Tuan-Duy H. Nguyen

Tuan-Duy H. Nguyen

ML + Data Engineer/Researcher

Biography

I graduated with highest honors (Excellent/Summa Cum Laude) from Vietnam National University Ho Chi Minh City - University of Science.

Previously, I was a Research Resident in machine learning robustness and interpretability at VinAI, where I also worked as an Engineering Resident focusing on automotive perception and user interfaces.

Under the supervision of amazing mentors, interdisciplinary experiences span implicit sensing, time-varying signal processing, knowledge discovery with graph neural networks, decentralized computing, and blockchain.

Looking forward, I want to develop trustworthy algorithms and devices that assist healthcare providers in enhancing the quality of care.

Interests
  • Trustworthy AI
  • Bioinformatics
  • Human-Computer Interaction
  • NeuroAI
Education
  • B.Sc., Advanced Program in Computer Science, 2021

    Vietnam National University Ho Chi Minh City - University of Science

Portfolio

A Vietnamese-English Neural Machine Translation System (INTERSPEECH 2022 Show & Tell)
Pre-trained VinAI Translate models vinai/vinai-translate-vi2en and vinai/vinai-translate-en2vi are state-of-the-art text translation models for Vietnamese-to-English and English-to-Vietnamese, respectively. Our demonstration system VinAI Translate employing these pre-trained models is available at: https://vinai-translate.vinai.io.
A Vietnamese-English Neural Machine Translation System (INTERSPEECH 2022 Show & Tell)
Robust Bayesian Recourse (UAI 2022)
We introduce a model-agnostic recourse that minimizes the posterior probability odds ratio along its min-max robust counterpart with the goal of hedging against future changes in the machine learning model parameters.
Robust Bayesian Recourse (UAI 2022)
Vietnamese Speech-based Question Answering over Car Manuals (DCAI@NeurIPS2021 and IUI 2022)
We develop QA-CarManual as a lightweight, real-time and interactive system that integrates state-of-the-art technologies in language and speech processing to (i) understand and interact with users via speech commands and (ii) automatically query the knowledge base and return answers in both forms of text and speech as well as visualization.
Vietnamese Speech-based Question Answering over Car Manuals (DCAI@NeurIPS2021 and IUI 2022)
SHREC 2021: Retrieval and classification of protein surfaces equipped with physical and chemical properties (CAG Vol. 99)
This paper presents the methods that have participated in the SHREC 2021 contest on retrieval and classification of protein surfaces on the basis of their geometry and physicochemical properties.
SHREC 2021: Retrieval and classification of protein surfaces equipped with physical and chemical properties (CAG Vol. 99)
SHREC 2021: Surface-based Protein Domains Retrieval (3DOR 2021)
This paper assesses the ability of five methods to retrieve similar protein surfaces, using either their shape only (3D meshes), or their shape and the electrostatic potential at their surface, an important surface property.
SHREC 2021: Surface-based Protein Domains Retrieval (3DOR 2021)
Surface-based protein domains retrieval methods from a SHREC2021 challenge (JMGM Vol. 111)
Five different groups participated in this contest using the shape-only dataset, and one group extended its pre-existing method to handle the electrostatic potential. Our comparative study reveals both the ability of the methods to detect related proteins and their difficulties to distinguish between highly related proteins.
Surface-based protein domains retrieval methods from a SHREC2021 challenge (JMGM Vol. 111)
Towards a Gratifying Interactive Modality for Smart Environments based on Ubiquitous Sensing (BSc. Thesis)
In this thesis, we set forth to extend this line of research for intuitive, effortless, and enjoyable computer interaction by employing a natural everyday-carry object – the smartwatch.
Towards a Gratifying Interactive Modality for Smart Environments based on Ubiquitous Sensing (BSc. Thesis)
HCMUS at MediaEval 2020: Image-Text Fusion for Automatic News-Images Re-Matching (MediaEval 2020)
We propose three multi-modal methods for mapping text and images of news articles to the shared space in order to perform efficient cross-retrieval.
HCMUS at MediaEval 2020: Image-Text Fusion for Automatic News-Images Re-Matching (MediaEval 2020)
Recognizing Families through Images with Pretrained Encoder (FG 2020)
We employ 3 methods, FaceNet, Siamese VGGFace, and a combination of FaceNet and VGG-Face models as feature extractors, to achieve the 9th standing for kinship verification and the 5th standing for kinship retrieval in the Recognizing Family in The Wild 2020 competition.
Recognizing Families through Images with Pretrained Encoder (FG 2020)
Towards a Robust WiFi-based Fall Detection with Adversarial Data Augmentation (CISS 2020)
This paper investigates a method of generalization through adversarial data augmentation.
Towards a Robust WiFi-based Fall Detection with Adversarial Data Augmentation (CISS 2020)