arXiv subjects

Bing Wang

Publications and source records attributed to Bing Wang.

At least 163 records · Page 9Linked to original sources

arXiv2024

FAGStyle: Feature Augmentation on Geodesic Surface for Zero-shot Text-guided Diffusion Image Style Transfer

The goal of image style transfer is to render an image guided by a style reference while maintaining the original content. Existing image-guided methods rely on specific style reference images, restricting their wider application and potentially compromising result quality. As a flexible alternative, text-guided methods allow users to describe the desired style using text prompts. Despite their versatility, these methods often struggle with maintaining style consistency, reflecting the described style accurately, and preserving the content of the target image. To address these challenges, we introduce FAGStyle, a zero-shot text-guided diffusion image style transfer method. Our approach enhances inter-patch information interaction by incorporating the Sliding Window Crop technique and Feature Augmentation on Geodesic Surface into our style control loss. Furthermore, we integrate a Pre-Shape self-correlation consistency loss to ensure content consistency. FAGStyle demonstrates superior performance over existing methods, consistently achieving stylization that retains the semantic content of the source image. Experimental results confirms the efficacy of FAGStyle across a diverse range of source contents and styles, both imagined and common.

cs.CV↗

arXiv2024

Early detection of inflammatory arthritis to improve referrals using multimodal machine learning from blood testing, semi-structured and unstructured patient records

Early detection of inflammatory arthritis (IA) is critical to efficient and accurate hospital referral triage for timely treatment and preventing the deterioration of the IA disease course, especially under limited healthcare resources. The manual assessment process is the most common approach in practice for the early detection of IA, but it is extremely labor-intensive and inefficient. A large amount of clinical information needs to be assessed for every referral from General Practice (GP) to the hospitals. Machine learning shows great potential in automating repetitive assessment tasks and providing decision support for the early detection of IA. However, most machine learning-based methods for IA detection rely on blood testing results. But in practice, blood testing data is not always available at the point of referrals, so we need methods to leverage multimodal data such as semi-structured and unstructured data for early detection of IA. In this research, we present fusion and ensemble learning-based methods using multimodal data to assist decision-making in the early detection of IA, and a conformal prediction-based method to quantify the uncertainty of the prediction and detect any unreliable predictions. To the best of our knowledge, our study is the first attempt to utilize multimodal data to support the early detection of IA from GP referrals.

cs.LG↗

arXiv2024

Harmfully Manipulated Images Matter in Multimodal Misinformation Detection

Nowadays, misinformation is widely spreading over various social media platforms and causes extremely negative impacts on society. To combat this issue, automatically identifying misinformation, especially those containing multimodal content, has attracted growing attention from the academic and industrial communities, and induced an active research topic named Multimodal Misinformation Detection (MMD). Typically, existing MMD methods capture the semantic correlation and inconsistency between multiple modalities, but neglect some potential clues in multimodal content. Recent studies suggest that manipulated traces of the images in articles are non-trivial clues for detecting misinformation. Meanwhile, we find that the underlying intentions behind the manipulation, e.g., harmful and harmless, also matter in MMD. Accordingly, in this work, we propose to detect misinformation by learning manipulation features that indicate whether the image has been manipulated, as well as intention features regarding the harmful and harmless intentions of the manipulation. Unfortunately, the manipulation and intention labels that make these features discriminative are unknown. To overcome the problem, we propose two weakly supervised signals as alternatives by introducing additional datasets on image manipulation detection and formulating two classification tasks as positive and unlabeled learning problems. Based on these ideas, we propose a novel MMD method, namely Harmfully Manipulated Images Matter in MMD (HAMI-M3D). Extensive experiments across three benchmark datasets can demonstrate that HAMI-M3D can consistently improve the performance of any MMD baselines.

cs.CL↗

arXiv2024

Why Misinformation is Created? Detecting them by Integrating Intent Features

Various social media platforms, e.g., Twitter and Reddit, allow people to disseminate a plethora of information more efficiently and conveniently. However, they are inevitably full of misinformation, causing damage to diverse aspects of our daily lives. To reduce the negative impact, timely identification of misinformation, namely Misinformation Detection (MD), has become an active research topic receiving widespread attention. As a complex phenomenon, the veracity of an article is influenced by various aspects. In this paper, we are inspired by the opposition of intents between misinformation and real information. Accordingly, we propose to reason the intent of articles and form the corresponding intent features to promote the veracity discrimination of article features. To achieve this, we build a hierarchy of a set of intents for both misinformation and real information by referring to the existing psychological theories, and we apply it to reason the intent of articles by progressively generating binary answers with an encoder-decoder structure. We form the corresponding intent features and integrate it with the token features to achieve more discriminative article features for MD. Upon these ideas, we suggest a novel MD method, namely Detecting Misinformation by Integrating Intent featuRes (DM-INTER). To evaluate the performance of DM-INTER, we conduct extensive experiments on benchmark MD datasets. The experimental results validate that DM-INTER can outperform the existing baseline MD methods.

cs.CL↗

arXiv2024

STS MICCAI 2023 Challenge: Grand challenge on 2D and 3D semi-supervised tooth segmentation

Computer-aided design (CAD) tools are increasingly popular in modern dental practice, particularly for treatment planning or comprehensive prognosis evaluation. In particular, the 2D panoramic X-ray image efficiently detects invisible caries, impacted teeth and supernumerary teeth in children, while the 3D dental cone beam computed tomography (CBCT) is widely used in orthodontics and endodontics due to its low radiation dose. However, there is no open-access 2D public dataset for children's teeth and no open 3D dental CBCT dataset, which limits the development of automatic algorithms for segmenting teeth and analyzing diseases. The Semi-supervised Teeth Segmentation (STS) Challenge, a pioneering event in tooth segmentation, was held as a part of the MICCAI 2023 ToothFairy Workshop on the Alibaba Tianchi platform. This challenge aims to investigate effective semi-supervised tooth segmentation algorithms to advance the field of dentistry. In this challenge, we provide two modalities including the 2D panoramic X-ray images and the 3D CBCT tooth volumes. In Task 1, the goal was to segment tooth regions in panoramic X-ray images of both adult and pediatric teeth. Task 2 involved segmenting tooth sections using CBCT volumes. Limited labelled images with mostly unlabelled ones were provided in this challenge prompt using semi-supervised algorithms for training. In the preliminary round, the challenge received registration and result submission by 434 teams, with 64 advancing to the final round. This paper summarizes the diverse methods employed by the top-ranking teams in the STS MICCAI 2023 Challenge.

cs.CV↗

arXiv2024

Compact ultra-broadband light coupling on chip via nonadiabatic pumping

Enlarging bandwidth capacity of the integrated photonic systems demands efficient and broadband light coupling among optical elements, which has been a vital issue in integrated photonics. Here, we have developed a compact ultra-broadband light coupling strategy based on nonadiabatic pumping in coupled optical waveguides, and experimentally demonstrated the designs in thin-film lithium niobate on insulator (LNOI) platform. We found that nonadiabatic transition would produce a decreased dispersion of the phases related to eigenstates in the waveguides. As a consequence, we realized high-efficiency directional transfer between edgestates for various wavelengths covering a 1-dB bandwidth of ~320 nm in experiment (>400 nm in simulation), with a coupling length (~50 μm) approximately 1/10 of that required in the adiabatic regime. Furthermore, we have constructed complex functional devices including beamsplitter and multiple-level cascaded networks for broadband light routing and splitting. Our work preserves significant advantages simultaneously in extending the operation bandwidth and minimizing the footprint, which demonstrates great potential for large-scale and compact photonic integration on chip.

physics.optics↗

arXiv2024

UniCoder: Scaling Code Large Language Model via Universal Code

Intermediate reasoning or acting steps have successfully improved large language models (LLMs) for handling various downstream natural language processing (NLP) tasks. When applying LLMs for code generation, recent works mainly focus on directing the models to articulate intermediate natural-language reasoning steps, as in chain-of-thought (CoT) prompting, and then output code with the natural language or other structured intermediate steps. However, such output is not suitable for code translation or generation tasks since the standard CoT has different logical structures and forms of expression with the code. In this work, we introduce the universal code (UniCode) as the intermediate representation. It is a description of algorithm steps using a mix of conventions of programming languages, such as assignment operator, conditional operator, and loop. Hence, we collect an instruction dataset UniCoder-Instruct to train our model UniCoder on multi-task learning objectives. UniCoder-Instruct comprises natural-language questions, code solutions, and the corresponding universal code. The alignment between the intermediate universal code representation and the final code solution significantly improves the quality of the generated code. The experimental results demonstrate that UniCoder with the universal code significantly outperforms the previous prompting methods by a large margin, showcasing the effectiveness of the structural clues in pseudo-code.

cs.CL↗

arXiv2024

Security of Partially Corrupted Repeater Chains

Quantum Key Distribution allows two parties to establish a secret key that is secure against computationally unbounded adversaries. To extend the distance between parties, quantum networks, and in particular repeater chains, are vital. Typically, security in such scenarios assumes the absolute worst case: namely, an adversary has complete control over all repeaters and fiber links in a network and is able to replace them with perfect devices, thus allowing her to hide her attack within the expected natural noise. In a large-scale network, however, such a powerful attack may be infeasible. In this paper, we analyze the case where the adversary can only corrupt a contiguous subset of a repeater chain connecting Alice and Bob, while some portion of the network near Alice and Bob may be considered safe from attack (though still noisy). We derive a rigorous finite key proof of security assuming this attack model and show that improved performance and noise tolerances are possible.

quant-ph↗

arXiv2024

Matching Query Image Against Selected NeRF Feature for Efficient and Scalable Localization

Neural implicit representations such as NeRF have revolutionized 3D scene representation with photo-realistic quality. However, existing methods for visual localization within NeRF representations suffer from inefficiency and scalability issues, particularly in large-scale environments. This work proposes MatLoc-NeRF, a novel matching-based localization framework using selected NeRF features. It addresses efficiency by employing a learnable feature selection mechanism that identifies informative NeRF features for matching with query images. This eliminates the need for all NeRF features or additional descriptors, leading to faster and more accurate pose estimation. To tackle large-scale scenes, MatLoc-NeRF utilizes a pose-aware scene partitioning strategy. It ensures that only the most relevant NeRF sub-block generates key features for a specific pose. Additionally, scene segmentation and a place predictor provide fast coarse initial pose estimation. Evaluations on public large-scale datasets demonstrate that MatLoc-NeRF achieves superior efficiency and accuracy compared to existing NeRF-based localization methods.

cs.CV↗

arXiv2024

McEval: Massively Multilingual Code Evaluation

Code large language models (LLMs) have shown remarkable advances in code understanding, completion, and generation tasks. Programming benchmarks, comprised of a selection of code challenges and corresponding test cases, serve as a standard to evaluate the capability of different LLMs in such tasks. However, most existing benchmarks primarily focus on Python and are still restricted to a limited number of languages, where other languages are translated from the Python samples (e.g. MultiPL-E) degrading the data diversity. To further facilitate the research of code LLMs, we propose a massively multilingual code benchmark covering 40 programming languages (McEval) with 16K test samples, which substantially pushes the limits of code LLMs in multilingual scenarios. The benchmark contains challenging code completion, understanding, and generation evaluation tasks with finely curated massively multilingual instruction corpora McEval-Instruct. In addition, we introduce an effective multilingual coder mCoder trained on McEval-Instruct to support multilingual programming language generation. Extensive experimental results on McEval show that there is still a difficult journey between open-source models and closed-source LLMs (e.g. GPT-series models) in numerous languages. The instruction corpora, evaluation benchmark, and leaderboard are available at \url{https://mceval.github.io/}.

cs.PL↗

arXiv2024

How time weathers galaxies: The temporal impact of the cluster environment on galaxy formation and evolution

We illuminate the altered evolution of galaxies in clusters compared to the field by tracking galaxies in the IllustrisTNG300 simulation as they enter isolated clusters of mass $10^{13} < M_{\rm 200, mean} / {\rm M}_\odot < 10^{15}$ (at $z=0$). We demonstrate significant trends in galaxy properties with residence time (time since first infall) and that there is a population of galaxies that remain star-forming even many Gyrs after their infall. By comparing the properties of galaxies at their infall time to their properties at $z=0$, we show how scaling relations, like the stellar-to-halo mass ratio, shift as galaxies live in the cluster environment. Galaxies with a residence time of 10 Gyr increase their stellar-to-halo mass ratio, by around 1 dex. As measurements of the steepest slope of the galaxy cluster number density profile ($R_{\rm st}$), frequently used as a proxy for the splashback radius, have been shown to depend strongly on galaxy selection, we show how $R_{\rm st}$ depends on galaxy residence time. Using galaxies with residence times less than one cluster crossing time ($\approx 5$ Gyr) to measure $R_{\rm st}$ leads to significant offsets relative to using the entire galaxy population. Galaxies must have had the opportunity to `splash back' to the first caustic to trace out a representative value of $R_{\rm st}$, potentially leading to issues for galaxy surveys using UV-selected galaxies. Our wok demonstrates that the evolution of cluster galaxies continues well into their lifetime in the cluster and departs from a typical field galaxy evolutionary path.

astro-ph.GA↗

arXiv2024

A Multi-Peak Solar Flare with a High Turnover Frequency of The Gyrosynchrotron Spectra from the Loop-Top Source

The origin of multiple peaks in lightcurves of various wavelengths remains illusive during flares. Here we discuss the flare of SOL2023-05-09T03:54M6.5 with six flux peaks as recorded by a tandem of new microwave and Hard X-ray instruments. According to its microwave spectra, the flare represents a high-turnover frequency (>15 GHz) event. The rather-complete microwave and HXR spectral coverage provides a rare opportunity to uncover the origin of such event together with simultaneous EUV images. We concluded that (1) the microwave sources originates around the top section of the flaring loops with a trend of source spatial dispersion with frequency;(2) the visible movement of the microwave source from peak to peak originates from the process of new flaring loops appearing sequentially along the magnetic neutral line; 3) the optically-thin microwave spectra are hard with the indices varying from -1.2 to -0.4, and the turnover frequency always exceeds 15 GHz; 4) higher turnover/peak frequency corresponds to stronger peak intensity and harder optically-thin spectra. Using the Fokker-Planck and GX simulator codes we obtained a good fit to the observed microwave spectra and spatial distribution of the sources at all peaks, if assuming the radiating energetic electrons have the same spatial distribution and single-power-law spectra but with the number density varying in a range of 30%. We conclude that the particle acceleration in this flare happens in a compact region nearing the looptop. These results provide new constraints on the acceleration of energetic electrons and the underlying flare intermittent reconnection process.

astro-ph.SR↗

arXiv2024

Efficient Multiparty Quantum Key Distribution over Quantum Networks

Multiparty quantum key distribution (QKD) is useful for many applications that involve secure communication or collaboration among multiple parties. While it can be achieved using pairwise QKD, a more efficient approach is to achieve it using multipartite entanglement distributed over quantum networks that connect the multiple parties. Existing studies on multipartite entanglement distribution, however, are not designed for multiparty QKD, and hence do not aim to maximize secret key generation rate. In this paper, we design efficient strategies for multiparty QKD over quantum networks. For 3-party QKD, we derive closed-form expressions for analyzing key distribution over quantum networks. We then use it to develop an efficient strategy for 3-party QKD by packing multiple stars that connect the 3 parties. For the general form of N-party QKD, we develop an approach that packs multiple trees to connect the N parties, while directly incorporating the estimated key rates on network paths. Extensive evaluation of our strategies, in both grid and random graphs, under a wide range of settings, demonstrates that our schemes achieve high key rate, which degrades gracefully when increasing the number of parties.

quant-ph↗

arXiv2024

Finite Key Security of Simplified Trusted Node Networks

Simplified trusted nodes (STNs) are a form of trusted node for quantum key distribution (QKD) networks which do not require running a full QKD stack every instance (i.e., they do not need to run error correction and privacy amplification each session). Such systems hold the advantage that they may be implemented with weaker computational abilities, than regular TNs, while still keeping up with key generation rate demands. The downside is that noise tolerance is lower. However, to get a better understanding of their suitability in various scenarios, one requires practical, finite-key security bounds for STN networks. So far, only theoretical asymptotic bounds are known. In this work we derive a new proof of security for STN chains in the finite key setting. We also derive a novel cost function allowing us to evaluate when STNs would be beneficial from a computational cost perspective, compared with regular TN networks.

quant-ph↗

arXiv2024

FreeReg: Image-to-Point Cloud Registration Leveraging Pretrained Diffusion Models and Monocular Depth Estimators

Matching cross-modality features between images and point clouds is a fundamental problem for image-to-point cloud registration. However, due to the modality difference between images and points, it is difficult to learn robust and discriminative cross-modality features by existing metric learning methods for feature matching. Instead of applying metric learning on cross-modality data, we propose to unify the modality between images and point clouds by pretrained large-scale models first, and then establish robust correspondence within the same modality. We show that the intermediate features, called diffusion features, extracted by depth-to-image diffusion models are semantically consistent between images and point clouds, which enables the building of coarse but robust cross-modality correspondences. We further extract geometric features on depth maps produced by the monocular depth estimator. By matching such geometric features, we significantly improve the accuracy of the coarse correspondences produced by diffusion features. Extensive experiments demonstrate that without any task-specific training, direct utilization of both features produces accurate image-to-point cloud registration. On three public indoor and outdoor benchmarks, the proposed method averagely achieves a 20.6 percent improvement in Inlier Ratio, a three-fold higher Inlier Number, and a 48.6 percent improvement in Registration Recall than existing state-of-the-arts.

cs.CV↗

arXiv2024

Knowledge Distillation via the Target-aware Transformer

Knowledge distillation becomes a de facto standard to improve the performance of small neural networks. Most of the previous works propose to regress the representational features from the teacher to the student in a one-to-one spatial matching fashion. However, people tend to overlook the fact that, due to the architecture differences, the semantic information on the same spatial location usually vary. This greatly undermines the underlying assumption of the one-to-one distillation approach. To this end, we propose a novel one-to-all spatial matching knowledge distillation approach. Specifically, we allow each pixel of the teacher feature to be distilled to all spatial locations of the student features given its similarity, which is generated from a target-aware transformer. Our approach surpasses the state-of-the-art methods by a significant margin on various computer vision benchmarks, such as ImageNet, Pascal VOC and COCOStuff10k. Code is available at https://github.com/sihaoevery/TaT.

cs.CV↗

arXiv2024

Improving Opioid Use Disorder Risk Modelling through Behavioral and Genetic Feature Integration

Opioids are an effective analgesic for acute and chronic pain, but also carry a considerable risk of addiction leading to millions of opioid use disorder (OUD) cases and tens of thousands of premature deaths in the United States yearly. Estimating OUD risk prior to prescription could improve the efficacy of treatment regimens, monitoring programs, and intervention strategies, but risk estimation is typically based on self-reported data or questionnaires. We develop an experimental design and computational methods that combine genetic variants associated with OUD with behavioral features extracted from GPS and Wi-Fi spatiotemporal coordinates to assess OUD risk. Since both OUD mobility and genetic data do not exist for the same cohort, we develop algorithms to (1) generate mobility features from empirical distributions and (2) synthesize mobility and genetic samples assuming an expected level of disease co-occurrence. We show that integrating genetic and mobility modalities improves risk modelling using classification accuracy, area under the precision-recall and receiver operator characteristic curves, and $F_1$ score. Interpreting the fitted models suggests that mobility features have more influence on OUD risk, although the genetic contribution was significant, particularly in linear models. While there exist concerns with respect to privacy, security, bias, and generalizability that must be evaluated in clinical trials before being implemented in practice, our framework provides preliminary evidence that behavioral and genetic features may improve OUD risk estimation to assist with personalized clinical decision-making.

q-bio.QM↗

arXiv2024

m3P: Towards Multimodal Multilingual Translation with Multimodal Prompt

Multilingual translation supports multiple translation directions by projecting all languages in a shared space, but the translation quality is undermined by the difference between languages in the text-only modality, especially when the number of languages is large. To bridge this gap, we introduce visual context as the universal language-independent representation to facilitate multilingual translation. In this paper, we propose a framework to leverage the multimodal prompt to guide the Multimodal Multilingual neural Machine Translation (m3P), which aligns the representations of different languages with the same meaning and generates the conditional vision-language memory for translation. We construct a multilingual multimodal instruction dataset (InstrMulti102) to support 102 languages. Our method aims to minimize the representation distance of different languages by regarding the image as a central language. Experimental results show that m3P outperforms previous text-only baselines and multilingual multimodal methods by a large margin. Furthermore, the probing experiments validate the effectiveness of our method in enhancing translation under the low-resource and massively multilingual scenario.

cs.CL↗