Research Profile

I develop structure-aware and resource-efficient methods for quantum, hybrid quantum–classical, and deep learning systems. My work preserves tensor and low-rank structure across model design, measurement, optimization, and deployment, with the goal of reducing parameter, data-processing, and quantum-resource costs while retaining predictive performance. I evaluate methods through theory, reproducible simulation, and experiments on IBM Quantum hardware using PyTorch, Qiskit, and PennyLane.

Research Interests

My research starts from a simple frustration: most hybrid quantum–classical models throw away the structure already sitting in their data, flattening images, sequences, and quantum measurements into vectors before a single parameter is learned, and then spend a quantum circuit’s scarce qubits and shots relearning what was already there. In Quantum-Based Tensor Contraction Layers (QTCL, SPIE 2026), I addressed this by placing a small trainable quantum circuit inside a compressed tensor bottleneck rather than the full image, tying its cost to the bottleneck width instead of the raw input. The design beat its classical counterpart on VGG-19/CIFAR-100 and, just as usefully, lost to it on AlexNet—the kind of regime-dependent result that turns “does quantum help?” into the more answerable question of when it does. That same instinct carried into purely classical territory in Multilinear Transformation Layers (MTL, Asilomar 2026) and its adaptive-rank extension, LR-MTL (under review, IEEE TAI), where late convolutions are replaced by mode-wise transformations and effective rank is learned rather than fixed by hand, and into optimization itself in QALR-CPD, where I formulated adaptive CP-rank selection as an interaction-aware QUBO so that a quantum solver decides which tensor components actually matter, rather than pruning by magnitude alone.

A well-placed quantum circuit is still only as useful as what you choose to measure from it, and most of the measurement-efficiency literature treats that choice as fixed rather than as part of the model. In Parameter-Efficient Multilinear Readout for Quantum Reservoir Computing (ML-QRC, QML@QCE 2026), I preserved the time–qubit–observable structure of reservoir measurements instead of flattening them, cutting readout parameters by roughly 12.1× without losing predictive quality. In Variational Quantum Bayesian Regression (VQBR, IEEE QCE 2026), I rebuilt a classical regression problem around exactly what a shallow circuit can measure, separating the target vector into a direction recovered by the circuit and a scale solved for analytically, recovering the true MAP direction with 0.998 cosine similarity under realistic finite-shot noise.

The open question connecting both lines of work is where any of this should actually run. A QPU being available is not the same as it being worth using once queueing, compilation, shots, and mitigation are counted honestly. I am currently building a resource-aware execution policy, grounded in a systematic review of the quantum-HPC integration literature, that weighs a QPU route’s complete cost against classical and simulated alternatives and is explicitly allowed to say no. Across all three lines, the goal is the same: not to make more things quantum, but to know, with evidence, exactly when it is worth it.

Education

The University of Texas at San Antonio (UT San Antonio)

Ph.D. Student, Electrical Engineering · San Antonio, TX, USA
August 2023–Present

  • Advisor: Prof. Panagiotis P. Markopoulos
  • Dissertation: Structure-Aware and Resource-Efficient Quantum–Classical Learning: Multilinear Architectures, Adaptive Measurement, and Resource-Aware Execution
  • GPA: 3.93/4.00

Vietnam National University HCMC–University of Information Technology

B.S. (Honors), Computer Science · Ho Chi Minh City, Vietnam
August 2018–January 2022

  • Graduated with Excellence; ranked 1st of 26 students; final grade: 9.05/10
  • Thesis: Evaluating Deep Learning Methods for Vietnamese Infographic Visual Question Answering
  • Advisors: Tien Do Van, M.Sc., and Thanh Duc Ngo, Ph.D.

Publications and Manuscripts

Accepted / To Appear

  1. Van Tien Nguyen and Panagiotis P. Markopoulos, “VQBR: Variational Quantum Bayesian Regression via Measurement-Based MAP Direction Recovery,” 2026 IEEE International Conference on Quantum Computing and Engineering (QCE), Quantum Machine Learning Technical Papers Track, Toronto, Canada, September 2026. Accepted.
  2. Van Tien Nguyen and Panagiotis P. Markopoulos, “Parameter-Efficient Multilinear Readout for Quantum Reservoir Computing,” Workshop on Quantum Machine Learning at IEEE Quantum Week 2026 (QML@QCE), Toronto, Canada, September 2026. Accepted for speed presentation and poster; to appear in workshop proceedings.
  3. Van Tien Nguyen, Mayur Dhanaraj, and Panagiotis P. Markopoulos, “Depth- and Rank-Selective Multilinear Transformation Layers for Compact CNNs,” 60th Asilomar Conference on Signals, Systems, and Computers, Pacific Grove, CA, October 2026. Accepted.

Peer-Reviewed Publications

  1. Van Tien Nguyen and Panagiotis P. Markopoulos, “Quantum-Based Tensor Contraction Layers,” Machine Learning from Challenging Data 2026, Proc. SPIE, vol. 14030, paper 140300I, 2026. doi:10.1117/12.3098024
  2. M. Tran, T. Pham, V. T. Nguyen, T. Do, and T. N. Duc, “A Robust Framework for Mathematical Formula Detection,” 2021 International Conference on Multimedia Analysis and Pattern Recognition (MAPR), pp. 1–6, 2021. doi:10.1109/MAPR53640.2021.9585197

Preprints and Manuscripts in Revision

  1. Van Tien Nguyen and Panagiotis P. Markopoulos, “Quantum Tensor Decomposition: A Review of Foundations, Algorithms, Challenges, and Future Directions,” manuscript in preparation, 2026.
  2. Mayur Dhanaraj*, Van Tien Nguyen*, and Panagiotis P. Markopoulos, “Adaptive Low-Rank Multilinear Transformations for Compact Convolutional Neural Networks,” manuscript in revision for resubmission to IEEE Transactions on Artificial Intelligence, 2026. *Equal contribution; supported by NSF Award 2332744.
  3. Van Tien Nguyen and Panagiotis P. Markopoulos, “Quantum Optimization for Adaptive Low-Rank CP Decomposition,” manuscript in revision, 2026.

Research Experience

Graduate Research Assistant · MILOS Lab, UT San Antonio

August 2023–Present · San Antonio, TX

  • Develop tensor-structured and measurement-efficient methods for quantum and hybrid quantum–classical learning under limited qubits, finite shots, and noisy hardware.
  • Designed VQBR, which separates a Bayesian regression coefficient vector into direction and scale, estimates a Gram-matrix-free objective through overlap and Pauli measurements, and recovers scale analytically; validated on simulators and IBM Quantum hardware.
  • Developed Tucker-factored readouts for quantum reservoir computing, reducing trainable parameters by approximately 12.1× relative to an MLP readout and achieving compression ratios up to 52.24× across tested reservoir sizes.
  • Co-developed depth- and rank-selective multilinear layers for compact CNNs and evaluated parameter, computation, and predictive-performance tradeoffs on CIFAR-10, DOTA-v1.0, and ImageNet-1K.
  • Formulated interaction-aware adaptive tensor-rank selection as a QUBO problem for structured CP decomposition, benchmarked against exhaustive search and classical heuristics.
  • Leading a literature-grounded synthesis of quantum-HPC integration, middleware, and workflow-orchestration systems to motivate a learning-aware execution policy for routing subtasks across CPU, GPU, simulator, and QPU resources under complete end-to-end cost.
  • Investigate quantum-assisted CP/PARAFAC decomposition and objective-aligned ambiguity resolution for recovering deployable classical predictors from quantum-encoded states.
  • Build reproducible workflows in PyTorch, Qiskit, and PennyLane; support shared GPU/CPU infrastructure, software environments, and remote experimentation across the laboratory.

Research Assistant · AI Club, VNUHCMC-UIT

2020–2022 · Ho Chi Minh City, Vietnam

  • Conducted deep-learning research in computer vision and NLP, including infographic visual question answering, instance retrieval, mathematical-formula detection, and reinforcement-learning game agents.
  • Contributed to dataset construction, model training, evaluation, technical writing, and research presentations.

Industry Experience

Research Engineer · Viettel High Technology Industries Corporation

March 2022–March 2023 · Vietnam

  • Developed NLP systems for named-entity recognition, entity linking, summarization, and timeline generation using internally constructed news and social-media datasets.
  • Packaged models as containerized microservices and integrated them into production workflows using version control, CI/CD, and reproducible model-serving practices.

Teaching Experience

Incoming Graduate Teaching Assistant · UT San Antonio

August 2026–May 2027 · Department of Electrical Engineering

  • Appointed for the 2026–2027 academic year; assigned to EE 2423 Network Theory for Fall 2026.
  • Responsibilities include grading, office hours, student support, course logistics, and preparation of instructional materials.

Teaching Assistant · VNUHCMC-UIT

September 2021–January 2022

  • Supported tutorials, grading, office hours, and student programming exercises for Introduction to Programming in C++ and Python.

Patents and Inventions

  • Quantum-to-Classical Ambiguity Resolution (QCAR), UT San Antonio invention disclosure, 2026. Objective-aligned recovery of missing sign and scale information from quantum-encoded model parameters to produce an explicit classical predictor.
  • Detecting Ships in Distress Based on Optical Satellite Images and Automatic Identification System Signals, Vietnamese patent, January 2023.

Honors and Awards

  • NSF Student Travel Grant, IEEE Quantum Week 2026; up to $1,250 in reimbursable travel support to attend and present accepted work at the IEEE International Conference on Quantum Computing and Engineering (QCE 2026), Toronto, Canada, July 2026.
  • Selected for Phase III, qBraid–MITRE–JonesTrading Quantum Challenge / Global Industry Challenge 2026, June 2026.
  • UT San Antonio Institutional Nominee, 2026 Google PhD Fellowship–Quantum Computing; selected as one of four university nominees, April 2026.
  • Quantum Rising Star Award, inaugural Quantum Student Success Summit (QS3), Rice University, March 2026.
  • Graduate School Academic Travel Fund, UT San Antonio; $750 supporting presentation at SPIE Machine Learning from Challenging Data 2026, Spring 2026.
  • Student of 5 Merits, University of Information Technology, 2019–2021 (three consecutive years).
  • Consolation Prize, AI Camera Inspection Challenge, 2020.
  • Consolation Prize, National Mathematical Olympiad Competition, 2019.

Leadership, Certification, and Service

  • Qiskit Advocate, IBM Quantum / Qiskit community, August 2025–Present; advanced to Tier 1 in July 2026.
  • Organizer, Qiskit Fall Fest, The University of Texas at San Antonio, 2026.
  • IBM Certified Quantum Computation using Qiskit v2.X Developer–Associate, June 2026.
  • Mentor-in-Training, Qiskit Global Summer School 2026, IBM Quantum.
  • Presenter, SPIE Defense + Commercial Sensing, Machine Learning from Challenging Data, National Harbor, MD, April 2026.
  • Poster presenter and lightning-talk speaker, inaugural Quantum Student Success Summit, Rice University, March 2026.
  • Participant, Qiskit Global Summer School, 2024 and 2025; earned Quantum Excellence completion badges.
  • Participant, NSF Spring AI School, San Antonio, TX, 2024–2026.
  • Scribe, EXAIL Workshop, Pittsburgh, PA, October 2024.

Selected Additional Technical Projects

  • Transformer-Based Quantum Error Mitigation. Benchmarked unmitigated estimation, zero-noise extrapolation, a histogram MLP, and a Transformer encoder on raw shot sequences; a shuffle ablation showed that the Transformer learned a permutation-invariant representation.
  • NV-Center Quantum Sensing for Biomedicine. Simulated nanodiamond NV-center T1 relaxometry for mitochondrial free-radical conditions using amplitude damping, inversion recovery, T1 fitting, and standard-quantum-limit ensemble scaling.
  • Quantum Neural Networks Benchmark Study. Implemented variational quantum classifiers and compared them with classical baselines across multiple benchmark datasets.
  • HHL for Linear Regression. Implemented and analyzed HHL-based linear regression on quantum simulators, including numerical stability, conditioning, data-loading assumptions, and resource requirements.

Technical Skills

  • Programming / machine learning: Python, C/C++, PyTorch, TensorFlow, scikit-learn, NumPy, SciPy, snnTorch
  • Quantum computing: Qiskit, PennyLane, Cirq, CUDA-Q, IBM Quantum Runtime; variational algorithms and quantum simulation
  • Research computing: Linux, Git, GitHub, Docker, CUDA/GPU workflows, Jupyter, LaTeX, CI/CD