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:

• Transferable annealing protocols on neutral atom processors were used to optimize smart-charging of electric vehicles.

• Quantum neural networks are applied to entity matching, a key task in data management and artificial intelligence.

• Researchers developed a quantum algorithm for simulating non-adiabatic vibronic dynamics and integrated it into a materials discovery workflow.

QUANTUM APPLICATION HEADLINES

⚡ A Quantum Power Shift — Pasqal and EDF Use Neutral Atom Processors to Optimize EV Smart-Charging

APPLICATION: Pasqal, in collaboration with EDF R&D, used transferable annealing protocols on neutral atom processors to optimize smart-charging of electric vehicles, with an overall goal to focus on adapting quantum optimization strategies for real-world applications in energy management.

SIGNIFICANCE: This study addresses the challenges of scaling quantum optimization algorithms while reducing computational costs by transferring optimized annealing parameters across graph instances with similar geometries. Said another way, optimization settings, or parameters, designed for one problem can be reused for others with similar structural properties. In this case, quantum annealing schedules optimized for one graph are shown to work effectively on other graphs that share similar geometries. This reduces the computational overhead of recalibrating parameters for every new instance, making the approach more practical and scalable for real-world quantum optimization problems. In practical terms, this approach could support EV charging infrastructure by efficiently scheduling charging intervals, in order to support renewable energy use while also minimizing grid strain.

HOW: Variational quantum annealing algorithms on Pasqal’s Orion platform were used to solve Maximum Independent Set problems on graphs. The team used Bayesian optimization to design quasi-adiabatic schedules for parameter transferability. Testing on up to 100 qubits validated the protocol's applicability, including mapping industrial datasets onto triangular graph geometries for smart-charging scenarios.

BY THE NUMBERS:

• 100 qubits – Scale of problems solved using the Orion Alpha platform.

• 98.3% – Success rate of the transferable annealing protocol in noiseless simulations.

• 33 graphs – Size range (9–23 nodes) tested for smart-charging applications using real-world EV charging data from Paris.

• 4.5 µs – The limit of device coherence time, which influenced experimental protocols.

🔗 Quantum Neural Networks Redefine Entity Matching for Big Data Challenges

APPLICATION:  Researchers at Zurich University of Applied Sciences explored the use of quantum neural networks for entity matching—identifying and linking related records—a key task in data management and artificial intelligence. Their hybrid approach combines classical and quantum layers to efficiently identify matching entities, with potential applications in e-commerce, healthcare, and scientific data integration.

SIGNIFICANCE: The study addresses challenges in managing large, heterogeneous datasets where traditional methods struggle with complex and ambiguous records. One example where this might be relevant is in combining patient records from multiple hospitals or clinics which can result in duplicates and variations due to differences in naming conventions, typos, or incomplete data. Through quantum machine learning, computational complexity can be reduced while still achieving comparable or superior performance to classical models with fewer parameters, serving as an example for using quantum computing for real-world problems in data integration.

HOW: The team developed a hybrid quantum neural network with a classical embedding layer and a quantum classification layer. Using handcrafted datasets, the HQNN outperformed some classical models in distinguishing matching entities under challenging conditions. Experiments were conducted on simulators and real quantum hardware to demonstrate the transferability of models between platforms. While TF-IDF had outstanding results, the study demonstrates that combining hybrid quantum neural networks with TF-IDF effectively addresses classification challenges by using HQNN for difficult matches and TF-IDF for simpler cases.

BY THE NUMBERS:

• 6 qubits – Used in the hybrid quantum neural network.

• 386 parameters – Total number of trainable parameters in the HQNN, with 72 in the quantum layer.

• 2,492 pairs – Total number of entity pairs in the training and testing dataset.

• 41.8 hours – Estimated training time required on quantum hardware, showcasing the cost benefits of simulation.

☀️ Quantum Simulations to Boost Solar Efficiency with Singlet Fission

APPLICATION: Researchers at Xanadu, the Vector Institute for Artificial Intelligence, and others developed a quantum algorithm for simulating non-adiabatic vibronic dynamics—related to the study of vibrational and electronic interactions within molecules—and integrating it into a materials discovery workflow for designing singlet fission-based organic solar cells with higher efficiency.

SIGNIFICANCE: Accurately simulating vibronic interactions between nuclear motion and electronic state is something classical methods struggle with because they require solving the time-dependent Schrödinger equation for systems with many degrees of freedom, such as multiple electronic states and vibrational modes which makes the computational cost scale exponentially with system size. This makes it impractical to achieve accurate results for larger or more complex molecules. Additionally, classical methods often rely on approximations that neglect quantum effects such as coherence, tunneling, and transitions that are essential for capturing the full vibronic dynamics. By using quantum computers, the algorithm provides a more comprehensive exploration of materials, such as the ones used photovoltaics which convert sunlight into usable electrical energy.

HOW: The researchers developed a quantum algorithm to simulate vibronic dynamics in order to understand processes like energy and charge transfer in materials. These simulations are essential for optimizing singlet fission, a process where one photon generates two excitons which potentially doubles solar cell efficiency. By modeling the behavior of chromophores (the molecules that absorb light and initiate singlet fission), the algorithm provides insights into designing materials for advanced photovoltaics. Using a trotterization scheme to simulate molecular evolution under a vibronic Hamiltonian, the method handles diverse electronic and vibrational configurations. Simulations demonstrated its effectiveness in analyzing singlet fission chromophores, with resource estimates of up to 1065 qubits for full-dimensional models.

BY THE NUMBERS:

• 1065 qubits – Required for full-dimensional simulations of complex systems like anthracene-fullerene interfaces.

• 200% efficiency – Potential improvement in solar cell efficiency using singlet fission materials.

• Few femtoseconds – Timescale for non-adiabatic processes critical to solar cell function.

RESEARCH HIGHLIGHTS

💭 IBM Quantum and ETH Zurich demonstrated the advantages of soft information decoding for quantum error correction, using analog data from superconducting qubits to improve logical error rates in repetition code experiments. By dynamically weighting decoding graphs based on analog signal ambiguity, soft decoding achieved up to 30 times lower error rates compared to traditional methods, with a 24.4% higher threshold.

🖼️ Researchers from the University of Shanghai present iHQGAN, a hybrid quantum-classical model for unsupervised image-to-image translation, using the invertibility of quantum computing to reduce model parameters. It uses two approximately reversible quantum generators paired with classical neural networks to maintain content consistency. Tested on multiple image tasks, iHQGAN demonstrated improved performance and generalization compared to classical methods while reducing parameter complexity.

🔋 A team from Forschungszentrum Jülich GmbH presents an optimization method using quantum annealing to efficiently sample ionic configurations in materials like lithium cobalt oxide, addressing challenges in simulating systems with exponential configuration spaces. By using a Legendre transformation, the method optimizes the Coulomb energy model while overcoming hardware limitations, successfully identifying ground-state configurations comparable to classical methods.

NEWS QUICK BYTES

🤝 SDT and Malaysia’s MIMOS have partnered to establish Malaysia’s first quantum computing center and a quantum cryptography network. The collaboration will focus on talent development, investment attraction, and integrating quantum technologies into industries to drive Malaysia’s “Quantum Valley” vision.

🎉 Eight startups have been selected under India’s National Quantum Mission and Cyber-Physical Systems Mission to drive advancements in quantum computing, communication, sensing, and materials, aligning with India’s vision for technological self-reliance by 2047. The startups will receive infrastructure, mentoring, and industry support to develop solutions like quantum-safe communication, superconducting hardware, atomic clocks, and advanced quantum materials.

🐈‍⬛ Alice & Bob launched Felis 1.0, the first logical qubit emulator, providing quantum computer scientists with tools to transition from NISQ to fault-tolerant algorithms by emulating logical qubits and predicting their behavior. Built on Qiskit and tailored for Alice & Bob's cat qubit architecture, Felis allows developers to optimize algorithms for fault-tolerant quantum computers, explore error correction techniques, and study hardware efficiency gains.

🧊 SemiQon has released the world’s first cryo-CMOS transistor, designed to function efficiently at cryogenic temperatures, reducing heat dissipation by 1,000x and consuming just 0.1% of the power of traditional transistors. This innovation allows control and readout electronics to operate directly inside cryostats, simplifying quantum computer scaling while reducing costs and energy consumption.

💰️ Atos has received a non-binding offer from the French State to acquire its advanced computing activities, including HPC, quantum, and AI divisions, for an enterprise value of up to €625 million (approximately $656 million). This proposed sale aligns with Atos’ ongoing financial restructuring, with an expected initial payment of €150 million (approximately $157 million) upon signing a Share Purchase Agreement by May 2025.

🚀 Aeluma has been awarded a NASA contract to develop quantum dot photonic integrated circuits on silicon, targeting next-generation space and aerospace applications such as free-space laser communication, autonomous navigation, and precision sensing. This collaboration is intended to enhance optical performance in challenging environments while extending the technology’s applications to AI infrastructure, high-performance computing, and cloud data centers.

QUANTUM MEDIA

WATCH

Jay Gambetta delivers the 2024 IBM Quantum State of the Union, highlighting developments which have enabled algorithm discovery and advanced quantum utility:

THAT’S A WRAP.