Quantum Computing / Quantum-enhanced AI

A half-day session introduces you to near-term quantum computing from the application perspective. The focus is not on fundamentals of quantum physics, but rather on the essential take-away messages for using quantum computing in business applications.

Required background: As we discuss the scope of the training beforehand with you, we can adapt to various backgrounds – may it be a management background, or a technical background like in engineering, computer science or physics.

Content:

• what is near-term quantum computing?
• what can quantum computing do? What is it unable to do?
• where can quantum computing be used?

Organization:

• length: ½ a day (e.g., an afternoon)
• up to 25 participants
• no prerequisites

Register now

Which use cases are worth exploring with near-term quantum computing in your company? The selection of use cases worth exploring with near-term quantum computing requires significant background knowledge. Therefore, we help you in the identification of suitable use cases with our expert knowledge.

Day 1 – the ideation phase – dives into the potential applications of QC and quantum-inspired algorithms from the areas of simulation, optimization and machine learning on an interactive basis. At the end of day 1, the participants will have identified use cases that will be analyzed in-depth on day 2 (the use case exploration).

A quantum business canvas designed specifically for this purpose covers both the business and technical side of the application to enable a prototypical implementation after the workshop. Throughout the workshop, a scientist from our team will be present to give advice from the technical side and provide unique insights into the current state of the research field. Of course, the workshop module can be adapted to your organizational needs and particular topic interests.

Content:

• which use cases are of interest for exploration with near-term quantum computing?
• how does the potential of quantum computing translate into business impact?

Organization:

• 2 days (1 day ideation phase, 1 day use case exploration)
• 3-15 participants
• prerequisite: basic knowledge of quantum computing (e.g., the introductory workshop module)
• format: input sessions and creative work groups

For these modules, we particularly value the confidentiality of your internal use cases. The details can be discussed prior to the workshop.

More Info

We build an example implementation that can serve as a first prototype for kick-starting the technical work on a particular use case. The tutorial highlights different options in terms of quantum algorithms and solution paths. This kick-starts the further quantum journey of your company.

Content:

• which quantum algorithms are suitable?
• how to integrate them into the company workflow?

Organization:

• prototypical tutorial implementation and demonstration
• target audience: software developers
• all details on a negotiation basis
  

Whether standalone or building on an initial workshop, we offer you regular touchpoints to advise you on your quantum journey. Our researchers will advise you – in regular technical discussions with your development team, for example – and make sure that your team finds the best direction in its progress towards a practical quantum advantage.

Content:

• feedback, reviews and quick evaluations
• consulting on the basis of state-of-the-art research

Organization:

• on a negotiation basis
• regular technical discussions and reviews with your development team

We evaluate customer use cases to estimate the application-specific potential of quantum-assisted methods. In the context of a use case, we aim to answer three questions:

• what specific problems or subproblems have the potential to be solved better with quantum-enhanced methods?
• what methods should be explored, developed and refined for the use case?
• what further improvements (e.g., in hardware technology) are needed to realize an advantage?

Our goal is to shift away from quantum supremacy claims based on specifically engineered problems to industry-relevant use cases. A thorough analysis needs to include both noisy intermediate-scale quantum (NISQ) methods as well as quantum-inspired algorithms that can already be used today.

A close collaboration with industry partners is the key for realizing a practical quantum advantage of any kind. Flexible modes of partnership are possible. Our past projects include:

• medical imaging: evaluation of hybrid quantum-classical convolutional neural networks for medical imaging tasks in collaboration with the University Hospital of the LMU as part of the Bavarian Competence Center Quantum Security and Data Science.
• supply chain: study of variational quantum algorithms for a route planning problem together with the Quantum Algorithms group at Infineon Technologies AG.

Content:

• potential analysis for specific use cases and/or application fields
• exploration of various quantum-enhanced and quantum-inspired methods
• application-driven benchmarking of solution methods and hardware

Organization:

• on a negotiation basis

We implement state-of-the-art hybrid machine learning methods for specific use cases. Our prototype solutions serve as a starting point for further experiments and scaling, ultimately serving as a blueprint for quantum-classical machine learning models with a real benefit.

Quantum Machine Learning is conjectured to have unique properties that allow handling complex datasets efficiently. An advantage may come in different directions:

• data-efficient learning in the presence of small datasets
• speeding up the computationally expensive training process
• improved modelling of complex probability distributions

In analogy with classical machine learning, we adopt different learning paradigms and architectures depending on the application, including:

• Quantum Convolutional Neural Networks (QCNN)
• Quantum Bayesian Neural Networks (QBNN)
• Quantum Reinforcement Learning (QRL)
• Quantum Kernel Methods (QKM)

 Content:

• implementation and optimization of quantum-enhanced machine learning methods
• application-oriented testing and performance analysis of the implemented solution

Organization:

• on a negotiation basis

• Robustness influences how well the quantum solution can deal with imperfections in the problem definition, underlying dataset, and the algorithm itself. Guaranteeing robustness is especially challenging due to the probabilistic nature of quantum computing and various sources of uncertainty.
• Accuracy refers to the quality of the solutions and their usability for the target application.
• Precision asks about the variance of the results that a particular algorithm or method produces when confronted with different problem instances.

Different fields have varying requirements: frequently, optimization problems benefit from analytical robustness guarantees. In quantum chemistry simulations, a reasonably low frequency of low-quality solutions is acceptable, assuming high-quality solutions are achieved as well. In a medical scenario, undetected low-quality solutions can be fatal.

We combine analytical considerations, numerical simulations and verification techniques to move towards reliable quantum computing while respecting the needs of the particular use case. Naturally, efficient verification and analysis methods are necessary that preserve the potential quantum advantage.

Content

• implementation of robust and reliable quantum-assisted solutions
• performance analysis of the implemented solution under special consideration of its reliability, robustness and expected solution quality

Organization:

• on a negotiation basis

We implement state-of-the-art optimization algorithms and methods and tailor them to your requirements. This includes the combination of quantum algorithms, high-performance computing, classical and quantum-inspired methods.

Combinatorial optimization problems are challenging to solve due to their large number of available options. Classical heuristics only produce approximate solutions. Quantum computing has the potential to explore the large solution space more efficiently by using superposition and entanglement. Its application to real-world use cases, however, remains an enormous challenge.

We design and explore solution paths that have the potential to solve optimization problems with a real benefit. All levels of the solution process are evaluated with the help of an application-specific performance metrics:

• Problem formulation: model a use case as a mathematical optimization problem.
• Decomposition: derive exact or approximate subproblems suitable for different kinds of algorithms and computing paradigms (quantum, classical, high-performance).
• Encoding: prepare the problem or subproblem for solution by a hybrid quantum-classical method.
• Algorithms: select the best hybrid, quantum and classical algorithms and tune them for a high-quality solution.
• Compilation and hardware: choose a suitable compiler and hardware setup from the available technologies (e.g., superconducting, trapped ion, neutral atom quantum computers).

Optimization problems exist in many forms – we provide solution methods that are adapted to your specific use case.

Content:

• implementation of hybrid quantum-classical optimization algorithms
• fine-tuning of the solution paths
• performance analysis of the implemented solution

Organization:

• on a negotiation basis

Whether standalone or building on an initial workshop, we offer you regular touchpoints to advise you on your quantum journey. Our researchers will advise you – in regular technical discussions with your development team, for example – and make sure that your team finds the best direction in its progress towards a practical quantum advantage.

Content:

• feedback, reviews and quick evaluations
• consulting on the basis of state-of-the-art research

Organization:

• on a negotiation basis
• regular technical discussions and reviews with your development team