Quantum^BW Colloquium

With the advance­ment of quantum technol­ogy, researchers aim to under­stand if and how quantum algorithms could have advan­tages compared to their classi­cal counter­parts, e.g., in the context of machine learn­ing. The inves­ti­ga­tion of possi­ble benefits of quantum compared to classi­cal machine learn­ing models requires thorough theoret­i­cal as well as empir­i­cal studies. In this talk, we will firstly present an overview over poten­tial advan­tages of quantum machine learn­ing (QML) over its classi­cal counter­part. Then, we will deep-dive into near-term quantum machine learn­ing algorithms that are based on short-depth, parame­ter­ized quantum circuits, which are well suited for execu­tion on near-term quantum hardware. These models are promis­ing candi­dates for a set of near-term empir­i­cal studies targeted to under­stand the applic­a­bil­ity of quantum machine learn­ing. However, as shown by a variety of research train­ing these models can become challeng­ing, especially at increas­ing scale.