Welcome to The Daily Qubit!

Get the latest in top quantum news and research Monday through Friday, summarized for quick reading so you stay informed without missing a qubit.

Have questions, feedback, or ideas? Fill out the survey at the end of the issue or email me directly at [email protected].

And remember—friends don’t let friends miss out on the quantum era. If you enjoy The Daily Qubit, pass it along to others who’d appreciate it too.

Happy reading and onward!

Cierra

Today’s issue includes:

• A hybrid classical-quantum neural network framework enhanced by quantum architecture search is used for detecting coronary artery stenosis.

• Quantum reservoir computing is applied to molecular property prediction, especially in terms of drug development.

• Quantum annealing for seismic traveltime inversion may lead to more accurate velocity models for geophysical applications.

QUANTUM APPLICATION HEADLINES

🩺 Quantum Enhanced Diagnostics with Hybrid Neural Networks for Coronary Artery Stenosis Detection

APPLICATION: A team from Zhengzhou University and the Chinese Academy of Science explores a hybrid classical-quantum neural network framework enhanced by quantum architecture search for detecting coronary artery stenosis.

SIGNIFICANCE: Detecting coronary artery stenosis—a narrowing of coronary arteries that reduces blood flow to the heart—requires analyzing complex medical images with high precision. This is a task that often overwhelms classical computational methods due to the high dimensionality of the data and the need for efficient processing. Classical models can struggle with balancing resource efficiency and diagnostic accuracy, especially when working with limited datasets. By integrating quantum circuits, this approach compresses and classifies features in a way that classical methods cannot, reducing computational overhead while maintaining or improving accuracy.

HOW: The framework begins with a pre-trained ResNet18 network, which extracts features from coronary angiography images. These features are resized by a dressed network to reduce dimensionality from 512 to 4, ensuring compatibility with quantum circuits. A distributed quantum circuit, constructed using MCTS, then processes the compressed features, leveraging its optimized architecture to classify the data. By integrating classical pre-training with iterative quantum circuit design, the system achieves high accuracy and generalization, while reducing computational and resource demands.

BY THE NUMBERS:

• 95.9% accuracy — The hybrid model demonstrates high precision in detecting stenosis and outperforms existing classical methods by a 4.1% margin.

• 20 trainable parameters — The quantum circuit layers strike a balance between complexity and generalization, avoiding overfitting while enhancing learning capacity.

• 250 images analyzed — The dataset, comprising equal parts stenosis and non-stenosis X-ray images, validates the model's efficacy in a challenging real-world scenario.

• Up to 3,000 quantum circuits tested — The MCTS algorithm evaluated thousands of architectures to identify the optimal design, ensuring robust and reliable performance.

⚛️ Quantum Reservoir Computing for Predicting Drug Molecule Activity

APPLICATION: A team from Deloitte, QuEra, Merck, and others apply quantum reservoir computing to molecular property prediction, specifically the biological activity of drug molecules. Using molecular fingerprints as input, the QRC framework addresses the limitations of classical machine learning models when working with small datasets.

SIGNIFICANCE: Drug discovery often faces limited data availability due to the high cost, time requirements, and complexity involved in collecting and curating experimental datasets. Additionally, the relationships between molecular properties can be challenging for classical models to decipher, especially when working with small datasets. By using quantum reservoir computing, this solution may contribute notably to stability and interpretability in such data-limited environments, making it valuable in fields where generating large datasets is impractical. The importance of QRC in improving machine learning performance for pharmaceuticals is in the precise and efficient predictions it may produce, which can accelerate drug development and reduce costs.

HOW: The workflow begins with feature selection, using SHAP analysis to identify the top 18 molecular descriptors from the Merck Molecular Activity Challenge dataset. These descriptors are then encoded into the quantum reservoir via the dynamics of neutral atom arrays, producing quantum-enhanced feature embeddings. Classical regression models, including Random Forest and Gaussian Process Regressor, are trained on these embeddings and compared to models trained on classical features. The study also uses the UMAP algorithm for dimensionality reduction, revealing the interpretability advantages of QRC embeddings in low-dimensional spaces.

BY THE NUMBERS:

• 18 features — Reduced from over 4,000 molecular descriptors to 18 critical variables identified via SHAP analysis. These features represent key molecular properties.

• Up to 800 data points — Subsampled data sizes ranged from 100 to 800 molecules. QRC embeddings demonstrated improved performance on smaller datasets (100–200 records), which are often encountered in pharmaceutical research due to the high cost and complexity of data collection.

• 15 molecular activity datasets — The study analyzed subsets of the Merck Molecular Activity Challenge dataset, which provides standardized benchmarks for predicting biological activities. The smallest five datasets were selected to explore QRC’s performance under data-scarce conditions.

• 0.74 accuracy — In binary classification tasks derived from UMAP-projected data, QRC embeddings achieved an accuracy of 74% compared to the 0.66 of classical features, highlighting their improved ability to distinguish between high and low molecular activity values in a simplified 2D space.

🌍️ Quantum Annealing Used in Seismic Velocity Modeling for Carbon Storage

APPLICATION: A team from the Colorado School of Mines demonstrates the use of quantum annealing for seismic traveltime inversion—the process of estimating subsurface velocity models from seismic wave travel times—in order to develop highly accurate velocity models for geophysical applications. The method converts seismic inversion problems into a QUBO format, which is well-suited for quantum annealers.

SIGNIFICANCE: Seismic traveltime inversion is relevant to applications such as carbon storage and subsurface exploration, both of which have significant environmental and economic implications. Carbon storage refers to the process of capturing and storing carbon dioxide underground in geological formations to reduce greenhouse gas emissions and combat climate change. Accurate subsurface models are needed for identifying suitable storage sites, such as depleted oil and gas reservoirs or saline aquifers, and for monitoring the integrity of these sites over time to prevent leaks. Subsurface exploration, on the other hand, involves mapping underground structures to locate valuable resources like oil, gas, minerals, and groundwater. It is also used in engineering projects, such as tunneling and foundation construction, and in environmental studies to assess geological hazards. Both applications rely on precise velocity models to interpret seismic data and make informed decisions. Classical methods often struggle with the noisy and computationally intensive datasets common to these fields. Quantum annealing may be able to more efficiently identify global minima in complex optimization problem.

HOW: The seismic inversion workflow begins with constructing a synthetic velocity model reflecting carbon storage conditions. Seismic traveltime data is collected using non-uniform source-receiver placement to enhance ray coverage. The data is then mapped into a QUBO format and processed using quantum annealing. Iterative inversion is performed, with velocity models refined over 10 iterations. Comparative testing is done against classical Tikhonov-regularized least squares methods under various noise conditions, demonstrating the quantum method's robustness and precision.

BY THE NUMBERS:

• 10 iterations — The seismic inversion process underwent 10 rounds of iterative refinement, each improving the velocity model’s accuracy by adjusting parameters to better match the observed data.

• 0.33% relative error — The relative error in the carbon storage zone, calculated as the deviation between the true and predicted velocities divided by the true velocity, was exceptionally low. This highlights the method's precision in modeling subsurface features critical for applications like carbon storage.

• 5% noise level — The study simulated seismic data with up to 5% random noise, representing real-world challenges in data collection. Despite this, the quantum annealing approach accurately reconstructed velocity models.

RESEARCH HIGHLIGHTS

💭 A study from Northwestern University and Purdue University focuses on improving quantum memory for telecom-band photons by developing a novel "interleaved pumping" method for erbium-doped crystals. The researchers successfully demonstrated efficient storage and retrieval of multidimensional qubits in frequency, time, and polarization bases with high fidelity, without relying on extreme operational conditions like ultra-high magnetic fields.

⚛️ Scientists from PlanQC, RWTH Aachen University, and Forschungszentrum Jülich introduce optimized measurement-free and fault-tolerant quantum error correction schemes for neutral atom platforms, addressing challenges with conventional measurement-based QEC protocols. By using redundant syndrome information and single-shot flags, the proposed method eliminates the need for measurements, reducing overhead and improving fault tolerance.

✅ Researchers from Kyoto University demonstrate the realization of a hybrid ytterbium atom tweezer array for quantum error correction, using nuclear spin qubits for data storage and optical clock qubits as ancillae for nondestructive readout. By minimizing crosstalk and optimizing atom placement, the system achieves high coherence retention (99.1%) and exceptional discrimination fidelity (99.92%) during imaging, making it promising for midcircuit quantum error correction protocols.

NEWS QUICK BYTES

🤝 Pasqal and Riverlane have announced a partnership to integrate Pasqal’s neutral atom quantum computing platform with Riverlane’s Deltaflow quantum error correction stack. This collaboration is intended to accelerate the development of fault-tolerant quantum computers by combining Pasqal’s scalable and stable platform with Deltaflow’s ability to detect and correct computational errors.

✨ Google Quantum AI has announced Willow, a state-of-the-art quantum chip that achieves exponential error reduction as qubits scale, marking a critical milestone in quantum error correction. Willow performed a benchmark computation in under five minutes, which would take today’s fastest supercomputers years, demonstrating its computational power.

🤖 Quantum Machines and Rigetti Computing successfully applied AI to automate the calibration of a 9-qubit Rigetti Novera QPU, demonstrating the potential of AI-powered tools to address one of the biggest challenges in scaling quantum systems: calibration complexity. Through their “AI for Quantum Calibration Challenge,” Quantum Elements and Qruise achieved high gate fidelities and significantly faster setup times, showcasing the effectiveness of automation compared to traditional manual tuning.

💰️ Arctic Instruments, a spinout from Finland's VTT Technical Research Centre, secured €2.35 (approximately $2.47 M) million in funding led by Lifeline Ventures to advance its superconducting microwave amplifiers, relevant for scaling quantum computers to 10,000 qubits and beyond. These near-quantum-limited amplifiers, designed to minimize noise and ensure accurate qubit state readouts, address a critical bottleneck in building large-scale quantum systems. With this funding.

🔬 Infleqtion, in collaboration with NVIDIA and leading universities, has demonstrated the first materials science application powered by logical qubits, achieving a 6x boost in computational accuracy using the NVIDIA CUDA-Q platform. This development, using Infleqtion's neutral atom quantum platform, showcases the potential of logical qubits to address complex problems such as designing next-generation materials, efficient batteries, and high-temperature superconductors.

QUANTUM MEDIA

WATCH

A panel on quantum communication and cryptography’s impact on secure global infrastructure at XPANSE 2024:

THAT’S A WRAP.