Wednesday, September 25th, 2024

Enjoy a nice cup of freshly brewed quantum news ☕️ 

Today’s issue includes:

• Researchers an ensemble framework of hybrid quantum-classical convolutional neural networks for classifying breast cancer histopathological images.

• Quantinuum scientists explore the use of a trapped-ion quantum computer to measure correlation and entanglement between molecular orbitals.

• A review from NVIDIA, Moderna, and Yale explores the intersection of quantum computing and machine learning in drug discovery.

• The G7 Cyber Expert Group released a statement urging proactive planning for quantum resilience and the potential vulnerabilities in encryption.

• Plus quantum reinforcement learning for drone fleets, the latest in research funding, entanglement at the hadron collider, and more.

And even more research, news, & events within quantum.

🩻 Breast Cancer Image Classification Using Hybrid Quantum-Classical Neural Networks

QUICK BYTE: Researchers from Jadavpur University propose an ensemble framework of hybrid quantum-classical convolutional neural networks for classifying breast cancer histopathological images, achieving improved accuracy over individual models, with the best ensemble performance reaching 86.72%.

DETAILS

• The authors implemented three hybrid quantum-classical neural network architectures combining classical convolutional neural networks for feature extraction with parameterized quantum circuits for final classification, applying these to breast cancer histopathology images from the BreakHis dataset.

• The best-performing individual model achieved 85.59% accuracy, while the ensemble techniques, particularly average probability between two models, improved accuracy to 86.72%, precision to 86.54%, recall to 82.49%, and F1-score to 84.04%.

• The ensemble approach used simple techniques such as majority voting and weighted probability averaging to combine predictions from different models, taking advantage of complementary insights from non-overlapping misclassifications across models.

• This work highlights the potential of hybrid quantum-classical models in medical image classification, though the increased complexity and longer training time for quantum neural networks due to quantum state preparation and gate execution remain limitations.

⚛️ Quantinuum Scientists Measure Orbital Correlations and Entanglement in Quantum Chemistry, Relevant to Battery Degradation

QUICK BYTE: Quantinuum scientists explore the use of a trapped-ion quantum computer to measure correlation and entanglement between molecular orbitals during the formation of dioxetane from vinylene carbonate and singlet oxygen, demonstrating that quantum computation can accurately estimate orbital correlations and entanglement relevant to quantum chemistry processes.

DETAILS

• The researchers used a Quantinuum trapped-ion quantum computer to calculate von Neumann entropies and mutual information, quantifying orbital correlations and entanglement in the vinylene carbonate and singlet oxygen reaction, which is relevant to lithium-ion battery degradation processes.

• By using the Jordan-Wigner transformation and a variational quantum eigensolver, they computed the orbital reduced density matrices with reduced measurement overheads, showing that quantum hardware can accurately estimate these quantities, with entropies in agreement with noiseless benchmarks.

• The study highlights the importance of accounting for fermionic superselection rules, which reduced the number of measurements needed and corrected for overestimated correlations by constraining quantum operators to respect fermionic symmetries.

• The findings reveal that one-orbital entanglement is negligible unless open-shell spin configurations are present, which provides valuable insights into the quantum mechanical nature of the molecular states involved in the dioxetane reaction.

🤖 NVIDIA, Moderna, and Yale Review Quantum Neural Networks in Drug Discovery, Highlighting Hybrid Quantum-Classical Approaches for Molecular Prediction

QUICK BYTE: A review from NVIDIA, Moderna, and Yale explores the intersection of quantum computing and machine learning in drug discovery, with a focus on the application of quantum neural networks for molecular property prediction and generation.

DETAILS: 

• The study discusses the foundational elements of quantum machine learning and the increasing use of hybrid quantum-classical methods, such as quantum neural networks, for tasks related to molecular property prediction and virtual screening in drug discovery.

• It highlights the potential benefits of quantum methods, with particular focus on quantum convolutional neural networks and quantum graph neural networks in tasks such as predicting molecular stability and protein-ligand binding affinities.

• The research emphasizes the need for hybrid approaches that integrate quantum and classical computing, using variational quantum circuits to manage current hardware constraints.

⚠️ International G7 Cyber Expert Group Urges Financial Sector to Prepare for Quantum Computing Risks and Encryption Vulnerabilities

QUICK BYTE: The G7 Cyber Expert Group released a statement outlining both the opportunities and risks posed by quantum computing for the financial sector, urging proactive planning for quantum resilience and the potential vulnerabilities in encryption.

DETAILS: 

• The G7 Cyber Expert Group, established in 2015, coordinates cybersecurity policy and strategy across G7 countries, focusing on improving the cyber resilience of the financial sector through incident response, preparedness, and threat mitigation. Members include financial authorities from Canada, France, Germany, Italy, Japan, the United Kingdom, the United States, and the European Union.

• The report released by the group emphasizes the need for financial institutions to monitor developments in quantum computing, noting that while quantum technology offers potential benefits, it also poses risks to current encryption methods used to secure financial systems.

• The risk to public-key cryptography is explicitly highlighted for the concerns around breaking widely used encryption algorithms, which could expose sensitive financial data.

• Financial entities are advised to begin assessing quantum risks, develop quantum resilience strategies, and coordinate with key stakeholders to prepare for a post-quantum environment, with a particular focus on replacing vulnerable technologies and maintaining cryptographic security.

Optimal sampling can be achieved with shallow circuits, even under realistic noise conditions, according to a study by scientists from Brookhaven National Laboratory, Argonne National Laboratory, and others. The researchers investigate the optimal depth of twirling circuits used in classical shadows protocols for estimating quantum states in the presence of noise. While shallow-depth circuits with local entangling gates minimize sample complexity, the effectiveness of deeper circuits is limited by noise. The study also derives noise thresholds and an upper bound on the circuit depth.

The EU has allocated €65M (approx. $73M) to support semiconductor research and innovation through the Chips Joint Undertaking, focusing on quantum chips for computing and sensing. This funding is part of a larger €200M (approx. $223M) investment over three years intended to support the development of quantum technologies and establishment of a European manufacturing supply chain. Proposals for funding can be submitted by 21 January 2025.

The U.S. Commerce Department is seeking increased investments and policy measures to strengthen the quantum information science sector and its technology supply chain, according to Deputy Secretary Don Graves. During a recent event, Graves emphasized the importance of strategic investments in quantum technologies and addressing supply chain vulnerabilities, such as rare earth material dependencies and specialized equipment like dilution refrigerators. The department’s new supply chain risk assessment tool, SCALE, will help identify and mitigate supply chain risks to ensure continued U.S. leadership in quantum technology development.

A white paper by SandboxAQ and EY discusses how AI and quantum-inspired technologies are transforming materials discovery by speeding up traditionally slow and costly R&D processes. AI-driven simulations, when combined with quantum mechanics, enable faster and more accurate predictions of material behavior, reducing the reliance on expensive lab experiments. These technologies democratize materials research, allowing smaller companies to compete by accelerating innovations in fields like battery degradation and energy storage.

Sandia National Laboratories, the University of Michigan, and Brigham Young University develop feedback-based quantum algorithms designed for efficient ground state preparation, particularly for Fermi-Hubbard and molecular Hamiltonians. Unlike traditional variational quantum algorithms which require complex classical optimization, FQAs use a feedback law based on quantum Lyapunov control to update circuit parameters. The study demonstrates the algorithm's convergence through simulations on the Fermi-Hubbard model and molecular systems, showing scalability, especially in noisy quantum computing environments.

Tech Mahindra and the University of Auckland signed an MoU to advance research in AI, machine learning, and quantum computing, focusing on industries like healthcare, finance, and government. The collaboration includes developing spiking neural networks, 1-bit large language models, and post-quantum cryptography, with applications in drug discovery and personalized digital biomarkers. This partnership also intends to support graduate employability through internships and hands-on training.

LISTEN

The most recent episode of the Physics World Weekly podcast, hosted by Hamish Johnston, features Elena Blokhina, chief scientific officer at Equal1, discussing the use of quantum dots as qubits in the company's silicon-based quantum–classical computing chips, and Brandon Grinkemeyer, a Harvard PhD student, who talks about his work on developing quantum processors using arrays of trapped atoms as qubits in Misha Lukin's research group.

ENJOY

Interest in quantum computing is spreading, with major banks such as JPMorgan Chase, Wells Fargo, and HSBC testing its potential to optimize portfolios and improve cybersecurity. JPMorgan Chase, for instance, has invested $100 million in Quantinuum and is developing quantum algorithms to solve business problems, while HSBC is using quantum technology to protect digital assets. Though quantum is still in its R&D phase, experts have encouraged the importance of building teams now to be "quantum-ready" and maintain a competitive edge.

WATCH

A look into how SandboxAQ’s AQBioSim is contributing to drug discovery, particularly in neurodegenerative and oncology research: