CURRICULUM VITAE

Full CV

Education

• Deakin University, Australia. PhD. in Information Technology, 2017-2020
• Hanoi University of Science and Technology, Vietnam. Master in Computer Science (drop out), 2016-2017
• Hanoi University of Science and Technology, Vietnam. Bachelor in Information Technology-Center for the Gifted Student, 2010-2015
• Hanoi Amsterdam High School for The Gifted Student, Vietnam. Mathematics, graduate of 2010

Work experience

• Deakin University, Australia. Senior Lecturer, 2026-now
• Deakin University, Australia. Research Lecturer, 2022-2025
• Deakin University, Australia. Associate Research Fellow, 2019-2022
• Cinnamon AI, Vietnam. AI Researcher, 2017
• Viettel R&D, Vietnam. ML Engineer, 2015-2017

Research Statistics

Grants (5)

ARC Training Centre for AI Governance, Oversight and Assurance

Complex Time-series System Forecasting Reinforced by Expert Knowledge

Multi-Agent Collaborative Language Model Alignment

Complex Time-series System Forecasting Reinforced by Expert Knowledge

Elucidating Human Brain Connectivity Through Deep Learning and Network Analysis

Publications (75)

SteerCast: Retrieval-Based Latent Steering for Decoder-Only Time Series Forecasting

CausalPlan: Empowering Efficient LLM Multi-Agent Coordination Through Causality-Driven Planning

SIMPA: Restricting What Experts See, Not What They Produce, in MoE-PEFT

Continual Fine-Tuning of Large Language Models via Program Memory

Creative Generation via Multi-Agent Debate: Does Debate Suppress Diversity?

Beyond Tokenization: Direct Timestep Embedding and Contrastive Alignment for Time-Series Question Answering

Reviving Error Correction in Modern Deep Time-Series Forecasting

Enhancing Model Explainability and Performance via Attention-guided Gradient Restriction

Retrieval-augmented Decoding for Improving Truthfulness in Open-ended Generation

Rethinking Sparse Mixture of Experts from a Unified Perspective

Eigenvectors of Experts are Training-free Non-collapsing Routers

SPaCe: Unlocking Sample-Efficient Large Language Models Training With Self-Pace Curriculum Learning

Do Domain-specific Experts exist in MoE-based LLMs?

Spectral Text Fusion: A Frequency-Aware Approach to Multimodal Time-Series Forecasting

Probabilities Are All You Need: A Probability-Only Approach to Uncertainty Estimation in Large Language Models

Accelerating Long-Term Molecular Dynamics with Physics-Informed Time-Series Forecasting

Federated Domain Generalization with Latent Space Inversion

Reasoning Under 1 Billion: Memory-Augmented Reinforcement Learning for Large Language Models

Sample Efficient Alignment Learning With Episodic Control

On the Role of Discrete Representation in Sparse Mixture of Experts

DmC: Nearest Neighbor Guidance Diffusion Model for Offline Cross-domain Reinforcement Learning

Hybrid Cross-domain Robust Reinforcement Learning

Navigation in Brain Networks Is Possible Using Only Local Information

Dynamic Steering With Episodic Memory For Large Language Models

Physics-Informed Machine Learning for Predicting the Ballistic Limit of Whipple Shields

ALIAS: DAG Learning with Efficient Unconstrained Policies

Beyond the Known: Decision Making with Counterfactual Reasoning Decision Transformer

Navigating Social Dilemmas with LLM-based Agents via Consideration of Future Consequences

Lagr-seq: Language-guided Reinforcement Learning with Sample-Efficient Querying

Accelerated Experimental Design Using a Human-Ai Teaming Framework

Automatic Prompt Selection for Large Language Models

SuperRAG: Beyond RAG with Layout-Aware Graph Modeling

Stable Hadamard Memory: Revitalizing Memory-Augmented Agents for Reinforcement Learning

Rapid Selection and Ordering of In-Context Demonstrations via Prompt Embedding Clustering

SimSMoE: Toward Efficient Training Mixture of Experts via Solving Representational Collapse

Navigating Social Dilemmas with LLM-based Agents via Consideration of Future Consequences

Multi-Reference Preference Optimization for Large Language Models

Fair Domain Generalization with Heterogeneous Sensitive Attributes Across Domains

Plug, Play, and Generalize: Length Extrapolation with Pointer-Augmented Neural Memory

Large Language Model Prompting With Episodic Memory

Revisiting the Dataset Bias Problem from a Statistical Perspective

VRDSynth: Synthesizing Programs for Multilingual Visually Rich Document Information Extraction

Variable-Agnostic Causal Exploration for Reinforcement Learning

Diversifying Training Pool Predictability for Zero-shot Coordination: A Theory of Mind Approach

Beyond Surprise: Improving Exploration Through Surprise Novelty

Universal Graph Continual Learning

Balanced Q-learning: Combining the Influence of Optimistic and Pessimistic Targets

Improving Domain Generalization with Interpolation Robustness

The Application of Machine Learning in Micrometeoroid and Orbital Debris Impact Protection and Risk Assessment for Spacecraft

Social Motivation for Modelling Other Agents under Partial Observability in Decentralised Training

Improving Out-of-distribution Generalization with Indirection Representations

Memory-Augmented Theory of Mind Network

Learning to Constrain Policy Optimization with Virtual Trust Region

Functional Indirection Neural Estimator for Better Out-of-distribution Generalization

Momentum Adversarial Distillation: Handling Large Distribution Shifts in Data-Free Knowledge Distillation

Improving Document Image Understanding with Reinforcement Finetuning

HYCEDIS: HYbrid Confidence Engine for Deep Document Intelligence System

Towards Effective and Robust Neural Trojan Defenses via Input Filtering

Neurocoder: General-Purpose Computation Using Stored Neural Programs

Make The Most of Prior Data: A Solution for Interactive Text Summarization with Preference Feedback

Generative Pseudo-Inverse Memory

Learning Theory of Mind via Dynamic Traits Attribution

Episodic Policy Gradient Training

Model-Based Episodic Memory Induces Dynamic Hybrid Controls

Robust Deep Reinforcement Learning for Extractive Legal Summarization

From Deep Learning to Deep Reasoning

DeepProcess: Supporting Business Process Execution Using a MANN-based Recommender System

A New Representation of Successor Features for Transfer across Dissimilar Environments

Self-attentive Associative Memory

LODENet: A Holistic Approach to Offline Handwritten Chinese and Japanese Text Line Recognition

Neural Stored-Program Memory

Learning to Remember More with Less Memorization

Variational Memory Encoder-Decoder

Dual Memory Neural Computer for Asynchronous Two-view Sequential Learning

Dual Control Memory Augmented Neural Networks for Treatment Recommendations

Talks (15)

Uncertainty Estimation in LLMs and What It’s Good For

Reliable Machine Learning Models via Uncertainty-Guided Checkpoint Selection

Stable Hadamard Memory: Revitalizing Memory-Augmented Agents for Reinforcement Learning

How Different Memory Types Help Agents Learn and Adapt

AI Horizons – Conversations with Australia’s leading and emerging researchers

Memory in Autonomous Agents: From Reinforcement Learning to Large Language Models

Towards Safe and Controlled Large Language Models

Recent Advanced in LLM Finetuning and Uncertainty, Confidence and Hallucination in LLMs

Unlocking Exploration: Self-Motivated Agents Thrive on Memory-Driven Curiosity

Memory-Based Reinforcement Learning

Memory for Lean Reinforcement Learning

Neural machine reasoning

From deep learning to deep reasoning

Generalization for Good: A story on OCR evolution

Theory behind conversational agents

Teaching (14)

Principle Supervisor of Bao Loc Pham

Principle Supervisor of Manh Tien Anh Nguyen

Associate Supervisor of Hai Chan Nguyen

Principle Supervisor of Yafeng Wu

Associate Supervisor of Linh Le Pham Van

Principle Supervisor of Huu Hiep Nguyen

Principle Supervisor of Manh Nguyen

Associate Supervisor of Truong Giang Do

Principle Supervisor of Van Dai Do

Principle Supervisor of Minh Hoang Nguyen

Associate Supervisor of Kha Pham

Associate Supervisor of Ragja Palakkadavath

Associate Supervisor of Bao Duong Nguyen

Teaching Assistant