Home
Technological innovation driven by quantum computers
Fujitsu's initiatives in quantum utilization
The benefits of Hybrid Quantum-Classical Technology
- Applications of quantum computing
- By applying quantum computing to complex and large-scale real-world problems that are difficult to analyze with classical computing, significant efficiency improvements can be expected.
- Hybridization with classical computing
- By leveraging the strengths of both quantum and classical computing, through measures such as reducing noise that negatively impacts quantum computation and providing assistance like optimization, we expect to apply them to real-world problems.
|
|
|
|
|
|
Technical overview
Target Industry/Users
Research and development companies seeking new quantum accelerated applications for computationally challenging problems that are currently intractable for existing computers from a computational complexity perspective, in the following areas:
• Materials discovery (e.g., development of new materials such as catalysts, semiconductors, and next-generation solar cells; computational chemistry)
• Motion control
• Drug discovery
• Time series prediction (e.g., aerodynamic characteristics, energy, healthcare, finance)
Challenges in Target Industry and Operations
For instance, in the crucial challenge of catalyst candidate material discovery within the materials field, it is necessary to analyze the interaction between candidate catalysts and reactants. However, traditional computers find it extremely difficult to comprehensively simulate the modeling of complex catalyst surfaces and the adsorption of molecules onto these surfaces. Consequently, the current approach involves prototyping materials to collect data, which is then analyzed by experts, leading to huge amounts of time and cost. Furthermore, consider robot motion control: multi-joint robots typically control movement through partial optimization, making more complex tasks harder to control, and presenting challenges in terms of computational load and precision.
Technical Challenges
In computational chemistry, such as catalyst discovery, problems exist that are intractable for conventional computers due to the exponentially increasing computational cost with problem size. Similarly, in robot motion control, an enormous amount of computation is required relative to the number of joints, making it difficult for conventional computers to handle. For such problems, the emergence of quantum-accelerated applications is anticipated, where quantum computers are expected to offer an advantage.
The benefits of Hybrid Quantum-Classical Technology（Detailed version）
We have established a scheme to perform analysis by combining quantum algorithms with optimization (AI) techniques and other methods for problems such as computational chemistry. For example, in a concrete case of the catalyst discovery problem, we utilize the Quantum Fourier Transform, one of the quantum algorithms, for modeling the structures exposed on the catalyst surface, and the adsorption energy calculation of molecules onto that surface is computed using a combinatorial optimization algorithm. Through this approach, we expect future quantum computers to improve the efficiency of exploration for catalysts with complex structures that are currently difficult to analyze with conventional computers. Furthermore, for motion control problems, taking the example of a pitching robot control, we represent joint rotations with qubits and leverage variational quantum algorithms. By combining this with precise optimization from classical computing, we control the motion of multi-joint robots. With this approach, we anticipate that future quantum computers will enhance the efficiency of integrated control for robots with many joints, which is currently challenging for conventional computers.
Fujitsu's Technological Advantage
We possess a wide range of advanced technologies, including quantum computers, HPC (High-Performance Computing), machine learning, and combinatorial optimization techniques, and we are actively promoting the utilization of each of these technologies across various business domains. Therefore, our advantage lies in our ability to develop schemes that can link and control these technologies to execute complex workflows with appropriate computational methods, thereby solving real-world problems.
Use Cases
- End users:
- In fields such as materials science, where empirical testing was previously necessary due to difficulties in simulation and modeling, the utilization of quantum computers and quantum algorithms is expected to enable in-silico modeling. This can lead to a drastic acceleration of materials discovery tasks.
- In motion control, the application of quantum algorithms is anticipated to enable holistic control of multi-joint robots, which were previously managed by combining partial controls due to the difficulty of simulating their complex movements. This opens up prospects for applications in various fields such as space development and precise manipulation.
- App Developers:
- Through the combination of various quantum and classical algorithms, we can expect the realization of an environment where complex real-world problems can be efficiently computed utilizing quantum algorithms.
Case studies
- By combining a quantum simulator and AI optimization in a catalyst discovery application, it is possible to output unique catalyst surfaces and further extract the states of molecules adsorbed onto these surfaces. Examples include surface structure modeling of CeO2 catalysts (La-doped) and PdZn catalysts, and the extraction of multi-molecular adsorption states.
- Through joint angle calculation using a quantum-classical hybrid method, it is possible to more than double the ball speed of a pitching robot compared to classical methods, while simultaneously reducing the load on the robot.