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 Research Fellow in Computer Science at the University of Torino specializing in Explainable Artificial Intelligence (XAI), Computer Vision, and Visual Anomaly Detection, with a focus on trustworthy and deployable AI for safety-critical and industrial environments.

My research spans explainable computer vision, Shapley-value-based attribution, normal-only anomaly detection, generative models, industrial robotics safety, and real-time AI systems. I have developed methods and systems including LIME Stratified, ShapBPT, XAD, and ADVIS, with research published at venues including AAAI, ACM ICPE/QualITA, and XAI.

Alongside my core research, I contribute to open-source machine learning software and collaborate on ongoing research in time-series and video anomaly detection, interpretable machine learning, cybersecurity, and trustworthy AI.

Last updated: 2026-05-30

➡️ See details on Google Scholar

Research Areas

• Explainable Artificial Intelligence (XAI) and Trustworthy AI
• Explainable Computer Vision and Feature Attribution
• Visual and Video Anomaly Detection
• Time-Series Anomaly Detection
• Shapley-value-based Explanations
• Generative Models and One-Class Learning
• Industrial AI and Smart Manufacturing
• Human–Robot Collaboration and Robotics Safety
• Edge AI and Real-Time Machine Learning
• Multimodal and Vision-Language AI

Experience

Research Fellow

University of Torino, Italy
Nov 2025 – Present

Working within the DistriMuSe EU project under the supervision of Prof. Elvio G. Amparore.

• Developing and validating visual anomaly detection systems for industrial robotics safety.
• Evaluating robustness across synthetic and real-world industrial environments.
• Developing RGB-based safety-area monitoring and normal-only anomaly detection pipelines.
• Supporting real-time integration using ROS2 and distributed robotics architectures.
• Investigating explainability, threshold calibration, and deployment of anomaly detection systems.

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. Supervised by:
  • Academic Supervisor: Prof. Elvio G. Amparore
    
  • Industrial Supervisor: Dr. Enrico Ferrari (RuleX Innovation Labs)
    

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

Independent / Upwork / Fiverr
2017 – 2023

• Delivered 10+ applied machine learning, computer vision, and image-analysis projects.
• Developed end-to-end solutions spanning data preprocessing, model development, evaluation, visualization, and GUI-based deployment.
• Worked across medical imaging, surveillance, agriculture, object recognition, and industrial image analysis.

Participation in Research Projects

DistriMuSe — EU Horizon Europe Project

Use Case 3: Safe Interaction and Cooperation with Robots

Role: Visual anomaly detection, AI demonstrator development, validation, and system integration

ADVIS-UniGra — Synthetic Industrial Environment

• Developed an RGB anomaly detection framework for collaborative robotics safety.
• Designed safety-area-specific anomaly detection for PLeft, PRight, RoboArm, and ConvBelt.
• Developed area-specific VAE-GAN models with normal-only training.
• Evaluated L1, L2, SSIM, RAVI, and spatially tolerant anomaly scoring.
• Achieved approximately 99.6% accuracy and 12.5 FPS inference.
• Contributed to validation activities and technical deliverables within DistriMuSe.
• Github Repository: ADVIS-UniGra

ADVIS-SR — Real Industrial Environment

• Extended ADVIS from simulation to a real robotic palletizing environment.
• Developed multi-area RGB monitoring for unexpected-condition detection.
• Integrated real-time inference with ROS2 and distributed communication.
• Developed visualization tools for anomaly timelines, reconstruction analysis, safety-area monitoring, and explainability.
• Evaluated the system on real multi-camera industrial data.
• Github Repository: ADVIS-SR

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.

Ongoing Research & Collaborations

Video Anomaly Detection

Developing a comprehensive study of the evolution of deep learning for video anomaly detection, from reconstruction- and prediction-based approaches to transformers, foundation models, and reasoning-oriented methods.

NIDS-SCNN — Explainable Network Intrusion Detection

Developing a lightweight CNN-based network intrusion detection framework using time-frequency representations and explainable AI for interpretable binary and multiclass attack classification.

Interpretable Machine Learning for Electrical Discharge Machining

Investigating hierarchical and interpretable machine learning for small-sample prediction of surface and tribological responses in advanced manufacturing.

Time-Series Anomaly Detection

Investigating deep learning and one-class approaches for time-series anomaly detection, including robust training, threshold calibration, and benchmark evaluation.

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

Authors: Muhammad Rashid, Elvio G. Amparore, Enrico Ferrari, Damiano Verda
QualITA Workshop @ ICPE 2026 (ACM)

PDF Code Details

2. ShapBPT: Image Feature Attributions using Data-Aware Binary Partition Trees

Authors: Muhammad Rashid, Elvio G. Amparore, Enrico Ferrari, Damiano Verda AAAI Conference on Artificial Intelligence (AAAI 2026)

PDF arXiv Code Tests PyPI User Study Poster Details

3. Can I Trust My Anomaly Detection System? A Case Study Based on Explainable AI

Authors: Muhammad Rashid, Elvio Amparore, Enrico Ferrari, Damiano Verda
World Conference on Explainable Artificial Intelligence (XAI 2024)

PDF Code Blog Details

4. Using Stratified Sampling to Improve LIME Image Explanations

Authors: Muhammad Rashid, Elvio G. Amparore, Enrico Ferrari, Damiano Verda
AAAI Conference on Artificial Intelligence (AAAI 2024)

PDF Code Examples PyPI Blog Slides Poster Details

See full publication list

Research Software

Selected Open-Source Contributions

• Xplique (DEEL) — Fixed tabular explanation colorbar/axes handling and added regression tests. Repository · PR #184

• Anomalib (Open Edge Platform) — Contributed PatchCore backbone documentation improvements and participated in discussions around explainability for visual anomaly detection. Repository · PR #3630 · Issue #1144

• Alibi Detect (Seldon) — Contributed a Keras 3 compatibility fix for anomaly/outlier detection functionality. Repository · PR #955

• FLAML (Microsoft) — Contributing native anomaly detection support, including the initial anomaly_detection task and Isolation Forest integration. Repository · PR #1567 · Issue #413

• Awesome Python for Data Science (Data-Centric AI Community) — Contributed an anomaly detection tutorial/resource to the community-maintained data science collection. Repository · PR #42 · Issue #28

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

• AAAI 2027
• 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

Technical Skills

Programming

Python, MATLAB, C++, SQL

Machine Learning & Deep Learning

PyTorch, TensorFlow/Keras, Scikit-learn, Torchvision, timm

Computer Vision & XAI

OpenCV, SHAP, LIME, Captum, Grad-CAM, Integrated Gradients, Shapley-value attribution, SAM/SAM2

Anomaly Detection

VAE/VAE-GAN, reconstruction-based anomaly detection, PatchCore, one-class learning, visual anomaly detection, time-series anomaly detection

Engineering & Deployment

Git/GitHub, Linux, ROS2, Zenoh, Docker, Raspberry Pi, HPC/SLURM, Jupyter, LaTeX

Data & Scientific Computing

NumPy, Pandas, SciPy, Matplotlib, Scikit-image

See details here - 📘 Courses & Training 🏅 Certifications

Languages