrete weight vector comprises performing quantum annealing to determine a discrete weight vector bit-depth;
in response to optimizing the primal problem, identifying one or more weak classifiers from the active weak classifier matrix with respective discrete weights greater than a threshold, and
optimizing l-regularized risk using the identified one or more weak classifiers and the training examples including determining a continuous weight vector, wherein each continuous weight in the continuous weight vector identifies a respective weight of a weak classifier in the one or more identified weak classifiers; and
determining the classifier as an ensemble identified by the weak classifiers in the active weak classifier matrix with respective continuous weights greater than a threshold and the continuous weight vector.
2. The system of claim 1 , wherein quantum annealing is performed on an Ising model or a restricted Ising model.
3. The system of claim 1 , wherein the plurality of qubits comprise superconducting loops.
4. The system of claim 1 , wherein each unit cell comprises 8 qubits.
5. The system of claim 1 , wherein the programmable couplers comprise inductive couplers.
6. The system of claim 1 , wherein optimizing the primal problem includes using, wherein each of the discrete weights is a non negative fixed point n-bit value, and each bit of each of the n-bit values is mapped to a corresponding quantum bit of the programmable quantum annealing chip.
7. A quantum-computer implemented method comprising:
obtaining, by a programmable quantum annealing chip, initialization data identifying a plurality of training examples, a dictionary of weak classifiers that includes a plurality of weak classifiers, and an active weak classifier matrix, wherein the programmable quantum annealing chip comprises a plurality of qubits arranged in a plurality of unit cells, and wherein the plurality of qubits are connected by programmable couplers;
performing, by the programmable quantum annealing chip, one or more iterations of a totally corrective boosting with cardinality penalization process until a termination condition is satisfied, wherein each iteration performs operations comprising:
selecting a weak classifier from the dictionary of weak classifiers that most violates a constraint of a dual of a primal problem,
including the selected weak classifier in the active weak classifier matrix,
optimizing the primal problem using the training examples and the active weak classifier matrix including determining a discrete weight vector comprising a plurality of discrete weights, wherein each discrete weight in the discrete weight vector identifies a respective weight of a weak classifier in the active weak classifier matrix, wherein the plurality of discrete weights are mapped to the plurality of qubits, and wherein determining the discrete weight vector comprises performing, by the programmable quantum annealing chip, quantum annealing to determine a discrete weight vector bit-depth,
in response to optimizing the primal problem, identifying one or more weak classifiers from the active weak classifier matrix with respective discrete weights greater than a threshold, and
optimizing l-regularized risk using the identified one or more weak classifiers and the training examples including determining a continuous weight vector, wherein each continuous weight in the continuous weight vector identifies a respective weight of a weak classifier in the one or more identified weak classifiers; and
determining the classifier as an ensemble identified by the weak classifiers in the active weak classifier matrix with respective continuous weights greater than a threshold and the continuous weight vector.
8. The quantum-computer implemented method of claim 7 , wherein quantum annealing is performed on an Ising model or a restricted Ising model.
9. The quantum-computer implemented method of claim 7 , wherein the plurality of qubits comprise superconducting loops.
10. The quantum-computer implemented method of claim 7 , wherein each unit cell comprises 8 qubits.
11. The quantum-computer implemented method of claim 7 , wherein the programmable couplers comprise inductive couplers.
12. The quantum-computer implemented method of claim 7 , wherein optimizing the primal problem includes using, wherein each of the discrete weights is a non negative fixed point n-bit value, and each bit of each of the n-bit values is mapped to a corresponding quantum bit of the programmable quantum annealing chip.