rel="stylesheet">
Creating and sharing knowledge in communications and information technology

IT Research Selected by IBM to Run Quantum AI Experiment on Real Quantum Hardware


by IT on 13-10-2026
Project Real Quantum Hardware Multimedia Signal Processing Information and Data Sciences
...

A research project involving the Instituto de Telecomunicações, Coimbra branch, and NYU Abu Dhabi has been selected by IBM’s Quantum Credit Program, receiving US$30,000 in IBM Quantum credits.

The selected work, “Enhancing Blood Cells Classification using Hybrid Quantum Neural Networks,” investigates the potential of quantum machine learning for the automated classification of blood cells. The study combines a conventional deep neural network with a quantum neural network, using controlled comparisons with equivalent classical models to determine the actual contribution of the quantum component.

The recognition is particularly significant because, among the papers accepted in the field, only a small number were selected to progress from simulation to experiments on real quantum processors. The team’s work is one of the selected studies.

In an interview, our senior researcher Gabriel Falcão and DEEC-UC Director Guilherme Cruz discuss the significance of the award, the challenges of moving from quantum simulation to physical quantum processors, and the potential of quantum-enhanced AI in medical imaging.

 

1 - What is the work selected by IBM, and what is the main novelty of the research?

The selected work, “Enhancing Blood Cells Classification using Hybrid Quantum Neural Networks,” investigates whether quantum computing can improve AI-based classification of microscopic blood cells. Our main contribution is a hybrid architecture that combines a conventional deep neural network with a quantum neural network. Importantly, we designed controlled ablation studies and comparisons with equivalent classical models, allowing us to isolate the contribution of the quantum component rather than simply attributing improvements to a more complex model.

 

2 - In practical terms, what does it mean to run this work on a real quantum computer rather than only through simulation?

Simulation reproduces the mathematical behavior of a quantum computer using classical hardware. Running the model on a real quantum processor is fundamentally different: the computation is performed by physical qubits, which are affected by noise, decoherence, and hardware imperfections. This allows us to test whether an algorithm that works in an idealized simulation remains viable under real quantum-computing conditions.

 

3 - Why was this work selected by IBM, and what does the award of US$30,000 in credits represent?

The award recognizes the relevance of exploring quantum machine learning in a concrete and socially important application domain such as medical imaging. The US$30,000 in IBM Quantum credits gives us valuable access to real quantum computing resources, allowing us to move beyond simulation and conduct substantially more experiments on quantum hardware. For us, the most important aspect is precisely this opportunity to test how these algorithms behave on current and future generations of quantum processors.

 

4 - What potential applications could the use of quantum neural networks have in the detection of blood cells?

Automated blood-cell analysis can support disease screening, diagnosis, and treatment monitoring. Our results suggest that quantum neural networks may be particularly useful when different cell types have subtle morphological differences and are therefore difficult for conventional AI models to distinguish. The objective is not to replace clinicians, but to investigate whether quantum-enhanced AI can eventually provide more accurate and robust computational tools to support clinical analysis.

 

5 - Why was the HPC infrastructure at NYU Abu Dhabi essential to this research?

Quantum machine-learning research currently relies heavily on classical high-performance computing. Before running an experiment on scarce and expensive quantum hardware, we need to train the neural networks, simulate quantum circuits, explore different configurations and perform extensive comparisons. GPU acceleration was particularly important for simulating the variational quantum circuits and for training the classical components of our hybrid architecture. HPC and quantum computing should therefore be seen as complementary technologies rather than competing ones. The paper uses GPU-accelerated quantum simulation through PennyLane's lightning.gpu backend.

 

6 - What does this achievement mean for Guilherme Cruz, as the first author, and for IT-Coimbra? 

For Guilherme, as the first author, this is an important recognition of the quality and relevance of his research at an early stage of his scientific career. For Instituto de Telecomunicações and the University of Coimbra, it demonstrates our capacity to conduct internationally competitive research in quantum computing and quantum AI, working closely with leading international institutions such as NYU Abu Dhabi. The paper itself is a good example of this collaboration, bringing together researchers from Coimbra and NYU Abu Dhabi.

 

7 - What are the next steps you hope to take based on this experience with real quantum hardware?

The next step is to use the IBM Quantum credits award to conduct a much broader experimental study on real quantum processors. We want to investigate larger and different quantum circuits, evaluate the effects of hardware noise, explore error-mitigation techniques, and understand which characteristics of quantum models can provide genuine advantages over classical approaches. Ultimately, the scientific question is not simply whether we can run AI on a quantum computer, but when and why quantum computation can offer an advantage for medically relevant problems.

 

8 - What does this experience demonstrate about the importance of computing infrastructures and international collaborations for research in quantum computing? 

Modern quantum-computing research requires an entire computational ecosystem. HPC and GPUs are needed for training, simulation, and large-scale experimentation, while quantum processors allow us to validate algorithms under real physical conditions. At the same time, these infrastructures are expensive and highly specialized, making international collaboration essential. This work illustrates that combination very well: expertise and research from Coimbra, high-performance computing and quantum expertise at NYU Abu Dhabi, and access to IBM quantum hardware.

 

9 - What societal impacts could this research have?

The potential societal impact is primarily in healthcare. More accurate and automated analysis of medical images could support earlier diagnosis, reduce the workload associated with repetitive image analysis, and help clinicians identify subtle patterns that are difficult to distinguish visually. However, quantum computing is still an emerging technology, so these results should be viewed as an important research step rather than an immediately deployable clinical solution. Our goal is to establish whether quantum-enhanced AI can eventually contribute to faster, more accurate, and more accessible diagnostic technologies.


SHARE: