PortfolioPortfolio
Portrait of Anh Dong Nguyen

Anh Dong Nguyen

Master of Research student, Ho Chi Minh City University of Technology (HCMUT – VNU-HCM)

I am a Master of Research student in Industrial Engineering at Ho Chi Minh City University of Technology (HCMUT – VNU-HCM), a collaborative research assistant with Prof. Mo El-Haj at VinUniversity, and in parallel a second-degree student in Electronics and Telecommunications at the Posts and Telecommunications Institute of Technology.

My research asks how systems that perceive the world should act when their perception is uncertain. I work on uncertainty-aware perception for video, structured episodic memory for long-horizon video and language, and graph learning under sparse supervision, and I build interactive systems that run on these ideas. I am moving toward wearable and egocentric interaction, where a system has to decide from noisy signals whether to act, ask or wait.

I am looking for PhD positions in human–computer interaction and computer vision.

Publications

* equal contribution (co-first authors). Newest first.

  1. COLINGFigure 1: retrieval rules on one EPBench question, comparing Hard-AND intersection with graded per-axis scoring

    What Makes Structured Episodic Retrieval Robust to Noisy Query Cues? A Controlled Study

    Nhan Thanh Ho*, Anh Dong Nguyen*, Thuan Duc Tran*, Mo El-Haj

    Under review at ACL Rolling Review (COLING 2027), 2026

    Abstract
    Structured memories for LLM agents index events by time, place, entity and topic, and answer a query by intersecting the matches on each axis. Intersection fails when an LLM extracts the events and the query cues, because one weak axis removes the right memory. We run a controlled study of retrieval rules on EPBench and LoCoMo. Hard intersection returns nothing for 63% of EPBench questions, and adding up the per-axis matches instead recovers most of the loss. How best to add them up depends on what the answer does with the retrieved set: routing questions to a selective rule recovers the retrieval loss. On LoCoMo, matching the speaker raises answer accuracy over dense retrieval by 7.9 points. All results carry paired bootstrap intervals, and our conclusions are limited to these two benchmarks.
    Project page PDF
  2. ACCVFigure 1: video observations become structured events with calibrated per-attribute confidences, retrieved by confidence-weighted matching

    Probabilistic Structured Event Modeling for Video Event Recognition under Perceptual Uncertainty

    Nhan Thanh Ho*, Anh Dong Nguyen*, Thuan Duc Tran*, Thang Van Nguyen, Bui Anh Thinh, Nguyen Bao Tran, Long Nhat Ho

    Asian Conference on Computer Vision (ACCV), 2026

    Abstract
    Long-duration video understanding systems summarize continuous observations as structured event representations to enable semantic retrieval. However, these events are produced by imperfect perceptual pipelines, meaning that every extracted attribute carries inherent uncertainty. This paper models structured events as probabilistic entities in which the correctness of each perceptual attribute is treated as a latent variable. A marginal-likelihood retrieval formulation integrates attribute-level uncertainty, allowing reliable attributes to compensate for uncertain ones instead of discarding an event after a single erroneous observation. For the spatial axis, detection covariance is propagated through the homography Jacobian to estimate uncertainty on the ground plane. Evaluated on an uncertainty-driven event retrieval benchmark using the indoor CAVIAR and outdoor VIRAT datasets, the framework consistently outperforms deterministic hard matching under increasing attribute noise, reaching 3.3× higher Recall@5 (0.394 vs. 0.120) at the highest corruption level.
    Project page PDF
    Bib
    @inproceedings{ho2026probabilistic,
      title     = {Probabilistic Structured Event Modeling for Video Event
                   Recognition under Perceptual Uncertainty},
      author    = {Ho, Nhan Thanh and Nguyen, Anh Dong and Tran, Thuan Duc and
                   Nguyen, Thang Van and {Bui Anh Thinh} and {Nguyen Bao Tran} and
                   Ho, Long Nhat},
      booktitle = {Asian Conference on Computer Vision (ACCV)},
      year      = {2026}
    }
  3. ICISNFigure 1: the 24H-GNN pipeline, from multimodal data and pseudo-labels to graph learning and hotspot ranking

    24H-GNN: A Pattern-Driven Pseudo-Label Fusion Graph Network for City-Wide Traffic Flow Prediction

    Thanh Nhan Ho*, Dong Anh Nguyen*, Minh Tai Pham Nguyen, Minh Khue Phan Tran, Hoang Minh Nguyen, Nguyen Bao Tran, Trong Nhan Le

    International Conference on Intelligent Systems and Networks (ICISN), Lecture Notes in Networks and Systems, Springer, 2026

    Abstract
    Urban traffic flow prediction is essential for effective transportation management but remains challenging due to heterogeneous and spatially inconsistent data sources. We propose a pseudo-label fusion framework that integrates loop detector data and taxi GPS trajectories using spatio-temporal similarity. Graph neural networks (GNNs) are employed to model network dependencies, while an adaptive fusion strategy balances reliable sensor data and trajectory-based estimates. The proposed approach achieves strong hotspot detection performance (hotspot@50 = 0.92, hotspot@100 = 0.97) and consistently outperforms baseline models, particularly under limited supervision.
    Project page PDF
    Bib
    @inproceedings{ho2026hgnn,
      title     = {24H-GNN: A Pattern-Driven Pseudo-Label Fusion Graph Network
                   for City-Wide Traffic Flow Prediction},
      author    = {Ho, Thanh Nhan and Nguyen, Dong Anh and Pham Nguyen, Minh Tai and
                   Phan Tran, Minh Khue and Nguyen, Hoang Minh and
                   {Nguyen Bao Tran} and Le, Trong Nhan},
      booktitle = {International Conference on Intelligent Systems and Networks (ICISN)},
      series    = {Lecture Notes in Networks and Systems},
      publisher = {Springer},
      year      = {2026}
    }

Projects

Personal projects with working demos and public code. Each has a write-up on my portfolio.

The Iris Look-to-Pay web app verifying an enrolled iris before a payment

Iris Look-to-Pay · 2023

Iris verification for a look-to-pay interaction on smart glasses: HOG features, PCA and cosine matching with an explicit threshold, and refusals that explain themselves. Ported to the browser and checked against the Python original.

The SeatScout live demo: detections on a café video, seat states and a staff alert

SeatScout · 2024

A live seat map from the cameras a café already has. YOLO11n runs on-device in the browser; a per-seat state machine (Free, Taken, Away, Check) survives missed detections and guests who step away. I designed the computer-vision pipeline and the seat rules.

The Cafe Finder app: an OpenStreetMap map of Saigon with ranked cafés and the evidence behind each label

Cafe Finder · 2025

Personalised search over 1,441 real cafés in Ho Chi Minh City from OpenStreetMap. A bilingual labeller turns what people say into labels with evidence and confidence; ranking treats unknown as unknown, not as no. I designed the personalisation.

CV

Research experience

Education

Honours and certificates

Skills

Programming
Python, C++, Objective-C, TypeScript
ML and vision
PyTorch, TensorFlow, TensorRT, ONNX Runtime, OpenCV, Ultralytics YOLO, graph neural networks
Hardware
Arduino, Raspberry Pi, ESP32, IMU and optical sensors, BLE, prototyping
Languages
Vietnamese (native), English (C1; IELTS 7.0)