arXiv subjects

Xiang Qian

Publications and source records attributed to Xiang Qian.

9 recordsLinked to original sources

arXiv2026

FaithMed: Training LLMs For Faithful Evidence-Based Medical Reasoning

Faithful reasoning is essential in medicine, where clinical decisions require transparent justification grounded in reliable evidence. Current medical LLMs either lack active access to evidence or use retrieved evidence without supervising how it should be appraised and applied during reasoning. To address this, we formalize evidence-based medicine principles as process-level criteria and introduce FaithMed, a framework that combines clinician-designed, automatically refined rubrics with reinforcement learning using step-level process reward assignment and advantage grouping. Across seven medical benchmarks, FaithMed improves over agentic-search baselines (+9% on average) and outcome-only RL (+5.8%), while raising average evidence-based medicine rubric scores over agentic-search Qwen3 baselines (+15.5%). This work demonstrates that explicit step-level supervision can improve both task success and the faithfulness of the reasoning process. Code is available at https://github.com/cxcscmu/FaithMed.

cs.CL↗

arXiv2025

Scaling Electronic Health Record Foundation Models for Population Health Management

Population health management requires scalable methods to identify individuals at risk of chronic diseases such as cardiovascular conditions and cancer, yet existing approaches rely on fragmented data and resource-intensive screening. We present Scaling Electronic Health Record Foundation Models for Population Health Management, an Electronic Health Record Foundation Model that performs large-scale chronic disease prediction using cross-site longitudinal medical records. We pretrain Scaling Electronic Health Record Foundation Models for Population Health Management on billions of medical events from over 5 million patients across Taiwan and the United States, leveraging a unified code alignment framework to address cross-system heterogeneity, and characterize its scaling behavior via IsoFLOP analysis, training compute-optimal models up to 2.4B parameters. Across 11 chronic disease prediction tasks, Scaling Electronic Health Record Foundation Models for Population Health Management demonstrates strong scaling and generalization, outperforming tree-based models and both general and biomedical language models, achieving over 40% and 70% sensitivity at 99% specificity in U.S. and Taiwan cohorts, respectively. On the EHRShot benchmark, Scaling Electronic Health Record Foundation Models for Population Health Management surpasses prior EHR foundation models trained on in-site data despite substantial distribution shifts, highlighting strong few-shot generalization. Finally, we show that aligned cross-system data provides more effective pretraining signal than duplicating single-site data under data-limited settings, underscoring the importance of alignment for scalable healthcare modeling. Our analysis demonstrates the robustness of EHR-FM in various patient distributions and the benefits of operating in the ICD code space. The code will be open-sourced.

cs.LG↗

arXiv2024

FCA-RAC: First Cycle Annotated Repetitive Action Counting

Repetitive action counting quantifies the frequency of specific actions performed by individuals. However, existing action-counting datasets have limited action diversity, potentially hampering model performance on unseen actions. To address this issue, we propose a framework called First Cycle Annotated Repetitive Action Counting (FCA-RAC). This framework contains 4 parts: 1) a labeling technique that annotates each training video with the start and end of the first action cycle, along with the total action count. This technique enables the model to capture the correlation between the initial action cycle and subsequent actions; 2) an adaptive sampling strategy that maximizes action information retention by adjusting to the speed of the first annotated action cycle in videos; 3) a Multi-Temporal Granularity Convolution (MTGC) module, that leverages the muli-scale first action as a kernel to convolve across the entire video. This enables the model to capture action variations at different time scales within the video; 4) a strategy called Training Knowledge Augmentation (TKA) that exploits the annotated first action cycle information from the entire dataset. This allows the network to harness shared characteristics across actions effectively, thereby enhancing model performance and generalizability to unseen actions. Experimental results demonstrate that our approach achieves superior outcomes on RepCount-A and related datasets, highlighting the efficacy of our framework in improving model performance on seen and unseen actions. Our paper makes significant contributions to the field of action counting by addressing the limitations of existing datasets and proposing novel techniques for improving model generalizability.

cs.CV↗

arXiv2024

Enhancing Industrial Transfer Learning with Style Filter: Cost Reduction and Defect-Focus

Addressing the challenge of data scarcity in industrial domains, transfer learning emerges as a pivotal paradigm. This work introduces Style Filter, a tailored methodology for industrial contexts. By selectively filtering source domain data before knowledge transfer, Style Filter reduces the quantity of data while maintaining or even enhancing the performance of transfer learning strategy. Offering label-free operation, minimal reliance on prior knowledge, independence from specific models, and re-utilization, Style Filter is evaluated on authentic industrial datasets, highlighting its effectiveness when employed before conventional transfer strategies in the deep learning domain. The results underscore the effectiveness of Style Filter in real-world industrial applications.

cs.LG↗

arXiv2023

Underwater and Surface Aquatic Locomotion of Soft Biomimetic Robot Based on Bending Rolled Dielectric Elastomer Actuators

All-around, real-time navigation and sensing across the water environments by miniature soft robotics are promising, for their merits of small size, high agility and good compliance to the unstructured surroundings. In this paper, we propose and demonstrate a mantas-like soft aquatic robot which propels itself by flapping-fins using rolled dielectric elastomer actuators (DEAs) with bending motions. This robot exhibits fast-moving capabilities of swimming at 57mm/s or 1.25 body length per second (BL/s), skating on water surface at 64 mm/s (1.36 BL/s) and vertical ascending at 38mm/s (0.82 BL/s) at 1300 V, 17 Hz of the power supply. These results show the feasibility of adopting rolled DEAs for mesoscale aquatic robots with high motion performance in various water-related scenarios.

cs.RO↗

arXiv2023

Model-based Transfer Learning for Automatic Optical Inspection based on domain discrepancy

Transfer learning is a promising method for AOI applications since it can significantly shorten sample collection time and improve efficiency in today's smart manufacturing. However, related research enhanced the network models by applying TL without considering the domain similarity among datasets, the data long-tailedness of a source dataset, and mainly used linear transformations to mitigate the lack of samples. This research applies model-based TL via domain similarity to improve the overall performance and data augmentation in both target and source domains to enrich the data quality and reduce the imbalance. Given a group of source datasets from similar industrial processes, we define which group is the most related to the target through the domain discrepancy score and the number of samples each has. Then, we transfer the chosen pre-trained backbone weights to train and fine-tune the target network. Our research suggests increases in the F1 score and the PR curve up to 20% compared with TL using benchmark datasets.

cs.CV↗

arXiv2022

MIPR:Automatic Annotation of Medical Images with Pixel Rearrangement

Most of the state-of-the-art semantic segmentation reported in recent years is based on fully supervised deep learning in the medical domain. How?ever, the high-quality annotated datasets require intense labor and domain knowledge, consuming enormous time and cost. Previous works that adopt semi?supervised and unsupervised learning are proposed to address the lack of anno?tated data through assisted training with unlabeled data and achieve good perfor?mance. Still, these methods can not directly get the image annotation as doctors do. In this paper, inspired by self-training of semi-supervised learning, we pro?pose a novel approach to solve the lack of annotated data from another angle, called medical image pixel rearrangement (short in MIPR). The MIPR combines image-editing and pseudo-label technology to obtain labeled data. As the number of iterations increases, the edited image is similar to the original image, and the labeled result is similar to the doctor annotation. Therefore, the MIPR is to get labeled pairs of data directly from amounts of unlabled data with pixel rearrange?ment, which is implemented with a designed conditional Generative Adversarial Networks and a segmentation network. Experiments on the ISIC18 show that the effect of the data annotated by our method for segmentation task is is equal to or even better than that of doctors annotations

eess.IV↗

arXiv2018

Accurate evaluation of the fractal dimension based on a single morphological image

Fractal dimension (D) is an effective parameter to represent the irregularity and fragmental property of a self-affine surface, which is common in physical vapor deposited thin films. D could be evaluated through the scaling performance of surface roughness by using atomic force microscopy (AFM) measurements, but lots of AFM images with different scales (L) are needed. In this study, a surface roughness prediction (SRP) method was proposed to evaluate D values of a single AFM image, in which the roughness at smaller L was estimated by image segmentation with flatten modification. Firstly, a series of artificial fractal surfaces with ideal dimension (Di) values ranging from 2.1 to 2.9 were generated through Weierstrass-Mandelbrot (W-M) function, in order to compare SRP method with traditional methods such as box counting method and power spectral density method. The calculated dimension (Dc) by SRP method was much closer to Di than the other methods, with a mean relative error of only 0.64%. Secondly, SRP method was utilized to deal with real surfaces, which were AFM images of amorphous alumina thin films with L of 1-70 μm. Dc obtained by SRP method based on a single AFM image was also close to the result in our previous study by multi-image analysis at L above 10 μm, while the larger Dc at smaller L was consisted with the actual surface feature. The validity of SRP method and the physics nature of real surfaces were discussed, which might be helpful to obtain more understandings of fractal geometry.

cond-mat.mtrl-sci↗

arXiv2015

A progressively reduced pretension method to fabricate Bradbury-Nielsen gates with uniform tension

A Bradbury-Nielsen gate (BNG) is often used to modulate ion beams. It consists of two interleaved and electrically isolated sets of wires with uniform tension, which can keep parallel, equidistant, and coplanar over a wide temperature range, making the BNG reliable and robust. We have previously analyzed the non-uniform problem of wire tensions with sequentially winding method, and developed a template-based transfer method to solve this problem. In this paper, we introduced a progressively reduced pretension method, which allows directly and sequentially winding wires on the substrate without using a template. Theoretical analysis shows that by applying proper pretension to each wire when fixing it onto the substrate, the final wire tensions of all wires can be uniform. The algorithm and flowchart to calculate the pretension sequence are given, and the fabrication process is introduced in detail. Pretensions are generated by weights combination with a homebuilt weaving device. A BNG with stainless steel wire and a printed circuit board substrate is constructed with this method. The non-uniformity of the final wire tensions is less than 2.5% in theory. The BNG is successfully employed in our homemade ion mobility spectrometer, and the measured resolution is 33.5 at a gate opening time of 350 μs. Compared to the template-based method, this method is simpler, faster and more flexible when making BNGs with different configurations.

physics.ins-det↗