arXiv subjects

Wenhan Zhang

Publications and source records attributed to Wenhan Zhang.

At least 19 recordsLinked to original sources

arXiv2026

RareLens: Towards End-to-End Rare Disease Care via Aligning Divergent Large Language Model Reasoning

Rare diseases represent one of the most challenging settings for clinical decision-making, where heterogeneous presentations, sparse evidence and limited expertise create persistent uncertainty throughout the care pathway. Although artificial intelligence could help, existing systems largely address isolated tasks, particularly diagnosis, and usually rely on downstream investigations rather than information available at initial presentation. Here we show that clinical AI performance under uncertainty can be improved not by scaling a single model, but by exploiting the diversity of multiple imperfect reasoning systems. Across heterogeneous large language models, we identify divergent reasoning trajectories with complementary error patterns and develop RareLens, which learns to reconcile these perspectives into actionable decisions across four stages of rare disease care: risk screening, diagnosis, treatment planning and prognosis prediction. Built on RarelensBench, a real-world dataset of 157,525 cases spanning all 33 Orphanet categories and more than 7,000 conditions, RareLens outperformed every frontier model tested, including GPT-5, DeepSeek-R1, Claude-3.7-Sonnet and Gemini-2.5-Pro, across all stages. It achieved an area under the curve of 0.917 for screening and top-1 accuracies of 65.5% and 89.8% for diagnosis and treatment. In an external evaluation involving 1,287 cases and 23 physicians, autonomous RareLens and physicians assisted by RareLens both outperformed unaided physicians, while demonstrating that effective human-AI collaboration requires more than simply providing model outputs. These findings establish divergent model reasoning as an exploitable source of information and suggest a general strategy for building AI systems that operate reliably under high clinical uncertainty.

cs.AI↗

arXiv2026

TextRich: A Multi-Domain Benchmark for Detecting AI-Generated Text-Rich Images from GPT-Image-2

Text-rich images often contain privacy-sensitive, transactional, or decision-relevant information. As recent multimodal image generation models become increasingly capable of synthesizing realistic textual content and structured visual designs, detecting AI-generated text-rich images has become an important challenge for digital trust and content authenticity. Existing benchmarks, however, largely focus on object-centric images and provide limited coverage of scenarios where textual semantics and layout organization are central. In this paper, we introduce TextRich, a multi-domain benchmark for detecting text-rich images generated by OpenAI's GPT-Image-2. The benchmark contains 12,095 images across six representative categories: commercial posters, infographic charts, academic posters, receipts, tables, and UI screenshots. Using this benchmark, we evaluate five representative AI-generated image detectors under a zero-shot setting and further explore the capability of a multimodal vision-language model for this task. Our results reveal substantial performance variations across text-rich domains, where existing AI-generated image detectors exhibit distinct strengths and failure modes. Although the strongest detector achieves competitive overall performance, it remains ineffective on certain structured categories and highly sensitive to JPEG compression. Vision-language models provide a promising complementary approach, but still struggle with highly structured text-rich images. These findings highlight the need for text- and layout-aware detection methods for modern AI-generated images. Our dataset is released at https://huggingface.co/datasets/Shuyiww/TextRich.

cs.CV↗

arXiv2024

Prompt Tuning for Item Cold-start Recommendation

The item cold-start problem is crucial for online recommender systems, as the success of the cold-start phase determines whether items can transition into popular ones. Prompt learning, a powerful technique used in natural language processing (NLP) to address zero- or few-shot problems, has been adapted for recommender systems to tackle similar challenges. However, existing methods typically rely on content-based properties or text descriptions for prompting, which we argue may be suboptimal for cold-start recommendations due to 1) semantic gaps with recommender tasks, 2) model bias caused by warm-up items contribute most of the positive feedback to the model, which is the core of the cold-start problem that hinders the recommender quality on cold-start items. We propose to leverage high-value positive feedback, termed pinnacle feedback as prompt information, to simultaneously resolve the above two problems. We experimentally prove that compared to the content description proposed in existing works, the positive feedback is more suitable to serve as prompt information by bridging the semantic gaps. Besides, we propose item-wise personalized prompt networks to encode pinnaclce feedback to relieve the model bias by the positive feedback dominance problem. Extensive experiments on four real-world datasets demonstrate the superiority of our model over state-of-the-art methods. Moreover, PROMO has been successfully deployed on a popular short-video sharing platform, a billion-user scale commercial short-video application, achieving remarkable performance gains across various commercial metrics within cold-start scenarios

cs.IR↗

arXiv2024

Filtered Randomized Smoothing: A New Defense for Robust Modulation Classification

Deep Neural Network (DNN) based classifiers have recently been used for the modulation classification of RF signals. These classifiers have shown impressive performance gains relative to conventional methods, however, they are vulnerable to imperceptible (low-power) adversarial attacks. Some of the prominent defense approaches include adversarial training (AT) and randomized smoothing (RS). While AT increases robustness in general, it fails to provide resilience against previously unseen adaptive attacks. Other approaches, such as Randomized Smoothing (RS), which injects noise into the input, address this shortcoming by providing provable certified guarantees against arbitrary attacks, however, they tend to sacrifice accuracy. In this paper, we study the problem of designing robust DNN-based modulation classifiers that can provide provable defense against arbitrary attacks without significantly sacrificing accuracy. To this end, we first analyze the spectral content of commonly studied attacks on modulation classifiers for the benchmark RadioML dataset. We observe that spectral signatures of un-perturbed RF signals are highly localized, whereas attack signals tend to be spread out in frequency. To exploit this spectral heterogeneity, we propose Filtered Randomized Smoothing (FRS), a novel defense which combines spectral filtering together with randomized smoothing. FRS can be viewed as a strengthening of RS by leveraging the specificity (spectral Heterogeneity) inherent to the modulation classification problem. In addition to providing an approach to compute the certified accuracy of FRS, we also provide a comprehensive set of simulations on the RadioML dataset to show the effectiveness of FRS and show that it significantly outperforms existing defenses including AT and RS in terms of accuracy on both attacked and benign signals.

cs.LG↗

arXiv2024

Finite element method coupled with multiscale finite element method for the non-stationary Stokes-Darcy model

In this paper, we combine the multiscale flnite element method to propose an algorithm for solving the non-stationary Stokes-Darcy model, where the permeability coefflcient in the Darcy region exhibits multiscale characteristics. Our algorithm involves two steps: first, conducting the parallel computation of multiscale basis functions in the Darcy region. Second, based on these multiscale basis functions, we employ an implicitexplicit scheme to solve the Stokes-Darcy equations. One signiflcant feature of the algorithm is that it solves problems on relatively coarse grids, thus signiflcantly reducing computational costs. Moreover, under the same coarse grid size, it exhibits higher accuracy compared to standard flnite element method. Under the assumption that the permeability coefflcient is periodic and independent of time, this paper demonstrates the stability and convergence of the algorithm. Finally, the rationality and effectiveness of the algorithm are verifled through three numerical experiments, with experimental results consistent with theoretical analysis.

math.NA↗

arXiv2024

Multiscale finite element method for Stokes-Darcy model

This paper explores the application of the multiscale finite element method (MsFEM) to address steady-state Stokes-Darcy problems with BJS interface conditions in highly heterogeneous porous media. We assume the existence of multiscale features in the Darcy region and propose an algorithm for the multiscale Stokes-Darcy model. During the offline phase, we employ MsFEM to construct permeability-dependent offline bases for efficient coarse-grid simulation, with this process conducted in parallel to enhance its efficiency. In the online phase, we use the Robin-Robin algorithm to derive the model's solution. Subsequently, we conduct error analysis based on $L^2$ and $H^1$ norms, assuming certain periodic coefficients in the Darcy region. To validate our approach, we present extensive numerical tests on highly heterogeneous media, illustrating the results of the error analysis.

math.NA↗

arXiv2023

Testing the variants of the Stokes-Einstein relation in the framework of self-consistent generalized Langevin equation theory

The two functional forms, D~1/tau and D~T/tau, are usually adopted as the variants of the Stokes-Einstein relation; where D is the diffusion constant, tau the relaxation time and T the temperature. The self-consistent generalized Langevin equation (SCGLE) theory is presented as an analytical tool to predict the long time dynamics of colloids and molecular liquids. In this work, taking truncated Lennard-Jones-like liquids as an example, the rationality of the two variants were tested in the framework of the SCGLE theory. Our results indicate that D~1/tau is a good variant of the Stokes-Einstein relation in the framework of SCGLE theory; however, D~T/tau is not a good one but taking a fractional from as D~(T/tau)^zeta with an exponent zeta is not equal to 1.0 even the Stokes-Einstein relation is established in SCGLE theory.

cond-mat.soft↗

arXiv2023

Design and Implementation of A Soccer Ball Detection System with Multiple Cameras

The detection of small and medium-sized objects in three dimensions has always been a frontier exploration problem. This technology has a very wide application in sports analysis, games, virtual reality, human animation and other fields. The traditional three-dimensional small target detection technology has the disadvantages of high cost, low precision and inconvenience, so it is difficult to apply in practice. With the development of machine learning and deep learning, the technology of computer vision algorithms is becoming more mature. Creating an immersive media experience is considered to be a very important research work in sports. The main work is to explore and solve the problem of football detection under the multiple cameras, aiming at the research and implementation of the live broadcast system of football matches. Using multi cameras detects a target ball and determines its position in three dimension with the occlusion, motion, low illumination of the target object. This paper designed and implemented football detection system under multiple cameras for the detection and capture of targets in real-time matches. The main work mainly consists of three parts, football detector, single camera detection, and multi-cameras detection. The system used bundle adjustment to obtain the three-dimensional position of the target, and the GPU to accelerates data pre-processing and achieve accurate real-time capture of the target. By testing the system, it shows that the system can accurately detect and capture the moving targets in 3D. In addition, the solution in this paper is reusable for large-scale competitions, like basketball and soccer. The system framework can be well transplanted into other similar engineering project systems. It has been put into the market.

cs.CV↗

arXiv2019

Spectromicroscopic measurement of surface and bulk band structure interplay in a disordered topological insulator

Topological insulators are bulk semiconductors that manifest in-gap massless Dirac surface states due to the topological bulk-boundary correspondence principle [1-3]. These surface states have been a subject of tremendous ongoing interest, due both to their intrinsic properties and to higher order emergence phenomena that can be achieved by manipulating the interface environment [4-11]. Here, angle resolved photoemission (ARPES) spectromicroscopy and supplementary scanning tunneling microscopy (STM) are performed on the model topological insulator Bi2Se3 to investigate the interplay of crystallographic inhomogeneity with the topologically ordered bulk and surface band structure. Quantitative analysis methods are developed to obtain key spectroscopic information in spite of a limited dwell time on each measured point. Band energies are found to vary on the scale of 50 meV across the sample surface, enabling single-sample measurements that are analogous to a multi-sample doping series (termed a "binning series"). Focusing separately on the surface and bulk electrons reveals a nontrivial hybridization-like interplay between fluctuations in the surface and bulk state energetics.

cond-mat.mtrl-sci↗

arXiv2019

Squared English Word: A Method of Generating Glyph to Use Super Characters for Sentiment Analysis

The Super Characters method addresses sentiment analysis problems by first converting the input text into images and then applying 2D-CNN models to classify the sentiment. It achieves state of the art performance on many benchmark datasets. However, it is not as straightforward to apply in Latin languages as in Asian languages. Because the 2D-CNN model is designed to recognize two-dimensional images, it is better if the inputs are in the form of glyphs. In this paper, we propose SEW (Squared English Word) method generating a squared glyph for each English word by drawing Super Characters images of each English word at the alphabet level, combining the squared glyph together into a whole Super Characters image at the sentence level, and then applying the CNN model to classify the sentiment within the sentence. We applied the SEW method to Wikipedia dataset and obtained a 2.1% accuracy gain compared to the original Super Characters method. For multi-modal data with both structured tabular data and unstructured natural language text, the modified SEW method integrates the data into a single image and classifies sentiment with one unified CNN model.

cs.CL↗

arXiv2019

Intrinsic superstructure near atomically clean armchair-type step edge of graphite

We investigated the electronic superstructure of graphite surface in the vicinity to monoatomic armchair step edges with scanning tunneling microscopy and spectroscopy. Only the ($\sqrt{3}\times\sqrt{3}$)$\textit{R}$30$^{\circ}$ superstructure is visualized near atomically clean armchair edges, while the honeycomb superstructure is absent. The spectroscopic mapping near the clean armchair edge clearly reveals the ($\sqrt{3}\times\sqrt{3}$)$\textit{R}$30$^{\circ}$ superstructure on both sides of the step edge. We have also visualized a mixture of ($\sqrt{3}\times\sqrt{3}$)$\textit{R}$30$^{\circ}$ and honeycomb superstructures near structurally defective armchair edges. Our results suggest that the honeycomb superstructure pattern results from superposition of two sets of ($\sqrt{3}\times\sqrt{3}$)$\textit{R}$30$^{\circ}$ superstructure with different phases. Our observation solves the mystery of the coexistence of two types of superstructures reported by prior studies.

cond-mat.str-el↗

arXiv2019

SuperTML: Two-Dimensional Word Embedding for the Precognition on Structured Tabular Data

Tabular data is the most commonly used form of data in industry. Gradient Boosting Trees, Support Vector Machine, Random Forest, and Logistic Regression are typically used for classification tasks on tabular data. DNN models using categorical embeddings are also applied in this task, but all attempts thus far have used one-dimensional embeddings. The recent work of Super Characters method using two-dimensional word embeddings achieved the state of art result in text classification tasks, showcasing the promise of this new approach. In this paper, we propose the SuperTML method, which borrows the idea of Super Characters method and two-dimensional embeddings to address the problem of classification on tabular data. For each input of tabular data, the features are first projected into two-dimensional embeddings like an image, and then this image is fed into fine-tuned two-dimensional CNN models for classification. Experimental results have shown that the proposed SuperTML method had achieved state-of-the-art results on both large and small datasets.

cs.CV↗

arXiv2019

SuperCaptioning: Image Captioning Using Two-dimensional Word Embedding

Language and vision are processed as two different modal in current work for image captioning. However, recent work on Super Characters method shows the effectiveness of two-dimensional word embedding, which converts text classification problem into image classification problem. In this paper, we propose the SuperCaptioning method, which borrows the idea of two-dimensional word embedding from Super Characters method, and processes the information of language and vision together in one single CNN model. The experimental results on Flickr30k data shows the proposed method gives high quality image captions. An interactive demo is ready to show at the workshop.

cs.CL↗

arXiv2019

System Demo for Transfer Learning across Vision and Text using Domain Specific CNN Accelerator for On-Device NLP Applications

Power-efficient CNN Domain Specific Accelerator (CNN-DSA) chips are currently available for wide use in mobile devices. These chips are mainly used in computer vision applications. However, the recent work of Super Characters method for text classification and sentiment analysis tasks using two-dimensional CNN models has also achieved state-of-the-art results through the method of transfer learning from vision to text. In this paper, we implemented the text classification and sentiment analysis applications on mobile devices using CNN-DSA chips. Compact network representations using one-bit and three-bits precision for coefficients and five-bits for activations are used in the CNN-DSA chip with power consumption less than 300mW. For edge devices under memory and compute constraints, the network is further compressed by approximating the external Fully Connected (FC) layers within the CNN-DSA chip. At the workshop, we have two system demonstrations for NLP tasks. The first demo classifies the input English Wikipedia sentence into one of the 14 ontologies. The second demo classifies the Chinese online-shopping review into positive or negative.

cs.CL↗

arXiv2019

SuperChat: Dialogue Generation by Transfer Learning from Vision to Language using Two-dimensional Word Embedding and Pretrained ImageNet CNN Models

The recent work of Super Characters method using two-dimensional word embedding achieved state-of-the-art results in text classification tasks, showcasing the promise of this new approach. This paper borrows the idea of Super Characters method and two-dimensional embedding, and proposes a method of generating conversational response for open domain dialogues. The experimental results on a public dataset shows that the proposed SuperChat method generates high quality responses. An interactive demo is ready to show at the workshop.

cs.CL↗

arXiv2019

2-bit Model Compression of Deep Convolutional Neural Network on ASIC Engine for Image Retrieval

Image retrieval utilizes image descriptors to retrieve the most similar images to a given query image. Convolutional neural network (CNN) is becoming the dominant approach to extract image descriptors for image retrieval. For low-power hardware implementation of image retrieval, the drawback of CNN-based feature descriptor is that it requires hundreds of megabytes of storage. To address this problem, this paper applies deep model quantization and compression to CNN in ASIC chip for image retrieval. It is demonstrated that the CNN-based features descriptor can be extracted using as few as 2-bit weights quantization to deliver a similar performance as floating-point model for image retrieval. In addition, to implement CNN in ASIC, especially for large scale images, the limited buffer size of chips should be considered. To retrieve large scale images, we propose an improved pooling strategy, region nested invariance pooling (RNIP), which uses cropped sub-images for CNN. Testing results on chip show that integrating RNIP with the proposed 2-bit CNN model compression approach is capable of retrieving large scale images.

cs.AI↗

arXiv2018

MRAM Co-designed Processing-in-Memory CNN Accelerator for Mobile and IoT Applications

We designed a device for Convolution Neural Network applications with non-volatile MRAM memory and computing-in-memory co-designed architecture. It has been successfully fabricated using 22nm technology node CMOS Si process. More than 40MB MRAM density with 9.9TOPS/W are provided. It enables multiple models within one single chip for mobile and IoT device applications.

eess.SP↗

arXiv2018

Super Characters: A Conversion from Sentiment Classification to Image Classification

We propose a method named Super Characters for sentiment classification. This method converts the sentiment classification problem into image classification problem by projecting texts into images and then applying CNN models for classification. Text features are extracted automatically from the generated Super Characters images, hence there is no need of any explicit step of embedding the words or characters into numerical vector representations. Experimental results on large social media corpus show that the Super Characters method consistently outperforms other methods for sentiment classification and topic classification tasks on ten large social media datasets of millions of contents in four different languages, including Chinese, Japanese, Korean and English.

cs.CL↗