My CV is available here.

Interests

Neuro-Symbolic AI • Knowledge-Aided ML • Deep Generative Models • Large Language Models

I build neuro-symbolic AI systems that inject background-knowledge requirements into neural networks during training and guarantee those requirements are satisfied. My focus is making these knowledge-aided frameworks practical for real-world use — synthesising tabular data, autonomous driving, and constraining large language models for theorem proving.

Education

• DPhil in Computer Science, University of Oxford (2021–2026)
  • Thesis: Knowledge-Refined Deep Neural Networks for Real-World Applications
  • Supervisor: Prof. Thomas Lukasiewicz
  • Areas: deep learning, generative modelling, autonomous driving, neuro-symbolic AI
• Master of Informatics with Honours, First Class, The University of Edinburgh (2014–2019)
  • Supervisor: Prof. Sharon Goldwater
  • Areas: NLP, speech-to-text machine translation, speech processing

Work Experience

• Research Associate at Imperial College London, UK (2025–present)
  • PI: Dr. Eleonora Giunchiglia
  • Autoformalisation: building methods to constrain language models so they translate informal mathematics into formal statements verifiable by proof assistants such as Lean.
  • Funded by Renaissance Philanthropy.
• Research Intern — Computer Vision & Autonomous Vehicles at FiveAI, Oxford, UK (2020–2021)
  • Supervisor: Dr. Tommaso Cavallari
  • 3D symmetry detection: designed a method that estimates planar reflective symmetries on 3D inputs by slicing along the height dimension and feeding slices to a 2D convolutional recurrent regressor, avoiding costly 3D convolutions.
  • Paper: Recurrently Estimating Reflective Symmetry Planes from Partial Pointclouds (CVPR 2021 Workshop on 3D Vision and Robotics).
  • Patent: 3D Perception, US 20240212189 A1 (Five AI Ltd, 2024; pending).
• Research Assistant — NLP & Machine Translation at The University of Edinburgh, UK (2019)
  • Supervisor: Prof. Sharon Goldwater
  • Speech-to-text translation: showed that ASR pretraining with target-language data augmentation improves translation performance, and that pretrained-ASR word error rate directly predicts translation quality.
  • Paper: Analyzing ASR pretraining for low-resource speech-to-text translation (Proc. of ICASSP 2020).
• Summer Research Fellow — Network Verification & Program Synthesis at ETH Zurich, Switzerland (2018)
  • Supervisors: Prof. Martin Vechev, Asst. Prof. Dana Drachsler Cohen
  • Program synthesis for P4: built a Python parser for P4 (Programming Protocol-Independent Packet Processors) and used the Z3 solver to encode and synthesise intended packet-processing behaviour.

Technical Skills

• Languages — Python (proficient); C++, Java, Bash/Shell (familiar)
• Deep learning — PyTorch, PyTorch Lightning, TensorFlow; Hugging Face, verl
• ML & data — scikit-learn, NumPy, SciPy, Pandas, Matplotlib, Seaborn; Jupyter
• Tooling — Linux/command line, Git, Docker, Slurm, Conda, Weights & Biases, TensorBoard, LaTeX

Software & Patents

• Software: PiShield — a PyTorch framework for integrating background-knowledge requirements into neural networks [github.com/mihaela-stoian/PiShield]
• Patent: 3D Perception, US 20240212189 A1 (Five AI Ltd, 2024; pending)

Awards

• Oxford PhD Runner-up Prize awarded by G-Research (2025)
• EPSRC Scholarship for Doctoral Studies awarded by University of Oxford (2021–2025)
• Women in Quant Finance Grant awarded by G-Research (2025)
• IJCAI Grant awarded by IJCAI-AIJ (2024)
• Conference Travel Grants awarded by St Hilda’s College, University of Oxford (2023, 2024)
• Best Paper Award at the AI4AD Workshop @ IJCAI (2022)

Selected Talks

• OxBridge Women in CS Conference, University of Oxford — selected for oral presentation (2026)
• Civil Service Leadership Group, Imperial College London — invitation-only (2025)
  • One of only four projects selected; presented to senior UK civil servants, incl. the Head of the Civil Service.
• Dagstuhl Seminar on Logic and Neural Networks — invitation-only (2025)
• IJCAI Doctoral Consortium (2024)
• SnT, University of Luxembourg — invited (2024)
  • Security, Reasoning and Validation reading group.
• Sony AI — invited (2024)
  • Barcelona and Tokyo reading groups.

Mentoring

• Tutorial: Beyond Soft Penalties: Hard Constraints for Neural Networks with Continuous Outputs — UAI 2026 (forthcoming)
  • Co-presented; introduces methods for enforcing hard constraints over continuous neural network outputs.
• Co-supervision:
  • A. Godun — MSc, TU Wien (current)
  • S. Lee — MSc, Imperial College London (2025)
  • L. Pejic — BSc, TU Wien (2025)
• Teaching — Class Tutor, University of Edinburgh (2017–2019)
  • Discrete Mathematics & Mathematical Reasoning
  • Algorithms, Data Structures & Learning
  • Processing Formal & Natural Languages

Service

• Workshop & challenge organiser
  • ROAD++: The Third Workshop & Challenge — ECCV 2024
  • ROAD-R: The Road Event Detection with Requirements Challenge — NeurIPS 2023
  • ROAD++: The Second Workshop & Challenge — ICCV 2023
• Reviewing
  • Conferences — NeurIPS, ICLR, ICML, IJCAI, NeSy
  • Journal — Machine Learning
  • Workshops — RepL4NLP @ ACL 2022, NeSy-GeMs @ ICLR 2023

Languages

English (proficient) • Romanian (native) • German (elementary)

Publications

Can I Have Your Order? Monte-Carlo Tree Search for Slot Filling Ordering in Diffusion Language Models

Joshua Ong Jun Leang, Yu Zhao, Mihaela C. Stoian, Wenda Li, Shay B. Cohen, Eleonora Giunchiglia. Can I Have Your Order? Monte-Carlo Tree Search for Slot Filling Ordering in Diffusion Language Models. In Proceedings of International Conference on Machine Learning (ICML), 2026.

A Survey on Deep Learning Approaches for Tabular Data Generation: Utility, Alignment, Fidelity, Privacy, Diversity, and Beyond

Mihaela C. Stoian, Eleonora Giunchiglia, Thomas Lukasiewicz. A Survey on Deep Learning Approaches for Tabular Data Generation: Utility, Alignment, Fidelity, Privacy, Diversity, and Beyond. Transactions on Machine Learning Research, 2026.

Beyond the Convexity Assumption: Realistic Tabular Data Generation under Quantifier-Free Real Linear Constraints

Mihaela C. Stoian and Eleonora Giunchiglia. Beyond the Convexity Assumption: Realistic Tabular Data Generation under Quantifier-Free Real Linear Constraints. In Proceedings of International Conference on Learning Representations (ICLR) 2025.

How Realistic Is Your Synthetic Data? Constraining Deep Generative Models for Tabular Data

Mihaela C. Stoian, Salijona Dyrmishi, Maxime Cordy, Thomas Lukasiewicz, Eleonora Giunchiglia. How Realistic Is Your Synthetic Data? Constraining Deep Generative Models for Tabular Data. In Proceedings of International Conference on Learning Representations (ICLR) 2024.

PiShield: A PyTorch Package for Learning with Requirements

Mihaela C. Stoian, Alex Tatomir, Thomas Lukasiewicz, Eleonora Giunchiglia. PiShield: A PyTorch Package for Learning with Requirements. In Proceedings of International Joint Conference on Artificial Intelligence (IJCAI), 2024.

Deep generative models as an adversarial attack strategy for tabular machine learning

Salijona Dyrmishi, Mihaela C. Stoian, Eleonora Giunchiglia, Maxime Cordy. Deep generative models as an adversarial attack strategy for tabular machine learning. In Proceedings of International Conference on Machine Learning and Cybernetics 2024.

CCN+: A neuro-symbolic framework for deep learning with requirements

Eleonora Giunchiglia‚ Alex Tatomir‚ Mihaela C. Stoian, Thomas Lukasiewicz. CCN+: A neuro-symbolic framework for deep learning with requirements. International Journal of Approximate Reasoning, 171, 109-124 (2024).

Deep Learning with Requirements in the Real World

Mihaela C. Stoian. Deep Learning with Requirements in the Real World. In Proceedings of International Joint Conference on Artificial Intelligence (IJCAI), 2024, Doctoral Consortium.

ROAD-Waymo: Action Awareness at Scale for Autonomous Driving

Salman Khan, Izzeddin Teeti, Reza Javanmard Alitappeh, Mihaela C. Stoian, Eleonora Giunchiglia, Gurkirt Singh, Andrew Bradley, Fabio Cuzzolin. ROAD-Waymo: Action Awareness at Scale for Autonomous Driving. arXiv preprint 2411.01683, 2024

Exploiting T-norms for Deep Learning in Autonomous Driving

Mihaela C. Stoian, Eleonora Giunchiglia, Thomas Lukasiewicz. Exploiting T-norms for Deep Learning in Autonomous Driving. In Proceedings of the 17th International Workshop on Neural-Symbolic Learning and Reasoning (NeSy), 2023.

ROAD-R: The Autonomous Driving Dataset with Logical Requirements

Eleonora Giunchiglia, Mihaela C. Stoian, Salman Khan, Fabio Cuzzolin, Thomas Lukasiewicz. ROAD-R: The Autonomous Driving Dataset with Logical Requirements. Machine Learning, 112, 3261–3291 (2023).

Deep Learning with Logical Constraints

Eleonora Giunchiglia, Mihaela C. Stoian, Thomas Lukasiewicz. Deep Learning with Logical Constraints. In Proceedings of International Joint Conference on Artificial Intelligence (IJCAI), 2022.

Recurrently Estimating Reflective Symmetry Planes from Partial Pointclouds

Mihaela C. Stoian, Tommaso Cavallari. Recurrently Estimating Reflective Symmetry Planes from Partial Pointclouds. In Conference on Computer Vision and Pattern Recognition (CVPR) Workshop on 3D Vision and Robotics, 2021.

Analyzing ASR pretraining for low-resource speech-to-text translation

Mihaela C. Stoian, Sameer Bansal, Sharon Goldwater. Analyzing ASR pretraining for low-resource speech-to-text translation. In Proceedings of International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2020.