arXiv subjects

Iris Xia

Publications and source records attributed to Iris Xia.

4 recordsLinked to original sources

arXiv2026

VisTa3D: A Dataset and Benchmark for Thin Object Reconstruction from Vision, Tactile, and 3D Point Clouds

State-of-the-art 3D reconstruction models, whether from visual, range, or both, tend to underperform on thin objects. This is partially due to the small amount of space such objects occupy in RGB images and in 3D point clouds. To test the extent of their errors, we collected the first thin object dataset comprising of synchronized RGB images, depth maps, and tactile response maps, where each frame is associated with inertial measurements, camera pose and calibration, and groundtruth depth and segmentation maps obtained from laser scanning of thin objects. We hypothesize that tactile data can aid in the reconstruction of thin objects as their response maps provide local shape and deformation information. Our dataset, termed VisTa3D, comprises of 387 scenes covering 70 thin objects over 17 environments. We benchmarked current 3D reconstruction models on VisTa3D and found that, indeed, they exhibit low fidelity on thin objects. To test if tactile data can help, we introduce the first visual-range-tactile 3D reconstruction model as a baseline. Code and data: https://huggingface.co/datasets/shaniaguo/VisTa3D.

cs.CV↗

arXiv2026

Medical thinking with multiple images

Large language models perform well on many medical QA benchmarks, but real clinical reasoning often requires integrating evidence across multiple images rather than interpreting a single view. We introduce MedThinkVQA, an expert-annotated benchmark for thinking with multiple images, where models must interpret each image, combine cross-view evidence, and answer diagnostic questions with intermediate supervision and step-level evaluation. The dataset contains 8,067 cases, including 720 test cases, with an average of 6.62 images per case, substantially denser than prior work, whose expert-level benchmarks use at most 1.43 images per case. On the test set, the best closed-source models, Claude-4.6-Opus, Gemini-3-Pro, and GPT-5.2-xhigh, reach only 57.2%, 55.3%, and 54.9% accuracy, while GPT-5-mini and GPT-5-nano reach 39.7% and 30.8%. Strong open-source models lag behind, led by Qwen3.5-397B-A17B at 52.2% and Qwen3.5-27B at 50.6%. Further analysis identifies grounded multi-image reasoning as the main bottleneck: models often fail to extract, align, and compose evidence across views before higher-level inference can help. Providing expert single-image cues and cross-image summaries improves performance, whereas replacing them with self-generated intermediates reduces accuracy. Step-level analysis shows that over 70% of errors arise from image reading and cross-view integration. Scaling results further show that additional inference-time computation helps only when visual grounding is already reliable; when early evidence extraction is weak, longer reasoning yields limited or unstable gains and can amplify misread cues. These results suggest that the key challenge is not reasoning length alone, but reliable mechanisms for grounding, aligning, and composing distributed evidence across real-world multimodal clinical inputs.

cs.CV↗

arXiv2025

From Scores to Steps: Diagnosing and Improving LLM Performance in Evidence-Based Medical Calculations

Large language models (LLMs) have demonstrated promising performance on medical benchmarks; however, their ability to perform medical calculations, a crucial aspect of clinical decision-making, remains underexplored and poorly evaluated. Existing benchmarks often assess only the final answer with a wide numerical tolerance, overlooking systematic reasoning failures and potentially causing serious clinical misjudgments. In this work, we revisit medical calculation evaluation with a stronger focus on clinical trustworthiness. First, we clean and restructure the MedCalc-Bench dataset and propose a new step-by-step evaluation pipeline that independently assesses formula selection, entity extraction, and arithmetic computation. Under this granular framework, the accuracy of GPT-4o drops from 62.7% to 43.6%, revealing errors masked by prior evaluations. Second, we introduce an automatic error analysis framework that generates structured attribution for each failure mode. Human evaluation confirms its alignment with expert judgment, enabling scalable and explainable diagnostics. Finally, we propose a modular agentic pipeline, MedRaC, that combines retrieval-augmented generation and Python-based code execution. Without any fine-tuning, MedRaC improves the accuracy of different LLMs from 16.35% up to 53.19%. Our work highlights the limitations of current benchmark practices and proposes a more clinically faithful methodology. By enabling transparent and transferable reasoning evaluation, we move closer to making LLM-based systems trustworthy for real-world medical applications.

cs.CL↗

arXiv2021

Anomalous Broad-Line Region Responses to Continuum Variability in Active Galactic Nuclei. I. H$β$ Variability

In the standard AGN reverberation-mapping model, variations in broad-line region (BLR) fluxes are predicted from optical continuum variability (taken as a proxy for the ionizing continuum) convolved with a response function that depends on the geometry. However, it has long been known that BLR variability can deviate from these predictions. We analyze both extensive long-term H$β$ and continuum monitoring of NGC 5548 and a large sample of high-quality H$β$ light curves of other AGNs to investigate the frequency and characteristics of anomalous responses of the BLR. We find that anomalies are very common and probably occur in every object. Onsets can be on a timescale only slightly longer than the light-crossing time and durations are of the order of the characteristic timescale of variability of the optical continuum to several times longer. Anomalies are larger when NGC 5548 is in a low state, but otherwise there is no correlation with continuum variability. There is abundant evidence for the optical continuum of AGNs varying independently of the higher-energy continua and this is sufficient to explain the anomalous responses of the total BLR flux. There are good reasons for believing that the frequent lack of correlation between spectral regions is due to anisotropic and non-axisymmetric emission. Rapid changes in line profiles and velocity-dependent lags are consistent with this. Motion of compact absorbing clouds across the line of sight is another possible cause of anomalies. The prevalence of anomalies should be considered when planning reverberation-mapping campaigns.

astro-ph.GA↗