As part of the DFG Priority Programme SPP2244 "2dMP", we are investigating atomically thin semiconductors that are stacked on top of each other at a specific twist angle. In the process, interlayer excitons can form - optically active, composite particles that are distributed between the layers of the heterostructure. Their electronic and optical properties can be specifically influenced by the choice of material combination. At the same time, the twist angle creates a superimposed potential landscape that enables analogies to the Hubbard physics of lattice models.

Together with our experimental partners at the University of Oldenburg and TU Berlin, we are investigating the tunable optical properties of moiré heterostructures in microresonators to study quantum phase transitions, exciton-polariton condensation and fermionisation of bosonic quasiparticles on a highly controllable semiconductor platform.

The BMBF's nationally funded Quantum Repeater.Link (QR.N) network aims to realise a quantum repeater prototype - a key technology for future quantum networks such as the quantum internet.

Our group acts as an interface between quantum information theory and the development of quantum repeater protocols on the one hand and the experimental semiconductor hardware platform on the other. We develop methods to simulate the implementation of quantum repeater protocols, both with conventionally entangled states and with multipartite entangled photon-based cluster states.

Quantum reservoir computing (QRC) is a novel computing concept at the interface of quantum computing and artificial neural networks. In contrast to gate-based quantum computing, the reservoir approach relaxes the strict requirements for complete control over the system. Instead, a loose network of quantum systems with a randomised coupling topology is trained to solve computational or classification tasks.

PhotonicQRC offers two major advantages due to the inherent quantum nature of the system: Firstly, quantum input can be processed directly without having to translate it into a classical form. Secondly, the exponential size of the Hilbert space enables the solution of complex tasks with extremely small physical networks.

PhotonicQRC is operated by a German-French consortium co-funded by the DFG and the ANR. The project links our group in Oldenburg with experimental partners at the TU Berlin, femto-ST (CNRS) and the ENS in Paris.

Quantum machine learning (QML) is becoming increasingly important, especially due to its applicability to NISQ technologies. In this project, we investigate the potential of QML algorithms that can be executed on available quantum computing hardware and compare their capabilities and performance with machine learning methods based on quantum artificial neural networks.

The project is funded by the Quantum Fellowship Programme by the DLR Quantum Computing Initiative and is being carried out in collaboration with Prof. Dr Meike List from the DLR Institute for Satellite Geodesy and Inertial Sensor Technology in Bremen.

Quantum Natural Language Processing (QNLP) leverages the properties of quantum systems to develop models for the processing of human natural languages.

Despite the rapid and powerful rise of intelligent natural language processing (NLP) applications, such as translators, text generators, etc, the methods used in these applications still face challenges in fully capturing the complexity of natural languages and reaching human-level capabilities, such as the diverse ways humans understand and use language. They still face the inability to understand complex linguistic contexts and perform difficult linguistic tasks (For text generation, these limitations are temporarily addressed by a technique known as prompt engineeringbut it is not highly efficient and is not considered a long-term solution). Another issue with intelligent text processing systems is time complexity, as these models are trained on massive text corpora, making the process highly time-consuming and computationally expensive.

Our project on QNLP aims at developing, implementing, and benchmarking quantum models to address the aforementioned issues. By harnessing quantum superposition and entanglement, these models could be more innovative and effective for intelligent applications in natural language processing, potentially outperforming classical methods in understanding context and reducing computational overhead. The project is funded by the DAAD (German Academic Exchange Service).