Welcome to the Quantum Realm.

Enjoy today’s breakdown of news, research, & events within quantum.

⚖️ Is two really better than one? Recent research explores the potential of multi-modality algorithms alongside CPU. Plus, Deloitte Italy is combatting digital commerce fraud QML, and a timely reminder of the importance of responsible quantum development.

🗓️UPCOMING

Sunday, July 21 |QTM-X Quantum Education Series 6 of 10: Quantum Hardware

📰QUANTUM QUICK BYTES

🛡️ Deloitte Italy is using QML-based fraud detection with Amazon Braket: As digital commerce grows, fraud detection is an absolute necessity for protecting businesses and consumers, and ML algorithms can analyze vast amounts of transactional data in real-time to identify fraudulent activity. With this in mind, Deloitte Italy has built a hybrid quantum neural network solution using Amazon Braket, integrating quantum computing algorithms with ML models to improve fraud detection in digital payment platforms. Deloitte's solution, although not yet ready for production, gives us a sneak peek into future use cases of QML.

🔗Riverlane and Atlantic Quantum partner to integrate quantum error correction with advanced qubit architectures: Riverlane and Atlantic Quantum announced a strategic partnership to advance quantum error correction. This collaboration will merge Riverlane's QEC stack, Deltaflow, with Atlantic Quantum's superconducting fluxonium-based qubit architecture to develop unique QEC techniques and drive progress toward quantum advantage. Leaders from both companies have emphasized the necessity of similar collaborations in achieving fault-tolerant quantum computing by using combined expertise to accelerate advancements.

🔐AppViewX launches a free online PQC Test Center: AppViewX has introduced the AppViewX PQC Test Center, a free online service that allows organizations to assess their readiness for post-quantum cryptography and begin implementing quantum-safe security measures. As quantum computing advances threaten traditional encryption methods, this platform enables users to generate and test Quantum-Safe certificates as well as provide tools to create a Quantum-Safe PKI hierarchy and test compatibility within their existing architectures.

⚠️Quantum tech is unfolding on a global level and requires regular responsible reflection: The United Nations has designated 2025 as the International Year of Quantum Science and Technology to raise awareness about its profound impact, already unfolding all around us. This technology race is pulling in major global powers and tech giants and is shaping economic and national security policies, with significant investments in quantum startups and infrastructure. However, the rapid development of quantum technologies also brings risks and ethical considerations, necessitating proactive discussions on their global impact, reminiscent of the historical lessons from nuclear technology. This is no profound insight if you’ve been immersed in the quantum field for any significant amount of time, but the importance of paying attention and actively working towards responsible development demands regular reiteration.

Plus, the second episode in IBM’s Quantum Computing in Practice just dropped. Perfect for beginners getting started with quantum circuits, or those needing to polish up their Qiskit 1.0 knowledge:

☕️FRESHLY BREWED RESEARCH

QUANTUM ANNEALER ACCELERATES THE VARIATIONAL QUANTUM EIGENSOLVER IN A TRIPLE-HYBRID ALGORITHM

QUICK BYTE: A triple-hybrid algorithm leveraging the strengths of quantum annealing and quantum gate-based computation is proposed and experimentally validated by finding the ground state energy of an H₂ molecule.

PRE-REQS:

• Multi-hybrid algorithms are presented in the paper as algorithms that run on a combination of at least three types of quantum computing units alongside classical computing, though triple-hybrid (two quantum computing units alongside classical) could also be categorized as such. It’s important to note that instances such as parallel annealing or parallel gate-based computing are not considered multi-hybrid as only one quantum modality is used.

• The graph-coloring problem is a problem within computer science and mathematics in which colors must be assigned to the vertices of a graph as long as two adjacent vertices do not share the same color. This is known to be NP-hard for large graphs.

SIGNIFICANCE: As quantum computation continues to be explored through companies that specialize in specific modalities, we see that different modalities are better suited to specific problem sets. If we temporarily put the question of computation cost aside, it’s logical to consider how the strengths of different modalities can be leveraged by distributing subtasks within the same problem across modalities to most effectively reach the solution. While each subtask could, theoretically, be optimally solved, it is important to note that this is not necessarily the most cost-effective approach for current state technology, but will become increasingly applicable as we improve technology.

While considering how a problem could be more effectively solved using multi-hybrid algorithms, it’s important to consider how subtasks can be distributed across different models of computation. Iterative models are defined by bidirectional input-output relationships, where each system works alternatively such that a loop is established. Sequential models do not establish a loop, but rather each system completes its work in a linear fashion. In the separative model, there is not a defined input-output relationship; the systems work independently of each other.

A selection of examples in which multi-hybrid algorithms would be advantageous are as follows: Using quantum annealing to reduce circuit depth for gate-based operations, using quantum annealing to optimize job scheduling for gate-based quantum and classical computations, and integrating quantum annealing in VQE to simultaneously measure all terms of a Hamiltonian for calculating expectation values.

The proposed triple-hybrid setup for finding the ground state energy involves: inputting the problem Hamiltonian as a string representation into a CPU and converting it into a QUBO problem, the quantum annealer solves the QUBO and outputs bitstrings to the CPU for validation and storage, initial parameters are provided to a classical optimization algorithm on the CPU and embedded into a quantum circuit to be passed to the gate-based quantum unit, the gate-based quantum unit executes the circuit and produces bitstrings that are returned to the CPU, which calculates the energy for all terms in all groups to determine the Hamiltonian's energy. This is experimentally demonstrated by finding the ground state energy of an H₂ molecule as well as solving large graph-coloring problems.

While the decrease in classical computation may not necessarily balance the computation cost of multiple current-state quantum computers, the key takeaway is leveraging the strengths of different quantum computing modalities in tackling specific subtasks can be more effective as opposed to relying on a single modality. This will become more practical as the quality of our technology improves.

RESULTS:

• The algorithm demonstrated a 50% reduction in the resources needed from gate-based quantum computers when evaluating the ground state energy of the H₂ molecule

• The algorithm successfully solved large graph-coloring problems, showing a significant speed-up in the VQE process by allowing simultaneous measurements within commuting groups of Pauli operators

• Future research inspiration: explore the quantitative time benefits of the triple-hybrid algorithm and identify optimization techniques to reduce computational time, integrate the triple-hybrid setup with other methods for accelerating VQE — such as improved embedding techniques and classical post-processing methods

HONORABLE RESEARCH MENTIONS:

A method is presented to mitigate noise in quantum computations, specifically for variational quantum eigensolvers, using a tailored probabilistic machine learning approach based on parametric Gaussian process regression. Demonstrated on a two-site Anderson impurity model and an eight-site Heisenberg model using IBM's Qiskit framework, the approach significantly improves the accuracy of VQE outputs while reducing the number of direct QPU energy evaluations. —> link to Error mitigation in variational quantum eigensolvers using tailored probabilistic machine learning

An investigation of how coherence measurements of a qubit influence its surrounding quantum environment reveals that repetitive Ramsey interferometry measurements steer the environment to the fixed points of an induced quantum channel. Depending on the commutativity between the noise operator and the environment Hamiltonian, three distinct steering effects are identified: polarization, depolarization, and metastable polarization. These effects are illustrated with examples of central spin models, demonstrating the practical impact on measurement statistics and potential applications in designing protocols for quantum state and dynamics engineering. —> link to How coherence measurements of a qubit steer its quantum environment

The vulnerability of QML models to reverse engineering attacks by untrusted third-party quantum cloud providers is explored. The authors present a method for reverse engineering transpiled QML circuits, demonstrating that multi-qubit classifiers can be reverse-engineered with minimal error. They propose countermeasures such as adding dummy rotation gates to increase the difficulty of reverse engineering and ultimately highlight the need for enhanced security measures in the deployment of QML models. —>link to The Quantum Imitation Game: Reverse Engineering of Quantum Machine Learning Models

A new quantum color image watermarking technique is designed to improve the security of digital watermarking by using quantum image representation. The proposed method converts both the watermark and the carrier image into the quantum domain using the New Quantum Binary Image Detector format, resulting in a highly secure and efficient embedding process. —> link to New quantum color image watermarking technique (NQCIWT)

UNTIL TOMORROW.