Seismic inversion-including post-stack, pre-stack, and full waveform inversion is compute and memory-intensive. Recently, several approaches, including physics-informed machine learning, have been developed to address some of these limitations. Motivated by the potential of quantum computing, we report on our attempt to map one such classical physics-informed algorithm to a quantum framework. The primary goal is to investigate the technical challenges of this mapping, given that quantum algorithms rely on computing principles fundamentally different from those in classical computing. Quantum computers operate using qubits, which exploit superposition and entanglement, offering the potential to solve classically intractable problems. While current quantum hardware is limited, hybrid quantum-classical algorithms-particularly in quantum machine learning (QML)-demonstrate potential for near-term applications and can be readily simulated. We apply QML to subsurface imaging through the development of a hybrid quantum physics-informed neural network (HQ-PINN) for post-stack and pre-stack seismic inversion. The HQ-PINN architecture adopts an encoder-decoder structure: a hybrid quantum neural network encoder estimates P- and S-impedances from seismic data, while the decoder reconstructs seismic responses using geophysical relationships. Training is guided by minimizing the misfit between the input and reconstructed seismic traces. We systematically assess the impact of quantum layer design, differentiation strategies, and simulator backends on inversion performance. We demonstrate the efficacy of our approach through the inversion of both synthetic and the Sleipner field datasets. The HQ-PINN framework consistently yields accurate results, showcasing quantum computing's promise for geosciences and paving the way for future quantum-enhanced geophysical workflows.
Seismic inversion-including post-stack, pre-stack, and full waveform inversion is compute and memory-intensive. Recently, several approaches, including physics-informed machine learning, have been developed to address some of these limitations. Motivated by the potential of quantum computing, we report on our attempt to map one such classical physics-informed algorithm to a quantum framework. The primary goal is to investigate the technical challenges of this mapping, given that quantum algorithms rely on computing principles fundamentally different from those in classical computing. Quantum computers operate using qubits, which exploit superposition and entanglement, offering the potential to solve classically intractable problems. While current quantum hardware is limited, hybrid quantum-classical algorithms-particularly in quantum machine learning (QML)-demonstrate potential for near-term applications and can be readily simulated. We apply QML to subsurface imaging through the development of a hybrid quantum physics-informed neural network (HQ-PINN) for post-stack and pre-stack seismic inversion. The HQ-PINN architecture adopts an encoder-decoder structure: a hybrid quantum neural network encoder estimates P- and S-impedances from seismic data, while the decoder reconstructs seismic responses using geophysical relationships. Training is guided by minimizing the misfit between the input and reconstructed seismic traces. We systematically assess the impact of quantum layer design, differentiation strategies, and simulator backends on inversion performance. We demonstrate the efficacy of our approach through the inversion of both synthetic and the Sleipner field datasets. The HQ-PINN framework consistently yields accurate results, showcasing quantum computing's promise for geosciences and paving the way for future quantum-enhanced geophysical workflows.
Abstract ABSTRACT Quantum computing harnesses the principles of quantum mechanics to solve problems that are intractable for classical computers. Quantum annealing, a specialized approach within quantum computing, is particularly effective for optimization tasks, as it leverages quantum tunneling to escape local minima and efficiently explore complex energy landscapes. In geosciences, many problems are framed as high-dimensional optimization problems, including seismic inversion, which aims to estimate subsurface impedances from seismic data for accurate geological interpretation and resource exploration. A novel application of quantum computing for seismic inversion is presented, marking the first instance of inverting seismic data to estimate both P-wave and S-wave impedances using a quantum annealer. Building upon the prior work, which demonstrated the estimation of acoustic impedances from post-stack data using a two-step framework, we propose an enhanced workflow capable of inverting both post-stack and pre-stack seismic data in a single step. This advancement significantly reduces the number of qubits per model parameter (from 20 to 5) while improving computational speed (from 20 seconds to 6.3 seconds). The seismic inversion is implemented using the D-Wave Leap hybrid solver, achieving impedance estimation within 4-9 seconds, with the quantum processing unit (QPU) contributing just 0.043-0.085 seconds. Comparative analysis with simulated annealing reveals that quantum annealing produces impedance models closely matching true values in a single epoch, whereas simulated annealing requires 10 epochs for improved accuracy. These findings underscore the transformative potential of quantum computing for real-time, high-precision seismic inversion, marking a crucial step toward fully quantum-driven geophysical solutions.
Full waveform inversion (FWI) reconstructs heterogeneous material properties from receiver data but remains computationally demanding. Physics-informed neural networks (PINNs) and their domain-decomposed variants (FBPINNs) offer a mesh-free alternative but face convergence challenges when representing complex velocity fields. We present a hybrid quantum-classical FBPINN for acoustic FWI, bringing together quantum computing and classical machine learning, in which the decomposed wavefield network and the global velocity network are implemented as classical-to-quantum pipelines terminating in parameterized quantum circuits (PQCs). The PQCs are realized as differentiable JAX statevector simulators, enabling end-to-end automatic differentiation through the classical PINN, the quantum circuit, and the physics-informed loss. On a geophysical anomaly benchmark, the quantum hybrid reaches a lower L1 velocity error than the primary classical FBPINN baseline in approximately 8x fewer training iterations, despite using approximately 33% fewer trainable parameters, and it outperforms all 15 classical hyperparameter variants tested. A second benchmark (checkerboard) demonstrates the generality of the inversion pipeline, confirming that the quantum hybrid architecture can recover structured spatial variations beyond the localized anomaly benchmark. Our framework is broadly applicable to wave-based inverse problems beyond geophysics, including medical ultrasound tomography and non-destructive evaluation.
Quantum machine learning (QML) has emerged as a promising paradigm for scientific computing, particularly when combined with Physics-Informed Neural Networks (PINNs) to solve partial differential equations. In this work, we introduce a Quantum Physics-Informed Neural Network (QPINN) framework to solve the Helmholtz equation in the context for scattered-wavefield. To overcome the numerical instabilities associated with point-source singularities commonly encountered in classical PINNs, we adopt the Lippmann–Schwinger integral formulation. The proposed hybrid architecture couples a classical neural-network encoder with a variational quantum circuit, enabling the learning of expressive, dynamically generated feature embeddings. Numerical experiments on a heterogeneous velocity model demonstrate that the proposed QPINN achieves high solution accuracy while requiring significantly fewer trainable parameters than analogous classical neural network architectures. Our results highlight the potential of hybrid quantum-classical approaches as efficient and scalable alternatives for wavefield simulation and for solving wave-based inverse problems.
Abstract This paper presents a groundbreaking application of Quantum Machine Learning (QML) in geophysics, specifically targeting seismic fault detection. By integrating quantum computing with machine learning, the study addresses critical limitations of traditional fault detection methods, such as the detection of subtle features, processing efficiency, and scalability. The primary objective is to enhance fault delineation accuracy and provide a robust framework suitable for large-scale seismic surveys and real-time applications. The methodology involves encoding seismic data into quantum states to efficiently handle high-dimensional features. A hybrid quantum-classical model is developed, utilizing quantum-enhanced algorithms for fault detection. This model incorporates Quantum Long Short-Term Memory (QLSTM) networks, which leverage quantum parallelism and interference to enhance the processing capabilities of traditional LSTM networks. The QLSTM architecture employs quantum gates and memory cells to maintain and regulate the flow of quantum information, enabling more efficient and accurate processing of sequential data. The workflow encompasses quantum-based feature extraction, model training with seismic datasets, and validation using both synthetic and real-world examples. Comparative studies with traditional machine learning methods assess performance metrics, including accuracy, noise suppression, and computational efficiency. Quantum hardware simulators ensure the method's feasibility for practical applications. Results demonstrate that QML, powered by QLSTM, significantly outperforms conventional methods in seismic fault detection, achieving higher accuracy, better resolution of subtle geological features, and reduced noise sensitivity. Comprehensive testing on synthetic and field datasets reveals QML's ability to capture complex fault patterns, particularly in geologically challenging settings. Additionally, QML's computational efficiency enables rapid analysis of large datasets, facilitating timely decision-making. In conclusion, QML emerges as a superior approach for seismic fault detection, addressing longstanding challenges and offering novel opportunities for real-time seismic interpretation. By leveraging quantum computing, QML provides unprecedented insights into subsurface structures, leading to more accurate predictions and informed decisions in exploration and production operations. This paper introduces a transformative approach to seismic fault detection, contributing significantly to existing literature and offering actionable insights for the petroleum industry, particularly in subsurface imaging and reservoir characterization.
Quantum computing is increasingly recognized as a transformative technology in geophysics, offering the potential to significantly enhance computational power and efficiency. This advancement holds promise for the simulation and processing of complex geological data at unprecedented speeds. Recent studies have begun exploring the application of quantum computing methods to simplified versions of seismic inversion problems, underscoring the technology’s capability to address real-world inverse problems. The primary aim of this study is to tackle a realistic, scalable, and business-relevant problem by estimating acoustic impedances from seismic trace data using a quantum computer. To the best of our knowledge, this is the first instance of seismic impedances being predicted from seismic data through quantum computing, with a discussion on the advantages of solving the inverse problem on quantum processing units (QPUs). In this paper, we leverage the D-Wave quantum annealer to tackle the poststack seismic inversion problem, employing a novel two-step workflow. In the first step, we utilize the quantum annealer to estimate reflectivities from seismic data. Subsequently, these estimated normal-incidence reflectivities serve as the basis for predicting acoustic impedances with the same quantum technology. To validate the effectiveness of our approach, we present five examples, juxtaposing the impedance predictions from the D-Wave quantum annealer against those obtained through Simulated Annealing—a stochastic global optimizer traditionally employed for seismic inversion. Remarkably, the impedances derived from the quantum annealer matched closely with the true values in just a single epoch, whereas simulated annealing required 10 epochs to achieve a similar accuracy. Furthermore, the QPUs within our hybrid solver took only about 0.08 seconds to estimate these seismic impedances. This is remarkably efficient when compared to the time needed by both the classical component of the hybrid solver and simulated annealing, which each took over 10 seconds. This highlights the potential of QPUs to entirely resolve the seismic inverse problem in under a second, underscoring the transformative impact that quantum computing can have in the field of geophysics.
The computational demands of geoscientific workflows are rapidly increasing in both complexity and scale, pushing the limits of classical computing. Quantum computing has emerged as a promising paradigm to address some of these challenges. This paper surveys the growing intersection between quantum computing and energy geosciences, highlighting recent developments and initiatives. To illustrate the practical relevance of this emerging field, we present a case study on seismic traveltime inversion using a variational quantum circuit. In this example, a hybrid quantum-classical workflow is used to reconstruct velocity models from noisy traveltime data, demonstrating how quantum circuits can solve geophysical inverse problems. While modest in scale, the study offers insights into the formulation and implementation of quantumassisted inversion. As quantum hardware advances and algorithmic frameworks mature, quantum computing is poised to become a transformative tool in the computational landscape of energy geosciences.