arXiv subjects

Tianyang Wang

Publications and source records attributed to Tianyang Wang.

At least 127 records · Page 7Linked to original sources

arXiv2024

Multi-dimension Transformer with Attention-based Filtering for Medical Image Segmentation

The accurate segmentation of medical images is crucial for diagnosing and treating diseases. Recent studies demonstrate that vision transformer-based methods have significantly improved performance in medical image segmentation, primarily due to their superior ability to establish global relationships among features and adaptability to various inputs. However, these methods struggle with the low signal-to-noise ratio inherent to medical images. Additionally, the effective utilization of channel and spatial information, which are essential for medical image segmentation, is limited by the representation capacity of self-attention. To address these challenges, we propose a multi-dimension transformer with attention-based filtering (MDT-AF), which redesigns the patch embedding and self-attention mechanism for medical image segmentation. MDT-AF incorporates an attention-based feature filtering mechanism into the patch embedding blocks and employs a coarse-to-fine process to mitigate the impact of low signal-to-noise ratio. To better capture complex structures in medical images, MDT-AF extends the self-attention mechanism to incorporate spatial and channel dimensions, enriching feature representation. Moreover, we introduce an interaction mechanism to improve the feature aggregation between spatial and channel dimensions. Experimental results on three public medical image segmentation benchmarks show that MDT-AF achieves state-of-the-art (SOTA) performance.

cs.CV↗

arXiv2024

Characterisation and simulation of stitched CMOS strip sensors

In high-energy physics, there is a need to investigate alternative silicon sensor concepts that offer cost-efficient, large-area coverage. Sensors based on CMOS imaging technology present such a silicon sensor concept for tracking detectors. The CMOS Strips project investigates passive CMOS strip sensors fabricated by LFoundry in a 150nm technology. By employing the technique of stitching, two different strip sensor formats have been realised. The sensor performance is characterised based on measurements at the DESY II Test Beam Facility. The sensor response was simulated utilising Monte Carlo methods and electric fields provided by TCAD device simulations. This study shows that employing the stitching technique does not affect the hit detection efficiency. A first look at the electric field within the sensor and its impact on generated charge carriers is being discussed.

physics.ins-det↗

arXiv2024

Privacy-Enhancing Collaborative Information Sharing through Federated Learning -- A Case of the Insurance Industry

The report demonstrates the benefits (in terms of improved claims loss modeling) of harnessing the value of Federated Learning (FL) to learn a single model across multiple insurance industry datasets without requiring the datasets themselves to be shared from one company to another. The application of FL addresses two of the most pressing concerns: limited data volume and data variety, which are caused by privacy concerns, the rarity of claim events, the lack of informative rating factors, etc.. During each round of FL, collaborators compute improvements on the model using their local private data, and these insights are combined to update a global model. Such aggregation of insights allows for an increase to the effectiveness in forecasting claims losses compared to models individually trained at each collaborator. Critically, this approach enables machine learning collaboration without the need for raw data to leave the compute infrastructure of each respective data owner. Additionally, the open-source framework, OpenFL, that is used in our experiments is designed so that it can be run using confidential computing as well as with additional algorithmic protections against leakage of information via the shared model updates. In such a way, FL is implemented as a privacy-enhancing collaborative learning technique that addresses the challenges posed by the sensitivity and privacy of data in traditional machine learning solutions. This paper's application of FL can also be expanded to other areas including fraud detection, catastrophe modeling, etc., that have a similar need to incorporate data privacy into machine learning collaborations. Our framework and empirical results provide a foundation for future collaborations among insurers, regulators, academic researchers, and InsurTech experts.

cs.LG↗

arXiv2024

Deep Active Learning with Noise Stability

Uncertainty estimation for unlabeled data is crucial to active learning. With a deep neural network employed as the backbone model, the data selection process is highly challenging due to the potential over-confidence of the model inference. Existing methods resort to special learning fashions (e.g. adversarial) or auxiliary models to address this challenge. This tends to result in complex and inefficient pipelines, which would render the methods impractical. In this work, we propose a novel algorithm that leverages noise stability to estimate data uncertainty. The key idea is to measure the output derivation from the original observation when the model parameters are randomly perturbed by noise. We provide theoretical analyses by leveraging the small Gaussian noise theory and demonstrate that our method favors a subset with large and diverse gradients. Our method is generally applicable in various tasks, including computer vision, natural language processing, and structural data analysis. It achieves competitive performance compared against state-of-the-art active learning baselines.

cs.LG↗

arXiv2023

DeepArt: A Benchmark to Advance Fidelity Research in AI-Generated Content

This paper explores the image synthesis capabilities of GPT-4, a leading multi-modal large language model. We establish a benchmark for evaluating the fidelity of texture features in images generated by GPT-4, comprising manually painted pictures and their AI-generated counterparts. The contributions of this study are threefold: First, we provide an in-depth analysis of the fidelity of image synthesis features based on GPT-4, marking the first such study on this state-of-the-art model. Second, the quantitative and qualitative experiments fully reveals the limitations of the GPT-4 model in image synthesis. Third, we have compiled a unique benchmark of manual drawings and corresponding GPT-4-generated images, introducing a new task to advance fidelity research in AI-generated content (AIGC). The dataset is available at: \url{https://github.com/rickwang28574/DeepArt}.

cs.CV↗

arXiv2023

Towards Inadequately Pre-trained Models in Transfer Learning

Pre-training has been a popular learning paradigm in deep learning era, especially in annotation-insufficient scenario. Better ImageNet pre-trained models have been demonstrated, from the perspective of architecture, by previous research to have better transferability to downstream tasks. However, in this paper, we found that during the same pre-training process, models at middle epochs, which is inadequately pre-trained, can outperform fully trained models when used as feature extractors (FE), while the fine-tuning (FT) performance still grows with the source performance. This reveals that there is not a solid positive correlation between top-1 accuracy on ImageNet and the transferring result on target data. Based on the contradictory phenomenon between FE and FT that better feature extractor fails to be fine-tuned better accordingly, we conduct comprehensive analyses on features before softmax layer to provide insightful explanations. Our discoveries suggest that, during pre-training, models tend to first learn spectral components corresponding to large singular values and the residual components contribute more when fine-tuning.

cs.CV↗

arXiv2023

Comprehensive investigation of Quantum Oscillations in Semimetal Using an ac Composite Magnetoelectric Technique with Ultrahigh Sensitivity

Quantum oscillation (QO), a physical phenomenon that reflects the characteristics of the Fermi surface and transport fermions, has been extensively observed in metals and semimetals through various approaches, like magnetostriction, magnetization, resistivity, and thermoelectric power. However, only some allowed oscillation frequencies can be revealed by each individual method, particularly in semimetals with intricate Fermi pockets and associated magnetic breakdown phenomena. In this paper, we present the application of an ac composite magnetoelectric (ME) technique to measure the QOs of a topological nodal-line semimetal, ZrSiS, which possesses six fundamental QO frequencies. By employing the ME technique with a maximum magnetic field of 13 T and a minimum temperature of 2 K, we are able to capture all the fundamental frequencies and most of the permissible magnetic breakdown frequencies. In comparison, some of the frequencies were missing in the aforementioned four methods under identical measurement conditions. Remarkably, a series of magnetic breakdown frequencies around 8000 T were revealed even in a magnetic field as low as 7.5 T. These findings highlight the ME technique as an ultrahigh-sensitive tool for studying Dirac Fermions and other topological semimetals with complex Fermi surfaces.

cond-mat.str-el↗

arXiv2023

DenseMP: Unsupervised Dense Pre-training for Few-shot Medical Image Segmentation

Few-shot medical image semantic segmentation is of paramount importance in the domain of medical image analysis. However, existing methodologies grapple with the challenge of data scarcity during the training phase, leading to over-fitting. To mitigate this issue, we introduce a novel Unsupervised Dense Few-shot Medical Image Segmentation Model Training Pipeline (DenseMP) that capitalizes on unsupervised dense pre-training. DenseMP is composed of two distinct stages: (1) segmentation-aware dense contrastive pre-training, and (2) few-shot-aware superpixel guided dense pre-training. These stages collaboratively yield a pre-trained initial model specifically designed for few-shot medical image segmentation, which can subsequently be fine-tuned on the target dataset. Our proposed pipeline significantly enhances the performance of the widely recognized few-shot segmentation model, PA-Net, achieving state-of-the-art results on the Abd-CT and Abd-MRI datasets. Code will be released after acceptance.

cs.CV↗

arXiv2022

Temporal Output Discrepancy for Loss Estimation-based Active Learning

While deep learning succeeds in a wide range of tasks, it highly depends on the massive collection of annotated data which is expensive and time-consuming. To lower the cost of data annotation, active learning has been proposed to interactively query an oracle to annotate a small proportion of informative samples in an unlabeled dataset. Inspired by the fact that the samples with higher loss are usually more informative to the model than the samples with lower loss, in this paper we present a novel deep active learning approach that queries the oracle for data annotation when the unlabeled sample is believed to incorporate high loss. The core of our approach is a measurement Temporal Output Discrepancy (TOD) that estimates the sample loss by evaluating the discrepancy of outputs given by models at different optimization steps. Our theoretical investigation shows that TOD lower-bounds the accumulated sample loss thus it can be used to select informative unlabeled samples. On basis of TOD, we further develop an effective unlabeled data sampling strategy as well as an unsupervised learning criterion for active learning. Due to the simplicity of TOD, our methods are efficient, flexible, and task-agnostic. Extensive experimental results demonstrate that our approach achieves superior performances than the state-of-the-art active learning methods on image classification and semantic segmentation tasks. In addition, we show that TOD can be utilized to select the best model of potentially the highest testing accuracy from a pool of candidate models.

cs.CV↗

arXiv2022

Characterization of Passive CMOS Strip Sensors

Recent advances in CMOS imaging sensor technology , e.g. in CMOS pixel sensors, have proven that the CMOS process is radiation tolerant enough to cope with certain radiation levels required for tracking layers in hadron collider experiments. With the ever-increasing area covered by silicon tracking detectors cost effective alternatives to the current silicon sensors and more integrated designs are desirable. This article describes results obtained from laboratory measurements of silicon strip sensors produced in a passive p-CMOS process. Electrical characterization and charge collection measurements with a 90Sr source and a laser with infrared wavelength showed no effect of the stitching process on the performance of the sensor.

physics.ins-det↗

arXiv2022

Imbalanced Malware Images Classification: a CNN based Approach

Deep convolutional neural networks (CNNs) can be applied to malware binary detection via image classification. The performance, however, is degraded due to the imbalance of malware families (classes). To mitigate this issue, we propose a simple yet effective weighted softmax loss which can be employed as the final layer of deep CNNs. The original softmax loss is weighted, and the weight value can be determined according to class size. A scaling parameter is also included in computing the weight. Proper selection of this parameter is studied and an empirical option is suggested. The weighted loss aims at alleviating the impact of data imbalance in an end-to-end learning fashion. To validate the efficacy, we deploy the proposed weighted loss in a pre-trained deep CNN model and fine-tune it to achieve promising results on malware images classification. Extensive experiments also demonstrate that the new loss function can well fit other typical CNNs, yielding an improved classification performance.

cs.CV↗

arXiv2022

Parameter-Free Style Projection for Arbitrary Style Transfer

Arbitrary image style transfer is a challenging task which aims to stylize a content image conditioned on arbitrary style images. In this task the feature-level content-style transformation plays a vital role for proper fusion of features. Existing feature transformation algorithms often suffer from loss of content or style details, non-natural stroke patterns, and unstable training. To mitigate these issues, this paper proposes a new feature-level style transformation technique, named Style Projection, for parameter-free, fast, and effective content-style transformation. This paper further presents a real-time feed-forward model to leverage Style Projection for arbitrary image style transfer, which includes a regularization term for matching the semantics between input contents and stylized outputs. Extensive qualitative analysis, quantitative evaluation, and user study have demonstrated the effectiveness and efficiency of the proposed methods.

cs.CV↗

arXiv2022

Boosting Active Learning via Improving Test Performance

Central to active learning (AL) is what data should be selected for annotation. Existing works attempt to select highly uncertain or informative data for annotation. Nevertheless, it remains unclear how selected data impacts the test performance of the task model used in AL. In this work, we explore such an impact by theoretically proving that selecting unlabeled data of higher gradient norm leads to a lower upper-bound of test loss, resulting in better test performance. However, due to the lack of label information, directly computing gradient norm for unlabeled data is infeasible. To address this challenge, we propose two schemes, namely expected-gradnorm and entropy-gradnorm. The former computes the gradient norm by constructing an expected empirical loss while the latter constructs an unsupervised loss with entropy. Furthermore, we integrate the two schemes in a universal AL framework. We evaluate our method on classical image classification and semantic segmentation tasks. To demonstrate its competency in domain applications and its robustness to noise, we also validate our method on a cellular imaging analysis task, namely cryo-Electron Tomography subtomogram classification. Results demonstrate that our method achieves superior performance against the state of the art. Our source code is available at https://github.com/xulabs/aitom/blob/master/doc/projects/al_gradnorm.md.

cs.LG↗

arXiv2021

Characterization of passive CMOS sensors with RD53A pixel modules

Both the current upgrades to accelerator-based HEP detectors (e.g. ATLAS, CMS) and also future projects (e.g. CEPC, FCC) feature large-area silicon-based tracking detectors. We are investigating the feasibility of using CMOS foundries to fabricate silicon radiation detectors, both for pixels and for large-area strip sensors. A successful proof of concept would open the market potential of CMOS foundries to the HEP community, which would be most beneficial in terms of availability, throughput and cost. In addition, the availability of multi-layer routing of signals will provide the freedom to optimize the sensor geometry and the performance, with biasing structures implemented in poly-silicon layers and MIM-capacitors allowing for AC coupling. A prototyping production of strip test structures and RD53A compatible pixel sensors was recently completed at LFoundry in a 150nm CMOS process. This presentation will focus on the characterization of pixel modules, studying the performance in terms of charge collection, position resolution and hit efficiency with measurements performed in the laboratory and with beam tests. We will report on the investigation of RD53A modules with 25x100 mu^2 cell geometry.

physics.ins-det↗

arXiv2021

Semi-Supervised Active Learning with Temporal Output Discrepancy

While deep learning succeeds in a wide range of tasks, it highly depends on the massive collection of annotated data which is expensive and time-consuming. To lower the cost of data annotation, active learning has been proposed to interactively query an oracle to annotate a small proportion of informative samples in an unlabeled dataset. Inspired by the fact that the samples with higher loss are usually more informative to the model than the samples with lower loss, in this paper we present a novel deep active learning approach that queries the oracle for data annotation when the unlabeled sample is believed to incorporate high loss. The core of our approach is a measurement Temporal Output Discrepancy (TOD) that estimates the sample loss by evaluating the discrepancy of outputs given by models at different optimization steps. Our theoretical investigation shows that TOD lower-bounds the accumulated sample loss thus it can be used to select informative unlabeled samples. On basis of TOD, we further develop an effective unlabeled data sampling strategy as well as an unsupervised learning criterion that enhances model performance by incorporating the unlabeled data. Due to the simplicity of TOD, our active learning approach is efficient, flexible, and task-agnostic. Extensive experimental results demonstrate that our approach achieves superior performances than the state-of-the-art active learning methods on image classification and semantic segmentation tasks.

cs.CV↗

arXiv2021

Radiation tolerant, thin, passive CMOS sensors read out with the RD53A chip

The radiation hardness of passive CMOS pixel sensors fabricated in 150 nm LFoundry technology is investigated. CMOS process lines are especially of interest for large-scale silicon detectors as they offer high production throughput at comparatively low cost. Moreover, several features like poly-silicon resistors, MIM-capacitors and several metal layers are available which can help enhance the sensor design. The performance of a 100 $\mathrmμ$m thin passive CMOS sensor with a pixel pitch of 50 $\mathrmμ$m at different irradiation levels, 5 $\times$ 10$^{15}$n$_{\mathrm{eq}}$cm$^{-2}$ and 1 $\times$ 10$^{16}$n$_{\mathrm{eq}}$cm$^{-2}$, is presented. The sensor was bump-bonded and read out using the RD53A readout chip. After the highest fluence a hit-detection efficiency larger than 99% is measured for minimum ionising particles. The measured equivalent noise charge is comparable to conventional planar pixel sensors. Passive CMOS sensors are thus an attractive option for silicon detectors operating in radiation harsh environments like the upgrades for the LHC experiments.

physics.ins-det↗

arXiv2020

An Evaluation of novel method of Ill-Posed Problem for the Black-Scholes Equation solution

It was proposed by Klibanov a new empirical mathematical method to work with the Black-Scholes equation. This equation is solved forwards in time to forecast prices of stock options. It was used the regularization method because of ill-posed problems. Uniqueness, stability and convergence theorems for this method are formulated. For each individual option, historical data is used for input. The latter is done for two hundred thousand stock options selected from the Bloomberg terminal of University of Washington. It used the index Russell 2000. The main observation is that it was demonstrated that technique, combined with a new trading strategy, results in a significant profit on those options. On the other hand, it was demonstrated the trivial extrapolation techniques results in much lesser profit on those options. This was an experimental work. The minimization process was performed by Hyak Next Generation Supercomputer of the research computing club of University of Washington. As a result, it obtained about 50,000 minimizers. The code is parallelized in order to maximize the performance on supercomputer clusters. Python with the SciPy module was used for implementation. You may find minimizers in the source package that is available on GitHub. Chapter 7 is dedicated to application of machine learning. We were able to improve our results of profitability using minimizers as new data. We classified the minimizer's set to filter for the trading strategy. All results are available on GitHub.

math.NA↗

arXiv2020

Conversion and Implementation of State-of-the-Art Deep Learning Algorithms for the Classification of Diabetic Retinopathy

Diabetic retinopathy (DR) is a retinal microvascular condition that emerges in diabetic patients. DR will continue to be a leading cause of blindness worldwide, with a predicted 191.0 million globally diagnosed patients in 2030. Microaneurysms, hemorrhages, exudates, and cotton wool spots are common signs of DR. However, they can be small and hard for human eyes to detect. Early detection of DR is crucial for effective clinical treatment. Existing methods to classify images require much time for feature extraction and selection, and are limited in their performance. Convolutional Neural Networks (CNNs), as an emerging deep learning (DL) method, have proven their potential in image classification tasks. In this paper, comprehensive experimental studies of implementing state-of-the-art CNNs for the detection and classification of DR are conducted in order to determine the top performing classifiers for the task. Five CNN classifiers, namely Inception-V3, VGG19, VGG16, ResNet50, and InceptionResNetV2, are evaluated through experiments. They categorize medical images into five different classes based on DR severity. Data augmentation and transfer learning techniques are applied since annotated medical images are limited and imbalanced. Experimental results indicate that the ResNet50 classifier has top performance for binary classification and that the InceptionResNetV2 classifier has top performance for multi-class DR classification.

cs.CV↗