Anh N. Nhu

Modeling how diverse communities respond to public health messages is critical for training students to design targeted communication materials. However, no method is currently available to iteratively evaluate how different populations may perceive their campaign designs before human-subject testing. Despite progress in VLM/LLM-based psychometric simulation, community simulators for public health education remain largely unexplored. We reframe VLMs as attribute-aware community response simulators and introduce CLAAS, to our knowledge the first multimodal response simulation tool for public-health higher education. CLAAS aligns open-source VLMs, such as Gemma3, Llama-3.2, and Qwen2.5-VL, with an in-house survey dataset of 33K+ responses from 1K+ participants to simulate trait-conditioned emotional, behavioral, and sentimental responses. Quantitatively, attribute-aligned VLMs consistently outperform zero-shot and random baselines and generalize to unseen communication strategies. Beyond dataset evaluation, our pilot studies in real course sessions across two semesters demonstrate CLAAS’s potential practical relevance in educational settings. Most students reported CLAAS provided useful feedback and generated responses aligned with their expectations. This study suggests attribute-aware VLM support experiential public health education and promote a socially impactful research direction: interactive population simulator for community-centered communication design.

We introduce OpenVO, a novel framework for Open-world Visual Odometry (VO) with temporal awareness under limited input conditions. OpenVO effectively estimates real-world-scale ego-motion from monocular dashcam footage with varying observation rates and uncalibrated cameras, enabling robust trajectory dataset construction from rare driving events recorded in dashcam. Existing VO methods are trained on fixed observation frequency (e.g., 10Hz or 12Hz), completely overlooking temporal dynamics information. Many prior methods also require calibrated cameras with known intrinsic parameters. Consequently, their performance degrades when (1) deployed under unseen observation frequencies or (2) applied to uncalibrated cameras. These significantly limit their generalizability to many downstream tasks, such as extracting trajectories from dashcam footage. To address these challenges, OpenVO (1) explicitly encodes temporal dynamics information within a two-frame pose regression framework and (2) leverages 3D geometric priors derived from foundation models. We validate our method on three major autonomous-driving benchmarks - KITTI, nuScenes, and Argoverse 2 - achieving more than 20 performance improvement over state-of-the-art approaches. Under varying observation rate settings, our method is significantly more robust, achieving 46%-92% lower errors across all metrics. These results demonstrate the versatility of OpenVO for real-world 3D reconstruction and diverse downstream applications.

In this work, we introduce the Time-Aware World Model (TAWM), a model-based approach that explicitly incorporates temporal dynamics. By conditioning on the time-step size, ∆t, and training over a diverse range of ∆t values – rather than sampling at a fixed time-step – TAWM learns both high- and low-frequency task dynamics across diverse control problems. Grounded in the information-theoretic insight that the optimal sampling rate depends on a system’s underlying dynamics, this time-aware formulation improves both performance and data efficiency. Empirical evaluations show that TAWM consistently outperforms conventional models across varying observation rates in a variety of control tasks, using the same number of training samples and iterations. Our code can be found online at: github.com/anh-nn01/Time-Aware-World-Model.