CV
Muhammad Rashid
PhD in Computer Science
Computer Vision | Explainable AI | Machine Learning | Visual Anomaly Detection
Turin, Italy
Website · Google Scholar · GitHub · LinkedIn
Professional Summary
I am a researcher in Computer Vision, Explainable AI (XAI), and Visual Anomaly Detection, with a focus on trustworthy AI systems for safety-critical industrial environments. My PhD research at the University of Torino and RuleX Innovation Labs focused on improving trust in black-box AI models through explainable visual attribution methods and robust anomaly detection frameworks for smart industries.
My work includes LIME Stratified, ShapBPT, explainable anomaly detection with VAE-GANs, and ADVIS, a real-time anomaly detection and visual intelligence system for human safety in collaborative robotics.
Research Metrics
| Source | h-index | Citations | Publications |
|---|---|---|---|
| Google Scholar | 9 | 788+ | 13 |
| Scopus | 7 | 539+ | 11 |
Research Areas
- Explainable Artificial Intelligence (XAI)
- Explainable Computer Vision
- Shapley-value-based explanations
- LIME, SHAP, ShapBPT
- Visual anomaly detection
- VAE-GAN and deep generative models
- Industrial AI and safety-critical systems
- Human–robot collaboration
- Medical image analysis
- Edge AI and model deployment
Experience
Research Fellow
University of Torino, Italy
Nov 2025 – Present
Working within the DistriMuSe project under the supervision of Prof. Elvio G. Amparore.
- Leading validation and optimization of anomaly detection pipelines for industrial robotics safety.
- Evaluating robustness on synthetic and real industrial datasets.
- Supporting integration of anomaly detection models into distributed robotics architectures.
- Improving ADVIS for real-time safety monitoring in collaborative robotic environments.
Visiting Doctoral Researcher
Valeria Lab, University of Granada, Spain
Jan 2025 – Aug 2025
- Completed an in-person research stay from 19 Jan 2025 to 19 Apr 2025, followed by remote collaboration until 31 Aug 2025.
- Worked on the DistriMuSe EU project, focusing on safe interaction with robots in smart industrial environments.
- Contributed to synthetic palletizing dataset development for Demo 3.2.
- Tested and validated the ADVIS anomaly detection framework on synthetic robotic scenarios.
- Supervisor: Prof. Jesús Garrido
Doctoral Researcher – R&D Projects
University of Torino & RuleX Innovation Labs, Italy
Nov 2022 – Oct 2025
- Conducted PhD research on trustworthy AI, explainable computer vision, and visual anomaly detection.
- Developed XAI methods including LIME Stratified and ShapBPT.
- Built explainable anomaly detection systems using VAE-GANs.
- Contributed to EU-funded projects including DistriMuSe and NextPerception.
- Completed the thesis:
Improving Trust in Safety-Critical AI Systems: Explainable AI and Anomaly Detection Frameworks for Human Safety in Smart Industries.
Research Assistant
HITEC University, Taxila, Pakistan
Aug 2021 – Jan 2023
- Designed and implemented machine learning and computer vision pipelines.
- Mentored undergraduate students on AI, computer vision, and data science projects.
- Supported research activities in medical imaging, surveillance, and pattern recognition.
Freelance Computer Vision & Machine Learning Developer
Upwork & Fiverr
2018 – 2023
- Delivered 10+ applied computer vision and machine learning projects.
- Developed GUI-based applications for object detection, classification, and image analysis.
- Worked on projects in medical imaging, surveillance, agriculture, and industrial inspection.
Participation in Research Projects
DistriMuSe — EU Project
Use Case 3: Safe Interaction with Robots
Role: Anomaly detection research, AI demonstrator development, and system integration
- Designed a real-time visual anomaly detection framework for collaborative robotics safety.
- Developed an end-to-end pipeline covering ROS streams, preprocessing, masking, safety-area segmentation, model training, threshold calibration, and real-time inference.
- Introduced localized anomaly detection across RoboArm, ConvBelt, PLeft, and PRight.
- Used ROS2 for critical safety messages and Zenoh for non-critical explainability streams.
- Developed GUI tools for anomaly timelines, reconstruction inspection, and system monitoring.
- Supported Demo 3.2 and Demo 3.3 with synthetic and real-world validation.
- Achieved 99.61% accuracy, 95.1% recall, 90.9% F1-score, and approximately 12.5 FPS real-time inference.
Repository: ADVIS DistriMuSe UC3
NextPerception — EU Project
Work Package 3: Distributed Intelligence
Role: Explainable AI research and demonstrator improvement
- Contributed to explainable AI methods for perception systems.
- Developed and evaluated improvements to LIME using stratified sampling.
- Improved explanation stability and coverage for high-dimensional image data.
Education
PhD in Computer Science
University of Torino, Italy
Nov 2022 – Nov 2025
Defended: 28 April 2026
Thesis: Improving Trust in Safety-Critical AI Systems: Explainable AI and Anomaly Detection Frameworks for Human Safety in Smart Industries
Supervisors: Prof. Elvio G. Amparore, Prof. Marco Botta, Dr. Enrico Ferrari
Research Focus: Explainable AI, Computer Vision, Visual Anomaly Detection, Industrial AI
Master of Science in Computer Science
COMSATS University Islamabad, Pakistan
2017 – 2019
- CGPA: 3.77/4.0
- Thesis: Object Detection and Classification Based on Feature Fusion and Deep Convolutional Neural Network
- Supervisor: Prof. Dr. Muhammad Sharif
- Research focus: object recognition, video surveillance, healthcare image analysis, feature fusion, and deep CNNs.
Bachelor of Science in Computer Science
Allama Iqbal Open University, Islamabad, Pakistan
2010 – 2016
- CGPA: 3.19/4.0
- Final Project: Online Venue Booking and Tour Planning
- Focus: secure web application development using CodeIgniter and MVC architecture.
Selected Publications
1. ShapBPT in Perspective: A Consolidated Review and an eXplainable Anomaly Detection Case Study
Muhammad Rashid, Elvio G. Amparore, Enrico Ferrari, Damiano Verda
QualITA Workshop @ ICPE 2026 (ACM)
2. ShapBPT: Image Feature Attributions using Data-Aware Binary Partition Trees
Muhammad Rashid, Elvio G. Amparore, Enrico Ferrari, Damiano Verda
AAAI Conference on Artificial Intelligence (AAAI 2026)
PDF arXiv Code Tests PyPI User Study Poster
3. Can I Trust My Anomaly Detection System? A Case Study Based on Explainable AI
Muhammad Rashid, Elvio Amparore, Enrico Ferrari, Damiano Verda
World Conference on Explainable Artificial Intelligence (XAI 2024)
4. Using Stratified Sampling to Improve LIME Image Explanations
Muhammad Rashid, Elvio G. Amparore, Enrico Ferrari, Damiano Verda
AAAI Conference on Artificial Intelligence (AAAI 2024)
PDF Code Examples PyPI Blog Details
Research Software
| Project | Description | Link |
|---|---|---|
| ShapBPT | Data-aware Shapley explanations using Binary Partition Trees | GitHub |
| ShapBPT Tests | Experimental evaluation of ShapBPT across vision tasks | GitHub |
| XAD | ShapBPT for explainable anomaly detection | GitHub |
| LIME Stratified | Improved LIME Image with stratified sampling | GitHub |
| LIME Stratified Examples | Experiments for LIME Stratified | GitHub |
| Explainable AD Case Study | VAE-GAN anomaly detection with XAI | GitHub |
| ADVIS DistriMuSe | Real-time anomaly detection for robotics safety | GitHub |
| AI on Edge Devices | AI deployment and optimization on Raspberry Pi | GitHub |
Teaching Activities
Teaching Collaboration
University of Torino, Italy
A.Y. 2023/2024
- Selected through a competitive departmental call for teaching support activities.
- Supported the course Sicurezza delle Reti e dei Sistemi.
- Contributed to exam sessions and student support activities.
Teaching Assistant
HITEC University, Pakistan
Sep 2019 – Jun 2021
- Assisted courses in Web Engineering, Digital Image Processing, Programming Fundamentals, and Data Structures and Algorithms.
- Supported laboratory sessions, assignments, and student evaluations.
- Mentored students on programming and computer vision projects.
Conferences and Presentations
- QualITA/ICPE 2026, Florence, Italy — Presented ShapBPT in Perspective.
- AAAI 2026, Singapore — Presented ShapBPT.
- XAI-World 2024, Valletta, Malta — Presented Can I Trust My Anomaly Detection System?
- AAAI 2024, Vancouver, Canada — Presented Using Stratified Sampling to Improve LIME Image Explanations.
- ECML-PKDD 2023, Turin, Italy — Attendee.
- icSoftComputing 2024, Remote — Attendee.
Academic Service
Program Committee Member
- ECML-PKDD 2026
- AAAI 2026
- ACDSA 2026
- ICLR 2025
- XAI-World 2026
- XAI-World 2025
- XAI-World 2024
- NLDB 2024
Journal Reviewer
- IEEE Transactions on Intelligent Transportation Systems
- Signal, Image and Video Processing, Springer
- Frontiers in Plant Science
Workshop Program Committee
- INSAIT Workshop @ ICIAP 2025
- DELTA Workshop @ ACM SIGKDD 2024
Research Network
- Confederation of Laboratories for AI Research in Europe (CLAIRE)
Awards and Scholarships
| Award | Organization | Year |
|---|---|---|
| Research Scholarship for DistriMuSe activities | University of Torino | 2025 |
| Erasmus+ Traineeship Scholarship | Erasmus+ | 2025 |
| Innovative Industrial Doctoral Scholarship | MUR / NRRP Italy | 2022 |
| National Laptop Award | Prime Minister Laptop Scheme | 2018 |
Technical Skills
Programming Languages
Python, MATLAB, C++, SQL, PHP
Frameworks and Libraries
PyTorch, TensorFlow, Keras, OpenCV, Scikit-learn, Scikit-image, NumPy, Pandas, Matplotlib, LIME, SHAP
Tools and Platforms
Git, LaTeX, ROS2, Zenoh, Raspberry Pi, Jupyter Notebook, Linux, Pixi
Domains
Machine Learning, Deep Learning, Computer Vision, Explainable AI, Visual Anomaly Detection, Generative AI, Edge AI, Distributed Intelligence, Robotics Safety
See details here - 📘 Courses & Training 🏅 Certifications
Languages
| Language | Level |
|---|---|
| 🇬🇧 English | C1 (Professional & Academic) |
| 🇵🇰 Urdu | Native |
| 🇮🇹 Italian | A1 (Basic Communication) |