Author name cluster

Tuan Do

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

3 papers

1 author row

AAAI Conference 2026 Short Paper

Graph Neural ODEs with Stability and Conservation Guarantees for Tumor Microenvironment Dynamics (Student Abstract)

Luong Doan
Tien Nguyen
Nhung Duong
Lap Nguyen
Tuan Do

We present Graph Neural ODEs (GNODEs) for modeling tumor microenvironment dynamics with mathematically guaranteed stability and conservation properties. Unlike bulk ODEs that miss spatial heterogeneity or discrete GNNs that inadequately capture continuous biological processes, GNODEs provide continuous-time evolution with explicit adjacency-aware dynamics while maintaining provable trajectory bounds. Our framework ensures: (1) existence and uniqueness of solutions under dynamic graph topology, (2) Lyapunov stability preventing unphysical states like negative cell counts, and (3) exact conservation of biological invariants through architectural constraints. Benchmarking on synthetic tumor data demonstrates that GNODE accurately captures the dynamics of the resistant cell fraction (0.282 predicted vs 0.242 true), whereas graph-free alternatives fail completely (0.000), underscoring the importance of stability-constrained local interactions for modeling emergent resistance.

PDF Details DOI

AAMAS Conference 2026 Conference Paper

Identifying Essential Rule Sets in Agent-Based Models Through Systematic Ablation: A Tumor Evolution Case Study

Nhung Duong
Luong Doan
Anh Do
Anh Truong
Ngoc Do
Tien Nguyen
Tuan Do

We present a systematic methodology for identifying essential rule sets in agent-based models, demonstrated through tumor evolution under therapeutic pressure. Our framework addresses a critical challenge: determining which interaction rules are necessary for reproducing emergent spatiotemporal patterns versus those representing auxiliary complexity. We couple autonomous cancer cell agents with reaction-diffusion fields, then systematically ablate mechanisms including paracrine signaling, metabolic competition, phenotypic plasticity, and spatial interactions. Our approach establishes empirically-calibrated significance thresholds from baseline system stochasticity, providing principled criteria for mechanism classification. A key methodological insight distinguishes ablatable biological rules from non-ablatable framework requirements: removing spatial diffusion caused model failure, revealing it encodes physical constraints rather than testable hypotheses. We demonstrate that complex patterns emerge from minimal rule sets, with several commonly-modeled mechanisms contributing negligibly to tissue-scale behavior. This methodology advances multi-agent simulation by providing objective model reduction techniques that enhance mechanistic interpretability and facilitate parameter calibration.

PDF

AAMAS Conference 2026 Conference Paper

Information Contagion in Climate-Stressed SME Networks: An Agent-Based Simulation Study

Luong Doan
Uyen Nguyen
Thao Duong
Ngan Duong
Phong Ho
Nhung Duong
Tuan Do

This paper investigates how heterogeneous adoption of environmental accounting standards generates information asymmetries that propagate through interconnected small and medium enterprise networks, creating emergent systemic vulnerabilities. We develop an agent-based model simulating hundreds of SMEs across multiple sectors, where firms adapt to climate stress based on differential access to environmental risk information. Our simulation framework captures the co-evolution of climate impacts, adaptation investments, and information diffusion through supply chain and credit network layers. Through extensive experiments, we identify a critical "valley of vulnerability" phenomenon where partial adoption of climate risk assessment tools temporarily increases systemic fragility before improving resilience. Spectral analysis reveals characteristic oscillation patterns in cascade dynamics that dampen with increased adoption. Robustness checks confirm this valley phenomenon emerges only when information signals are sufficiently precise to create behavioral divergence between informed and uninformed agents. Our findings challenge assumptions about gradual technology diffusion and provide quantitative insights for understanding information externalities in complex economic systems.

PDF