# Difference between revisions of "SC²S Colloquium - February 3, 2017"

(Created page with "{| class="wikitable" |- | '''Date:''' || February 3, 2017 |- | '''Room:''' || 02.07.023 |- | '''Time:''' || 3:00 pm, s.t. |- |} Category:ShowComingUp Category:news") |
|||

Line 8: | Line 8: | ||

|- | |- | ||

|} | |} | ||

+ | |||

+ | == Felix Thimm: Implementation and Evaluation of a tensornetwork-based Machine Learning Algorithm == | ||

+ | It has been known for some time, that there is a close relation between tensor networks and machine learning. Several machine learning algorithms using tensor networks have recently been proposed in different papers. This thesis focuses on evaluating one particular approach, which was introduced in the paper "Supervised Learning with Quantum-Inspired Tensor Networks". The proposed algorithm uses an MPS tensor network in order to store the data, which is necessary to learn the features of a given data set. The learning performance of the algorithm is evaluated for different data sets, in order to determine the properties and limits of the algorithm. Furthermore, possible extensions are presented to extend the algorithms performance for particular types of data sets. | ||

## Latest revision as of 11:05, 12 January 2017

Date: |
February 3, 2017 |

Room: |
02.07.023 |

Time: |
3:00 pm, s.t. |

## Felix Thimm: Implementation and Evaluation of a tensornetwork-based Machine Learning Algorithm

It has been known for some time, that there is a close relation between tensor networks and machine learning. Several machine learning algorithms using tensor networks have recently been proposed in different papers. This thesis focuses on evaluating one particular approach, which was introduced in the paper "Supervised Learning with Quantum-Inspired Tensor Networks". The proposed algorithm uses an MPS tensor network in order to store the data, which is necessary to learn the features of a given data set. The learning performance of the algorithm is evaluated for different data sets, in order to determine the properties and limits of the algorithm. Furthermore, possible extensions are presented to extend the algorithms performance for particular types of data sets.