Welcome to the Quantum Realm. 

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

🌹 Rose-tinted optimism abounds today with the increasing (or seemingly so) number of practical applications coupled with the initiatives of certain entities — Riverlane announces a brief timeline for one million reliable operations and Spectral Capital fills a deep need in startup land by providing coveted hardware access and training to quantum businesses.

🗓️ UPCOMING

Sunday, July 14 | QTM-X Quantum Education Series 6 of 10: Quantum Hardware

📰 NEWS QUICK BYTES

♾️ Riverlane is progressing fault-tolerant systems with the MegaQuOp milestone: As innovations in error correction technology continue to progress, the fault-tolerance milestone seems to close in, with expectations of such systems by 2035. Riverlane, a name well-known in the error correction space, has just announced a roadmap to reach their MegaQuOp milestone by 2026. This involves performing one million reliable quantum operations and is fueled by developments in qubit quality, algorithms, and quantum error correction from institutions such as Quantinuum, ETH Zürich, and Google. The MegaQuOp scale, requiring thousands of physical qubits, represents one stepping stone toward the TeraQuOp systems that will unlock even more diverse applications in science and engineering.

🌉 Spectral Capital launches program to accelerate the development and commercialization of quantum computing startups: Spectral Capital Corporation has just announced the launch of its Quantum Bridge Program for startups, which will provide comprehensive support across technological, financial, and educational domains. Key features include access to advanced quantum hardware, development tools, collaborative research initiatives, financial backing through seed funding and investor networks, and educational support through workshops and mentorship programs. The program also offers sector-specific benefits to drive innovation and efficiency in industries such as energy, materials, industrials, and healthcare.

🚢 CMA CGM partners with Pasqal for quantum maritime logistics: CMA CGM Group has announced a strategic partnership with Pasqal to integrate quantum computing technologies into its maritime operations by optimizing container management and other logistics operations. This collaboration includes the creation of a Quantum Center of Excellence at TANGRAM, equipped with Pasqal's quantum processor, and quantum computing training for CMA CGM staff. But, not to fret, the partnership is symbiotic: Pasqal, in turn, sees this collaboration as an opportunity to develop practical use cases and advance the understanding of quantum technologies in real-world applications.

🤑 Quantum projected to have a $2 trillion end-user impact by 2035: Though the practical applications of quantum computing are limited, its potential impact across industries could generate a $2 trillion end-user benefit by 2035. Key sectors like chemicals, life sciences, finance, and mobility are set to benefit first, with quantum algorithms providing machine learning models with faster, more accurate data analysis. Other notable applications include those with limited data (such as rare disease detection and defect identification), the optimization of supply chain logistics and materials development through advanced simulations, and the integration of quantum computing into digital twins for improved simulation accuracy and complexity. As companies do what they do — hire talented people to provide creative solutions — it’s more and more possible to see the projected market growth.

💥 Exciton-polariton lasers for low-power quantum applications: Researchers at FLEET have demonstrated that exciton-polariton lasers can achieve an ultra-narrow linewidth of 56MHz, smaller than previously thought. This is especially significant as this narrow linewidth corresponds to a coherence time of 5.7ns, which would make these lasers suitable for quantum computing applications. The long coherence time would allow for the manipulation of the quantum state of the laser as well as the potential for quantum information processing. Additionally, the exciton-polariton lasers can maintain this narrow linewidth even under conditions previously thought to introduce noise, indicating their promise for low-power, high-efficiency applications.

☕️ FRESHLY BREWED RESEARCH

Shadows of quantum machine learning: Intermediate shadow models are proposed to address the challenge of requiring continuous access to quantum computers for quantum machine learning. These models are trained on quantum computers and evaluated classically using information gathered during a shadowing phase, providing a practical solution that retains some quantum advantages while making QML more accessible and practical for real-world applications. Breakdown here.

A Novel Hybrid Quantum Architecture for Path Planning in Quantum-Enabled Autonomous Mobile Robots: The robotics market is projected to reach $349.8 billion by 2032, with autonomous robots being widely used in industries such as logistics, healthcare, and household applications. This research addresses the challenge of path planning in quantum-enabled autonomous mobile robots by proposing a Hybrid Quantum Ant Colony Optimization algorithm. The HQACO algorithm successfully automates optimal path formation with an average error percentage of 6.985% from optimal solutions. Breakdown here.

Entanglement distribution based on quantum walk in arbitrary quantum networks: Quantum walks are used for efficient entanglement distribution in quantum networks, proposing schemes to generate high-dimensional entangled states like Bell states and GHZ states. By using quantum repeaters and quantum walks with multiple coins, the study demonstrates methods to achieve long-distance entanglement in complex quantum networks. This provides a foundation for constructing more advanced quantum networks, with practical applications including quantum fractal networks and multiparty quantum secret sharing protocols.

Variational quantum cloning machine on a photonic integrated interferometer: This is the first experimental implementation of a variational quantum cloning machine on a photonic integrated platform. By using a programmable 6-mode universal interferometer and classical feedback, the study demonstrates near-optimal cloning performances for both phase-covariant and state-dependent cloning of dual-rail encoded photonic qubits. The results highlight the potential of integrated photonic platforms for self-learning and executing quantum algorithms.

Towards Photon-Number-Encoded High-dimensional Entanglement from a Sequentially Excited Quantum Three-Level System: A sequential two-photon resonant excitation process in a solid-state three-level quantum system is experimentally implemented using semiconductor quantum dots. By using energy-resolved and time-resolved correlation experiments, the study successfully generates high-dimensional entangled photon states encoded in the photon-number basis. This has applications in quantum information processing, including dense information encoding and advanced quantum communication protocols.

Efficient and secure quantum secret sharing for eight users: An efficient and secure quantum secret sharing protocol for eight users is demonstrated using a continuous-variable eight-partite bound entangled state. By using advanced entanglement generation techniques, precise phase control, and fiber distribution with polarization-division multiplexing, the research achieves high key rates and flexible, scalable multiuser quantum communication. The results show that this QSS system can efficiently and securely distribute quantum secrets among multiple users.

UNTIL TOMORROW.

BREAKDOWN

Shadows of quantum machine learning

🔍️ SIGNIFICANCE: 

• If you regularly keep up with the literature, despite the rather narrow application space as compared to classical computation, it can seem that there are infinite ways in which quantum computing is being explored across domains. However, one subdomain is consistently touted as one of the most promising for practical applications — quantum machine learning.

• What makes quantum machine learning one of the more obvious approaches for quantum advantage comes down to a handful of features, including the use of Grover’s or Shor’s algorithms for speeding up search, quantum kernel methods that map data into high-dimensional Hilbert spaces, and the ability to learn from fewer data points. However, a significant limitation is the need for access to a quantum computer for initial training as well as updating with new data. Access to real quantum computers is not only expensive, it’s severely limited (if not nonexistent) in many places.

• The goal of this research is to develop a new spin on training QML models that bypass the need for continual quantum computer access. To achieve this, we first consider the two different training methods. For both classical training and quantum training, both types follow a similar blueprint, where they are trained and then deployed to evaluate new data. For QML models, both phases require quantum computer access. As a solution, the researchers propose adding a shadowing phase between training and deployment. During this shadowing phase, a quantum computer gathers information on the quantum model, which a classical computer can then use to evaluate new data.

• While previous work suggested shadow models as a compromise, it also indicated that classical data might be trained on the same data and achieve similar performance, thus negating the quantum advantage. Inspired by these insights, the researchers have developed a general definition of a shadow model to explore and test whether this approach still provides a quantum advantage.

🧪 METHODOLOGY: 

• Shadow tomography is used to create shadow models. In shadow tomography, a quantum state’s properties are estimated using a minimal number of measurements (creating a “shadow” for multiple queries without direct measurement). These models are trained on a quantum computer but evaluated classically on new data using information collected during the training phase.

• The key technique involves using flipped models, where the roles of quantum states and observables are reversed compared to conventional models. This flipping is what assists in the generation of shadow models that can be efficiently evaluated classically.

• Quantum information theory and computational complexity assumptions are used to illustrate the practicality and advantages of shadow models.

📊 OUTCOMES & OUTLOOK: 

• Overall, the shadow models were shown to achieve a quantum advantage over fully classical models in certain learning tasks. However, it’s important to explicitly note that they have restricted learning capacities compared to fully quantum models.

• Shadow tomography is a practical and efficient method for constructing shadow models. So, creating the shadow models for a lesser degree of quantum advantage is still practical enough to provide merit.

• Providing a pathway to gain smaller, though still important quantum advantages without the continuous need for quantum hardware is important in that it can accelerate the integration of quantum machine learning while being mindful of computational cost, as well as allowing more communities to benefit where access is limited.

BREAKDOWN

A Novel Hybrid Quantum Architecture for Path Planning in Quantum-Enabled Autonomous Mobile Robots

🔍️ SIGNIFICANCE: 

• The robotics market is estimated to reach $349.8 billion worldwide by 2032, according to a recent report from Allied Market Research. Autonomous robots are used across industries, including in logistics and warehouses for inventory management and order fulfillment, within healthcare as surgical and autonomous disinfection robots, and as everybody’s favorite household pet — autonomous vacuums, to name a few. In many use cases, especially when considering order fulfillment, the Traveling Salesman Problem is a common framework used to map the problem.

• If you’re not familiar with the Traveling Salesman Problem, consider the following general problem statement: If given a list of greographical locations, how would you determine the shortest route possible that allows you to visit each location exactly once, while ending at the starting location? This is considered an NP-hard problem in computer science — which equates to the potential to see quantum advantage over purely classical methods.

• While theoretical, the researchers consider the use case where mobile autonomous robots, required to pull inventory and deliver to multiple locations with a return to the starting point, are enabled by quantum computing techniques to construct an optimal path such that time and energy consumption are mitigated.

• One more aspect to consider is that this process must be automated. Since solving via quantum computation would require previous consideration of how to map to qubits, simply solving with quantum computation on its own is not an option. Not to mention, the NISQ era is not the time for mobile QPUs due to the impracticality caused by noise. That being said, the goal of this exploration is to solve for the autonomous part in order to provide insight for future applications.

• As a solution to the above, the researchers consider the Ant Colony Optimization, a commonly used and efficient classical computation algorithm. The proposal is to implement classical computation so that it handles most of the ACO algorithm, and then quantum computation is used for the probabilistic selection step, ensuring that the QPU is used only where necessary. This novel Hybrid Quantum Ant Colony Optimization algorithm automates the path formation, using the combined strengths of quantum and classical computing to overcome the limitations of manual qubit conversion and noise.

🧪 METHODOLOGY: 

• A hybrid classical-quantum system that implements ACO was designed. Quantum circuits are used to generate quantum states corresponding to the probabilities of different paths, followed by measurements to select the next node in the path.

• The classical processor handles arithmetic calculations and updates pheromone values, while the quantum processor generates states with the required probabilities to enhance the randomness and efficiency of the selection process.

• A quantum selector is initialized with parameters derived from the probability distributions of paths and iteratively refines the path based on quantum measurements and classical calculations.

📊 OUTCOMES & OUTLOOK: 

• Notably, this approach does not require manual conversion of problem elements into quantum bits, which is a significant improvement over the existing methods mentioned in the literature review.

• The simulations of the proposed HQACO algorithm on several instances of the Traveling Salesman Problem indicate promising results with an average error percentage from optimal results of only 6.985%. The results suggest that HQACO can effectively automate path planning in quantum-enabled mobile robots and improve their operational efficiency, provided that QPUs within mobile devices become practical themselves.