Research interest

My research focuses on Machine Learning and Cryptography, which I enjoy for their mathematical foundations.

In Machine Learning, I work on Contrastive Learning, Federated Learning, Object Detection, and Reinforcement Learning. 

In Cryptography, I am interested in classical topics like RSA, (perceptual) Hashing, and Homomorphic Encryption (not classical).

Publications

• Elderly Activity Recognition in the Wild: Results from the EAR Challenge (CV4Smalls @ WACV 2025)
  Anh-Kiet Duong

• Hierarchical Document Parsing via Large Margin Feature Matching and Heuristics (DocUI @ AAAI 2025)
  Anh-Kiet Duong

• Scalable Framework for Classifying AI-Generated Content Across Modalities (AAAI 2025)
  Anh-Kiet Duong, Petra Gomez-Krämer

• Addressing Out-of-Label Hazard Detection in Dashcam Videos: Insights from the COOOL Challenge (WACV 2025)
  Anh-Kiet Duong, Petra Gomez-Krämer

• Action recognition using temporal shift module and ensemble learning (ICPR 2024)
  Anh-Kiet Duong, Petra Gomez-Krämer

• FedShip: Federated Learning for Ship Detection from Multi-Source Satellite Images (GRSL)
  Anh-Kiet Duong, Tran-Vu La,   Hoàng-Ân Lê, and Minh-Tan Pham

• Leveraging Feature Communication in Federated Learning for Remote Sensing Image Classification (IGRSS 2024)
  Anh-Kiet Duong, Hoàng-Ân Lê, Minh-Tan Pham

• Federated Learning: Centralized and P2P for a Siamese Deep Learning Model for Diabetes Foot Ulcer Classification (Applied Sciences)
  Toofanee Mohammud Shaad All, Hamroun Mohamed, Dowlut Sabeena, Tamine Karim, Petit Vincent, Duong Anh Kiet, Sauveron Damien

• DFU-Helper: An Innovative Framework for Longitudinal Diabetic Foot Ulcer Diseases Evaluation Using Deep Learning  (Applied Sciences)
  Toofanee Mohammud Shaad All, Hamroun Mohamed, Dowlut Sabeena, Tamine Karim, Petit Vincent, Duong Anh Kiet, Sauveron Damien

• DFU-SIAM a Novel Diabetic Foot Ulcer Classification with Deep Learning (IEEE Access)
  Toofanee Mohammud Shaad All, Hamroun Mohamed, Dowlut Sabeena, Tamine Karim, Petit Vincent, Duong Anh Kiet, Sauveron Damien

• Large Margin Cotangent Loss for Deep Similarity Learning (ACOMPA)
  Duong Anh-Kiet, Hoang-Lan Nguyen, and Toan-Thinh Truong

• HCMUS at MediaEval 2020: Image-Text Fusion for Automatic News-Images Re-Matching (ACMMM)
  Nguyen-Quang Thuc, Nguyen Tuan-Duy H, Nguyen-Ho Thang-Long, Duong Anh-Kiet, Hoang-Xuan Nhat, Nguyen-Truong Vinh-Thuyen, Nguyen Hai-Dang, Tran Minh-Triet

Education

• Master's student, Cryptology and Information Security (CRYPTIS), University of Limoges.
  CRYPTIS-INFO
  Vingroup Scholarship Program.
  GPA: 16.2/20. Rank: 1st.

• Bachelor’s degree in Information Technology (Honor Program), University of Science, Vietnam National University – Ho Chi Minh City.
  Honors program.
  GPA: 8.64/10.0.

Experience

• Research Internship at IRISA, University Bretagne Sud, France.
  - Topic: Towards Privacy-preserving Federated Learning in Multi-Task Learning from Multiple Datasets.
  - Worked with Membership Inference Attack, Machine Unlearning, Federated Learning, YOLO, Pytorch.
  - Supervised by Dr. Hoang An Le and Prof. Minh Tan Pham.

• Research Internship at XLIM, University of Limoges, France.
  - Topic: SaNitization Algorithm under attacK ...ε. Development of white-box membership inference attacks against differentially-private classification models.
  - Supervised by Prof. Cristina Onete.

• Research Project Master 1 CRYPTIS INFO.
  - Topic: Federated Learning Peer-to-Peer. Using Federated Learning and Contrastive Learning.
  - Worked with Tensroflow, Federated Learning, Similar Learning, Metric Learning, Graph Neural Networks.
  - Supervised by Dr. Toofanee Shaad and Dr. Tamine Karim. 

• Junior Machine Learning Engineer at MoMo.
  - MoMo is one of the leading e-wallets in Vietnam.
  - Worked with Python, Kotlin, Java, Contrastive Learning, Mobile, credit score, and Google Cloud. 

• Software Engineer Intern at OPSWAT.
  - Build an intelligent system to detect text-based file types.
  - Build an intelligent system to classify pdf/docx files as benign/malicious.
  - Worked with C++, tree-based methods, and Explainable AI.
  - Supervised by Dr. Nhut Minh Ngo.

• Student Researcher at SELab Vietnam National University HCMC.
  - We won first in task NewsImage of MediaEval 2020. The task is about text-image matching.
  - Worked with Image Captioning, GAN network, BERT, doc2vec, NLP, and fuzzy matching.
  - Supervised by Dr. Truong Toan Thinh and Prof. Tran Minh Triet. 

Achievements

Artificial Intelligence

• First at Elderly Action Recognition Challenge at WACV2025

• First at AAAI-25 Visually-Rich Document(VRD-IU)

• First at ConCon Challenge - AAAI 2025

• 2nd at Defactify 4 - AAAI 2025
  - 2nd CT2: AI-Generated Text Detection.
  - 2nd CT2 : AI-Generated Image Detection.

• First at Challenge Of Out-Of-Label in Autonomous Driving - WACV 2025

• First at Multi-Modal Visual Pattern Recognition Challenge Track 3 - ICPR 2024
  - Track 3: Multi-Modal Action Recognition.
  - Multi-Modal Visual Pattern Recognition Workshop at ICPR 2024.

• 2nd at iBioHash 2024-FGVC11
  - A task of large-scale zero-shot fine-grained image hashing.
  - This competition is part of the Fine-Grained Visual Categorization FGVC11 workshop at the CVPR 2024.

• Google Landmark Challenge
  - Google Landmark Retrieval: 2021 Silver medal (30th/263).
  - Google Landmark Recognition: 2021 Silver medal (23th/383); 2020 Bronze (100/736).
  - There are more than 81K classes in this challenge, zero-shot on test set.
  - This competition is part of the Instance-Level Recognition workshop at ICCV 2021 and ECCV 2020.

• Silver medal in Kaggle Rock, Paper, Scissors (56th/1662)
  - Create an AI to play against others in many rounds of this classic game.
  - A strong AI can consistently beat predictable AI.

• Consolation Prize in Ho Chi Minh city AI Hackathon 2020.

• First in task NewsImage of MediaEval 2020.

Information Security

• Bronze NSUCRYPTO olympiad 2024.

• 404CTF (Individual 67th/3800).

• European Cyber Week (ECW) CTF 2023.
  - Individual: 11th/8000 (general) and 7th/8000 (student).
  - Team: 8th Final.

• European Cyber Week (ECW) CTF 2022.
  - Individual: 51th/8000 (general) and 40th/8000 (student).
  - Team: 13th Final.

• CTF Challenge Author: Designed challenges for HCMUS-VNU CTF competitions (2022–2024).

• Archived Vingroup Scholarship Program for M. S. in CRYPTIS.

• Consolation Prize in ASEAN Student Contest on Information Security 2021.

Programming Contest

• Google Code Jam Farewell Rounds: 250th/40.000.

• Fourth Hash Code Archive - Drone Delivery (4th/130).

• Codeforce Candidate Master.

• Consolation Prize in ICPC Vietnam National.
  - In the team to participate in the regional Asia Pacific ICPC.

• Top 20 winners Bengaluru Tech Summit Global Hackathon 2018.

Mathematics