A5046808924- Quantum Computing Algorithms and Architecture
- Quantum Information and Cryptography
- Neural Networks and Reservoir Computing
- Management and Optimization Techniques
- Optimization and Packing Problems
- Advancements in Semiconductor Devices and Circuit Design
- Advanced Memory and Neural Computing
- Computability, Logic, AI Algorithms
- Scientific Computing and Data Management
- Parallel Computing and Optimization Techniques
- Catalysis for Biomass Conversion
- Ferroelectric and Negative Capacitance Devices
- Machine Learning in Materials Science
- Quantum-Dot Cellular Automata
- Quantum and electron transport phenomena
- Advanced Neural Network Applications
- Optimization and Search Problems
Volkswagen Group (Germany)
2021-2024
Data:Lab Munich (Germany)
2021-2023
Leiden University
2021-2023
Volkswagen Group (United States)
2021
Ames Research Center
2020
Research Institute for Advanced Computer Science
2020
We introduce TensorFlow Quantum (TFQ), an open source library for the rapid prototyping of hybrid quantum-classical models classical or quantum data. This framework offers high-level abstractions design and training both discriminative generative under supports high-performance circuit simulators. provide overview software architecture building blocks through several examples review theory neural networks. illustrate TFQ functionalities via basic applications including supervised learning...
Abstract With the increased focus on quantum circuit learning for near-term applications devices, in conjunction with unique challenges presented by cost function landscapes of parametrized circuits, strategies effective training are becoming increasingly important. In order to ameliorate some these challenges, we investigate a layerwise strategy circuits. The depth is incrementally grown during optimization, and only subsets parameters updated each step. We show that when considering...
Quantum machine learning (QML) has been identified as one of the key fields that could reap advantages from near-term quantum devices, next to optimization and chemistry. Research in this area focused primarily on variational algorithms (VQAs), several proposals enhance supervised, unsupervised reinforcement (RL) with VQAs have put forward. Out three, RL is least studied it still an open question whether can be competitive state-of-the-art classical based neural networks (NNs) even simple...
Abstract Variational quantum algorithms are the leading candidate for advantage on near-term hardware. When training a parametrized circuit in this setting to solve specific problem, choice of ansatz is one most important factors that determines trainability and performance algorithm. In machine learning (QML), however, literature ansatzes motivated by data structure scarce. work, we introduce an tasks weighted graphs respects graph symmetry, namely equivariance under node permutations. We...
Abstract Variational quantum machine learning algorithms have become the focus of recent research on how to utilize near-term devices for tasks. They are considered suitable this as circuits that run can be tailored device, and a big part computation is delegated classical optimizer. It has also been hypothesized they may more robust hardware noise than conventional due their hybrid nature. However, effect training models under influence hardware-induced not yet extensively studied. In work,...
The binary paint shop problem (BPSP) is an APX-hard optimization of the automotive industry. In this work, we show how to use Quantum Approximate Optimization Algorithm (QAOA) find solutions BPSP and demonstrate that QAOA with constant depth able beat classical heuristics on average in infinite size limit $n\rightarrow\infty$. For BPSP, it known no algorithm can exist which approximates polynomial runtime. We introduce a instance hard solve QAOA, numerically investigate its performance...
Variational quantum algorithms are the leading candidate for advantage on near-term hardware. When training a parametrized circuit in this setting to solve specific problem, choice of ansatz is one most important factors that determines trainability and performance algorithm. In machine learning (QML), however, literature ansatzes motivated by data structure scarce. work, we introduce an tasks weighted graphs respects graph symmetry, namely equivariance under node permutations. We evaluate...
Variational quantum machine learning algorithms have become the focus of recent research on how to utilize near-term devices for tasks. They are considered suitable this as circuits that run can be tailored device, and a big part computation is delegated classical optimizer. It has also been hypothesized they may more robust hardware noise than conventional due their hybrid nature. However, effect training models under influence hardware-induced not yet extensively studied. In work, we...
Image recognition is one of the primary applications machine learning algorithms. Nevertheless, models used in modern image systems consist millions parameters that usually require significant computational time to be adjusted. Moreover, adjustment model hyperparameters leads additional overhead. Because this, new developments and hyperparameter optimization techniques are required. This paper presents a quantum-inspired technique hybrid quantum-classical for supervised learning. We...