arXiv subjects

Thao M. Dang

Publications and source records attributed to Thao M. Dang.

3 recordsLinked to original sources

arXiv2026

Who Holds the Pen? Let Specifications, Not Agents, Sign Off

Large language model agents increasingly combine generation, decision-making, execution, and self-evaluation within a single agentic loop. Although they operate under external specifications such as task instructions, guidelines, output schemas, and reusable skills, these specifications typically remain context for the same model that acts and declares completion, leaving no independent specification authority boundary. We identify two resulting gaps. The understanding--execution gap arises when a requirement is understood but not satisfied in execution; the state--authority gap arises when an agent's interpretation or completion claim does not establish the required state. On SkillsBench, using only agent-visible prompts, workspace information, and injected skill specifications, we extract 509 source-grounded task directions. Across seven models, only 79.6%--86.4% are satisfied, while completion-claim rates exceed official evaluator pass rates by 28.7--37.9 percentage points. We therefore separate agent proposals from authoritative state. Agents may plan, act, and request completion, but only admissible evidence from qualified providers may establish specification-governed state. SpecHarness operationalizes this principle by compiling visible specifications into source-linked obligations and governing execution and finalization through versioned obligation state. Verifiable requirements are mediated or validated at runtime, while ambiguous or subjective requirements remain advisory. Experiments on guideline-following and artifact-generation tasks show that specifications can serve not merely as behavioral guidance, but as authority over compliant execution and completion.

cs.AI↗

arXiv2026

Histopathology Multi-modal Embedding for Pathology Composed Retrieval

To overcome the black-box nature of predictive AI and the hallucination risks of generative models, retrieval-based models offer an interpretable, evidence-based paradigm for pathology clinical workflow. However, real-world clinical queries are inherently interleaved (e.g., pathology images and text). Current dual-encoders suffer from an \textbf{Architectural Mismatch}, lacking the mechanism to fuse such composed queries. To address this, we formalize the task of Pathology Composed Retrieval (PCR). While Multimodal Large Language Models (MLLMs) offer deep-fusion capabilities, directly applying them exposes a \textbf{Task Mismatch} and a \textbf{Domain Mismatch}. To resolve these challenges, we propose HOMIE, a model-agnostic adaptation framework that transforms any generative MLLM into a specialized pathology retrieval expert. Evaluated on our newly introduced PCR Benchmark, a lightweight 2B-parameter HOMIE variant substantially outperforms existing paradigms, surpassing specialized 7B pathology MLLMs and dual-encoders by large margins on composed retrieval, while maintaining strong performance on traditional simple retrieval. The project page is available at https://qfchou.github.io/HOMIE_page/.

cs.CV↗

arXiv2025

Text-Guided Multi-Instance Learning for Scoliosis Screening via Gait Video Analysis

Early-stage scoliosis is often difficult to detect, particularly in adolescents, where delayed diagnosis can lead to serious health issues. Traditional X-ray-based methods carry radiation risks and rely heavily on clinical expertise, limiting their use in large-scale screenings. To overcome these challenges, we propose a Text-Guided Multi-Instance Learning Network (TG-MILNet) for non-invasive scoliosis detection using gait videos. To handle temporal misalignment in gait sequences, we employ Dynamic Time Warping (DTW) clustering to segment videos into key gait phases. To focus on the most relevant diagnostic features, we introduce an Inter-Bag Temporal Attention (IBTA) mechanism that highlights critical gait phases. Recognizing the difficulty in identifying borderline cases, we design a Boundary-Aware Model (BAM) to improve sensitivity to subtle spinal deviations. Additionally, we incorporate textual guidance from domain experts and large language models (LLM) to enhance feature representation and improve model interpretability. Experiments on the large-scale Scoliosis1K gait dataset show that TG-MILNet achieves state-of-the-art performance, particularly excelling in handling class imbalance and accurately detecting challenging borderline cases. The code is available at https://github.com/lhqqq/TG-MILNet

cs.CV↗