Applying Quantum Machine Learning to Diverse Quantum Data

OrbitAll: A Unified Quantum Mechanical Representation Deep Learning Framework for All Molecular Systems Accurately modeling chemical systems across diverse charges, spin states, and environments remains a central challenge in molecular machine learning. No existing machine learning–based methods can simultaneously handle molecules with varying charges, spins, and environments. A few recently developed approaches address one or two of these factors individually by designing task-specific architectures, but this limits their applicability to broader chemical scenarios. OrbitAll is the first deep learning-based method that can simultaneously incorporate spin, charge, and environmental information using consistent and physically grounded quantum mechanical features. It has superior accuracy, generalization, and data efficiency on diverse chemical systems. We introduce a unified quantum mechanical representation that naturally incorporates spin, charge, and environmental effects within a single, physics-informed framework. Specifically, OrbitAll utilizes spin-polarized orbital features from the underlying quantum mechanical method, and combines it with graph neural networks satisfying SE(3)-equivariance. This enables our model, OrbitAll, to achieve accurate, robust, and data-efficient predictions across a wide range of chemical systems–including charged and open-shell species, as well as solvated molecules–without the need for domain-specific tuning. OrbitAll achieves chemical accuracy using 10 times fewer training data than competing AI models, with a speedup of more than thousand times compared to density functional theory. It can extrapolate to molecules more than 10times larger than those in training data. This universality distinguishes our approach from current deep learning models.

I’d like to draw your attention to a new paper on arXiv, “Shallow-circuit Supervised Learning on a Quantum Processor”, from IBM and Qognitive that develops a Hamiltonian-based framework for quantum machine learning. Instead of fixed amplitude or angle encodings used in many prior approaches, our method learns a local Hamiltonian embedding for classical data. https://lnkd.in/ejcxYstW We are very interested in new approaches to QML as we deal with recurring bottlenecks like expensive classical data loading and difficult training dynamics in parameterized circuit models. Here, both the feature operators and the label operator are learned during training, with predictions obtained from measurements on an approximate ground state. This aims to avoid those bottlenecks. A key enabler is Sample-based Krylov Quantum Diagonalization (SKQD), which approximates low-energy states by sampling from time-evolved Krylov states and then diagonalizing the Hamiltonian in the sampled subspace. SKQD was recently employed to estimate low-energy properties of impurity models (https://lnkd.in/epwCrG5R). In our setting, restricting to 2-local Hamiltonian embeddings keeps the required time-evolution circuits relatively shallow, which helps make the approach practical on current quantum processors. The team demonstrates end-to-end training on IBM Heron processor up to 50 qubits, with non-vanishing gradients and strong proof-of-concept performance on a binary classification task. There are many exciting next steps here, including testing on broader datasets, using more expressive operator ansatz, and performing systematic comparisons to strong classical baselines to pinpoint when Hamiltonian-based encodings offer the right inductive bias. I encourage the community to try out this approach and explore where it can be extended in meaningful ways.

𝗤𝘂𝗮𝗻𝘁𝘂𝗺 𝗣𝗿𝗼𝗯𝗮𝗯𝗶𝗹𝗶𝘁𝘆 × 𝗟𝗟𝗠 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝖰𝗎𝖺𝗇𝗍𝗎𝗆 𝖺𝗆𝗉𝗅𝗂𝗍𝗎𝖽𝖾𝗌 𝗋𝖾𝖿𝗂𝗇𝖾 𝗅𝖺𝗇𝗀𝗎𝖺𝗀𝖾 𝗉𝗋𝖾𝖽𝗂𝖼𝗍𝗂𝗈𝗇 𝖯𝗁𝖺𝗌𝖾 𝖺𝗅𝗂𝗀𝗇𝗆𝖾𝗇𝗍 𝖾𝗇𝗋𝗂𝖼𝗁𝖾𝗌 𝖼𝗈𝗇𝗍𝖾𝗑𝗍𝗎𝖺𝗅 𝗇𝗎𝖺𝗇𝖼𝖾 Classical probability treats token likelihoods as isolated scalars, but quantum computation reimagines them as amplitude vectors whose phases encode latent context. By mapping transformer outputs onto Hilbert spaces, we unlock interference patterns that selectively amplify coherent meanings while cancelling noise, yielding sharper posteriors with fewer samples. Variational quantum circuits further permit gradient‑based training of unitary operators, allowing language models to entangle distant dependencies without the quadratic memory overhead of classical self‑attention. The result is not simply faster or smaller models, but a fundamentally richer probabilistic grammar where superposition captures ambiguity and measurement collapses it into actionable insight. As qubit counts rise and error rates fall, the convergence of quantum linear algebra and deep semantics promises a new era in which language understanding is limited less by data volume than by our willingness to rethink probability itself. #quantum #ai #llm

⚛️ Parallel Data Processing in Quantum Machine Learning 🧾 We propose a Quantum Machine Learning (QML) framework that leverages quantum parallelism to process entire training datasets in a single quantum operation, addressing the computational bottleneck of sequential data processing in both classical and quantum settings. Building on the structural analogy between feature extraction in foundational quantum algorithms and parameter optimization in QML, we embed a standard parameterized quantum circuit into an integrated architecture that encodes all training samples into a quantum superposition and applies classification in parallel. This approach reduces the theoretical complexity of loss function evaluation from O(N^2) in conventional QML training to O(N), where N is the dataset size. Numerical simulations on multiple binary and multi-class classification datasets demonstrate that our method achieves classification accuracy comparable to conventional circuits while offering substantial training time savings. These results highlight the potential of quantum-parallel data processing as a scalable pathway to efficient QML implementations. ℹ️ Ramezani et al - 2025

How should we measure if there is any benefit from using quantum machine learning methods in QNNs? Raw accuracy? I discovered an interesting study that for sure will spark some ideas about how to answer this question. I recently came across an insightful study titled "QMetric: Benchmarking Quantum Neural Networks Across Circuits, Features, and Training Dimensions". This paper introduces a much-needed tool in the evolving field of hybrid quantum-classical machine learning. The core challenge addressed by this research is the lack of principled, interpretable, and reproducible tools for evaluating hybrid quantum-classical models beyond traditional metrics like raw accuracy. These standard diagnostics don't capture crucial quantum characteristics such as circuit expressibility, entanglement structure, barren plateaus, or the sensitivity of quantum feature maps. To bridge this gap, the authors present QMetric, a modular and extensible Python package (for now available for Qiskit and PyTorch). QMetric offers a comprehensive suite of interpretable scalar metrics designed to evaluate quantum neural networks (QNNs) across three complementary dimensions: * Quantum circuit behavior: Metrics like Quantum Circuit Expressibility (QCE), Quantum Circuit Fidelity (QCF), Quantum Locality Ratio (QLR), Effective Entanglement Entropy (EEE), and Quantum Mutual Information (QMI) assess a circuit's representational capacity, robustness to noise, balance of gate operations, and internal correlations. * Quantum feature space: This category includes Feature Map Compression Ratio (FMCR), Effective Dimension of Quantum Feature Space (EDQFS), Quantum Layer Activation Diversity (QLAD), and Quantum Output Sensitivity (QOS), which analyze how classical data is encoded into quantum states, the geometry of the resulting feature space, and its robustness to perturbations. * Training dynamics: Metrics such as Training Stability Index (TSI), Training Efficiency Index (TEI), Quantum Gradient Norm (QGN), and Barren Plateau Indicator (BPI), along with relative metrics like Relative Quantum Layer Stability Index (RQLSI) and Relative Quantum Training Efficiency Index (r-QTEI), provide insights into convergence behavior, parameter efficiency, and gradient issues. This study illustrates how QMetric (or this suggested set of metrics) can help researchers diagnose bottlenecks, compare architectures, and validate empirical claims beyond raw accuracy, guiding more informed model design in quantum machine learning. Here the article: https://lnkd.in/dnZdufdu Here the repo: https://lnkd.in/dpE3sB2W #qml #quantum #machinelearning #ml #quantumcomputing #datascience

🚀 New Paper on arXiv! I’m excited to share our latest work: “Learning to Program Quantum Measurements for Machine Learning” 📌 arXiv: https://lnkd.in/euRhBQJM 👥 With Huan-Hsin Tseng (Brookhaven National Lab), Hsin-Yi Lin (Seton Hall University), and Shinjae Yoo (BNL) In this paper, we challenge a long-standing limitation in quantum machine learning: static measurements. Most QML models rely on fixed observables (e.g., Pauli-Z), limiting the expressivity of the output space. We take this one step further--by making the quantum observable (Hermitian matrix) a learnable, input-conditioned component, programmed dynamically by a neural network. 🧠 Our approach integrates: 1. A Fast Weight Programmer (FWP) that generates both VQC rotation parameters and quantum observables 2. A differentiable, end-to-end architecture for measurement programming 3. A geometric formulation based on Hermitian fiber bundles to describe quantum measurements over data manifolds 🧪 Experiments on noisy datasets (make_moons, make_circles, and high-dimensional classification) show that our dual-generator model outperforms all traditional baselines—achieving faster convergence, higher accuracy, and stronger generalization even under severe noise. We believe this work opens the door to adaptive quantum measurements and paves the way toward more expressive and robust QML models. If you're working on QML, differentiable quantum programming, or quantum meta-learning, I’d love to connect! #QuantumMachineLearning #QuantumComputing #QML #FastWeightProgrammer #DifferentiableQuantumProgramming #arXiv #HybridAI #AI #Quantum

*Is there a path to achieving a quantum computational advantage for machine learning with NISQ systems on problems of practical interest?* The community consensus seems to be _no_ for almost all problems involving classical data. But, as has been pointed out often over the past few years, if you combine quantum sensing with quantum machine learning, then there appears to be a path to a combined *quantum sensing-computational advantage*. An exciting possibility is that it might be feasible to achieve this kind of quantum advantage for processing extremely weak microwave signals, e.g., those comprising just a few photons per microsecond. In this case the quantum sensor would be an extremely sensitive detector of microwave photons, which can be made using superconducting circuits. While the photons are being received, quantum processing can be performed directly on the few-photon signal in the same circuit. We have implemented a proof-of-principle of the quantum-processing part of this direction: we have shown that we can use a superconducting circuit as an analog quantum dynamical system to perform quantum reservoir computing directly on analog microwave signals. Our paper presenting our scheme and experiments is out in Nature Communications today: https://lnkd.in/ePxqqSet . The number of microwave photons inside our superconducting circuit at any given time is small (<10), and yet we are able to accurately classify signals with a few thousand shots -- showing that it is possible in principle to classify very weak signals. We haven't yet demonstrated a quantum advantage, but we think this is a promising direction. Future work will involve building and integrating a subsystem capable of high-quantum-efficiency detection of microwave photons, and then exploring if there is a regime of advantage. Congratulations to Alen Senanian and all the other authors, including our collaborators from Xiaodi Wu's group and Valla Fatemi's group! Finally, I can also recommend reading the paper from Hakan Tureci's group on quantum reservoir computing, also out today in Nature Communications.

Trying to enter QML in 2026? This is the path I’d take, step by step. A Quantum Machine Learning roadmap should build three pillars in parallel: 1)Mathematics & Classical ML foundations 2)Quantum Computing foundations 3)Hybrid Quantum-Classical ML implementation → Advance QML Models Think of QML as ML + Linear Algebra + Quantum Mechanics + Optimization Step 1: Mathematics, Python, ML Stack, & ML Basics Linear Algebra - vectors, matrices, eigenvalues, tensor products Probability & Statistics - distributions, expectation, variance Optimization - gradient descent, loss functions Python - NumPy, SciPy, Matplotlib PyTorch or TensorFlow, Scikit-learn Supervised, Unsupervised Learning Regression, Classification, Overfitting, Regularization Neural Networks, CNN basics Goal: You should be comfortable building classical ML pipelines Step 2: Quantum Computing Foundations Qubits, superposition, measurement, Bloch sphere Quantum gates, Entanglement and Bell states Quantum circuits, Interference Quantum Algorithms - Deutsch-Jozsa, Grover’s Algorithm, Quantum Fourier Transform, Variational Quantum Algorithms Qiskit, Cirq, Q#(1 of them) Goal: You must think in circuits before doing QML Step 3: Bridge to QML Parameterized Quantum Circuits Variational circuits Classical-quantum feedback loop Cost functions Barren plateaus Expressibility & trainability Difference between: Quantum data → quantum model Classical data → quantum embedding PennyLane, TensorFlow Quantum, Qiskit ML Goal: Understand QML is optimization on quantum parameters Step 4: Core QML Models Quantum Data Encoding Angle embedding Amplitude encoding Basis encoding Quantum Models Variational Quantum Classifier Quantum Neural Networks Quantum Kernel Methods Quantum Support Vector Machines Data re-uploading circuits Compare: Classical NN vs VQC Classical SVM vs Quantum Kernel Goal: Show measurable learning, not just circuit execution Step 5: Advanced QML Concepts Barren Plateaus Noise-aware training Hardware-efficient ansatz Quantum Convolutional Neural Networks Quantum Autoencoders QGANs QML for anomaly detection NISQ Constraints - Noise, Shot statistics, Error mitigation Goal: You understand real-world limitations and research gaps Step 6: Research Grade QML Read Papers Schuld & Killoran (Quantum ML theory) Havlíček et al. (Quantum kernel methods) McClean et al. (Barren plateaus) Cerezo et al. (Variational algorithms) Hybrid classical-quantum architectures Quantum kernels vs classical kernels Data-efficient QML Noise-resilient QML QML benchmarking 5–8 serious QML projects Implement: One paper reproduction One modification or improvement Happy Learning! Save this post for later. Repost ♻️ for Quantum & AI Learners! Check my profile for more resources on Quantum & AI Tech Follow Kiran Kaur Raina here: 📌LinkedIn: https://lnkd.in/gEpKMQ7z 📌YouTube: https://lnkd.in/gTTv2ewB 📌Topmate: https://lnkd.in/gDj-kmYW 📌Medium: https://lnkd.in/gWBppT7G 📌Instagram: https://lnkd.in/g8qZKHe7

#quantum + #SciML = #QuaSciML. 📄 Solving differential equations on quantum computers offers a great power, as we can represent solutions on an exponentially large and fine grid. At times it even 𝘵𝘰𝘰 powerful. Here’s a challenge: how do we even read out this vast amount of information?  In our recent work [https://lnkd.in/gW8ZSYqK], we explored this question and realized something simple yet exciting — 𝐐𝐃𝐢𝐟𝐟 𝐬𝐨𝐥𝐯𝐞𝐫𝐬, which generate solutions as quantum states, are a perfect source of 𝐪𝐮𝐚𝐧𝐭𝐮𝐦 𝐝𝐚𝐭𝐚. And to extract useful information from these states we can use specialized tools. This is where quantum scientific machine learning steps in, providing adaptive measurement operators to analyze solutions in the problem-specific way.  For example, in turbine modelling, you don’t need every tiny detail of the pressure curve. Instead, you care about actionable questions: Is the turbine faulty? What is the critical temperature for failure? For such tasks of classification and regression one can build decision boundaries by training readout operators based on few supplied labelled solutions. As a simple example, by examining computational fluid dynamics equations we classified shock waves and distinguished between turbulent and laminar flow regimes with high accuracy. Moreover, we introduced a dual quantum neural network (QNN) structure, demonstrating its efficiency in learning correlations between solutions and formulating hypotheses for these classifications.  Our findings show that analyzing quantum data from QDiff solvers could be a powerful mode of operation for QuaSciML, combining computational and learning advantages. Huge qudos to our collaboration Pasqal and Siemens Digital Industries Software, Chelsea Williams for doing hard work, and the team Daniel Berger Antonio Andrea Gentile Stefano Scali for the support. #quantumcomputing #machinelearning #quantumCFD

Joint work "Fermi-Dirac thermal measurements: A framework for quantum hypothesis testing and semidefinite optimization" with Nana Liu now available on arXiv: https://lnkd.in/gMn6aYZ3 Key contributions: Our paper introduces a novel model for quantum machine learning, called Fermi–Dirac machines, which consists of parameterized Fermi–Dirac thermal measurements whose parameters can be learned by means of gradient-based, hybrid quantum–classical algorithms. This presents an alternative paradigm to quantum Boltzmann machines, the latter being based on parameterized thermal states. The strong connection of Fermi–Dirac thermal measurements to sigmoid/logistic functions thus represents an alternative way for generalizing concepts from neural networks to quantum machine learning. Our approach opens up a new paradigm for solving semidefinite programs on quantum computers, alternative to the conventional approach relying on thermal state preparation. In support of this, we propose quantum algorithms for realizing Fermi–Dirac thermal measurements when given access to samples of states needed to form them. We apply all of our findings to quantum hypothesis testing and binary classification, thus providing a number of new insights for this domain of quantum information theory. In particular, we show how near-optimal measurements for hypothesis testing, of the Fermi–Dirac thermal form, can be learned via gradient ascent algorithms.