arXiv subjects

Sheng Wang

Publications and source records attributed to Sheng Wang.

At least 145 records · Page 8Linked to original sources

arXiv2025

Jodes: Efficient Oblivious Join in the Distributed Setting

Trusted execution environment (TEE) has provided an isolated and secure environment for building cloud-based analytic systems, but it still suffers from access pattern leakages caused by side-channel attacks. To better secure the data, computation inside TEE enclave should be made oblivious, which introduces significant overhead and severely slows down the computation. A natural way to speed up is to build the analytic system with multiple servers in the distributed setting. However, this setting raises a new security concern -- the volumes of the transmissions among these servers can leak sensitive information to a network adversary. Existing works have designed specialized algorithms to address this concern, but their supports for equi-join, one of the most important but non-trivial database operators, are either inefficient, limited, or under a weak security assumption. In this paper, we present Jodes, an efficient oblivious join algorithm in the distributed setting. Jodes prevents the leakage on both the network and enclave sides, supports a general equi-join operation, and provides a high security level protection that only publicizes the input sizes and the output size. Meanwhile, it achieves both communication cost and computation cost asymptotically superior to existing algorithms. To demonstrate the practicality of Jodes, we conduct experiments in the distributed setting comprising 16 servers. Empirical results show that Jodes achieves up to a sixfold performance improvement over state-of-the-art join algorithms.

cs.CR↗

arXiv2025

BiomedCLIP: a multimodal biomedical foundation model pretrained from fifteen million scientific image-text pairs

Biomedical data is inherently multimodal, comprising physical measurements and natural language narratives. A generalist biomedical AI model needs to simultaneously process different modalities of data, including text and images. Therefore, training an effective generalist biomedical model requires high-quality multimodal data, such as parallel image-text pairs. Here, we present PMC-15M, a novel dataset that is two orders of magnitude larger than existing biomedical multimodal datasets such as MIMIC-CXR, and spans a diverse range of biomedical image types. PMC-15M contains 15 million biomedical image-text pairs collected from 4.4 million scientific articles. Based on PMC-15M, we have pretrained BiomedCLIP, a multimodal foundation model, with domain-specific adaptations tailored to biomedical vision-language processing. We conducted extensive experiments and ablation studies on standard biomedical imaging tasks from retrieval to classification to visual question-answering (VQA). BiomedCLIP achieved new state-of-the-art results in a wide range of standard datasets, substantially outperforming prior approaches. Intriguingly, by large-scale pretraining on diverse biomedical image types, BiomedCLIP even outperforms state-of-the-art radiology-specific models such as BioViL in radiology-specific tasks such as RSNA pneumonia detection. In summary, BiomedCLIP is a fully open-access foundation model that achieves state-of-the-art performance on various biomedical tasks, paving the way for transformative multimodal biomedical discovery and applications. We release our models at https://aka.ms/biomedclip to facilitate future research in multimodal biomedical AI.

cs.CV↗

arXiv2025

Normalized ground states for the mass supercritical Schrödinger-Bopp-Podolsky system: existence, uniqueness, limit behavior, strong instability

This paper concerns the normalized ground states for the nonlinear Schrödinger equation in the Bopp-Podolsky electrodynamics. This equation has a nonlocal nonlinearity and a mass supercritical power nonlinearity, both of which have deep impact on the geometry of the corresponding functional, and thus on the existence, limit behavior and stability of the normalized ground states. In the present study, the existence of critical points is obtained by a mountain-pass argument developed on the $L^2$-spheres. To be specific, we show that normalized ground states exist, provided that spherical radius of the $L^2$-spheres is sufficiently small. Then, by discussing the relation between the normalized ground states of the Schrödinger-Bopp-Podolsky system and the classical Schrödinger equation, we show a precise description of the asymptotic behavior of the normalized ground states as the mass vanishes or tends to infinity. Moreover, we discuss the radial symmetry and uniqueness of the normalized ground states. Finally, the strong instability of standing waves at the mountain-pass energy level is studied by constructing an equivalent minimizing problem. Also, as a byproduct, we prove that the mountain-pass energy level gives a threshold for global existence based on this equivalent minimizing problem.

math.AP↗

arXiv2025

Accurate RNA 3D structure prediction using a language model-based deep learning approach

Accurate prediction of RNA three-dimensional (3D) structure remains an unsolved challenge. Determining RNA 3D structures is crucial for understanding their functions and informing RNA-targeting drug development and synthetic biology design. The structural flexibility of RNA, which leads to scarcity of experimentally determined data, complicates computational prediction efforts. Here, we present RhoFold+, an RNA language model-based deep learning method that accurately predicts 3D structures of single-chain RNAs from sequences. By integrating an RNA language model pre-trained on ~23.7 million RNA sequences and leveraging techniques to address data scarcity, RhoFold+ offers a fully automated end-to-end pipeline for RNA 3D structure prediction. Retrospective evaluations on RNA-Puzzles and CASP15 natural RNA targets demonstrate RhoFold+'s superiority over existing methods, including human expert groups. Its efficacy and generalizability are further validated through cross-family and cross-type assessments, as well as time-censored benchmarks. Additionally, RhoFold+ predicts RNA secondary structures and inter-helical angles, providing empirically verifiable features that broaden its applicability to RNA structure and function studies.

q-bio.QM↗

arXiv2024

OCTCube-M: A 3D multimodal optical coherence tomography foundation model for retinal and systemic diseases with cross-cohort and cross-device validation

We present OCTCube-M, a 3D OCT-based multi-modal foundation model for jointly analyzing OCT and en face images. OCTCube-M first developed OCTCube, a 3D foundation model pre-trained on 26,685 3D OCT volumes encompassing 1.62 million 2D OCT images. It then exploits a novel multi-modal contrastive learning framework COEP to integrate other retinal imaging modalities, such as fundus autofluorescence and infrared retinal imaging, into OCTCube, efficiently extending it into multi-modal foundation models. OCTCube achieves best performance on predicting 8 retinal diseases, demonstrating strong generalizability on cross-cohort, cross-device and cross-modality prediction. OCTCube can also predict cross-organ nodule malignancy (CT) and low cardiac ejection fraction as well as systemic diseases, such as diabetes and hypertension, revealing its wide applicability beyond retinal diseases. We further develop OCTCube-IR using COEP with 26,685 OCT and IR image pairs. OCTCube-IR can accurately retrieve between OCT and IR images, allowing joint analysis between 3D and 2D retinal imaging modalities. Finally, we trained a tri-modal foundation model OCTCube-EF from 4 million 2D OCT images and 400K en face retinal images. OCTCube-EF attains the best performance on predicting the growth rate of geographic atrophy (GA) across datasets collected from 6 multi-center global trials conducted in 23 countries. This improvement is statistically equivalent to running a clinical trial with more than double the size of the original study. Our analysis based on another retrospective case study reveals OCTCube-EF's ability to avoid false positive Phase-III results according to its accurate treatment effect estimation on the Phase-II results. In sum, OCTCube-M is a 3D multi-modal foundation model framework that integrates OCT and other retinal imaging modalities revealing substantial diagnostic and prognostic benefits.

eess.IV↗

arXiv2024

Joinable Search over Multi-source Spatial Datasets: Overlap, Coverage, and Efficiency

The search for joinable data is pivotal for numerous applications, such as data integration, data augmentation, and data analysis. Although there have been many successful joinable search studies for table discovery, the study of finding joinable spatial datasets for a given query from multiple spatial data sources has not been well considered. This paper studies two cases of joinable search problems from multiple spatial data sources. In addition to the overlap joinable search problem (OJSP), we also propose a novel coverage joinable search problem (CJSP) that has not been considered before, motivated by many real-world applications in the field of spatial search. To support two cases of joinable search over multiple spatial data sources seamlessly, we propose a multi-source spatial dataset search framework. Firstly, we design a DIstributed Tree-based Spatial index structure called DITS, which is used not only to design acceleration strategies to speed up joinable searches, but also to support efficient communication between multiple data sources. Additionally, we prove that the CJSP is NP-hard and design a greedy approximate algorithm to solve the problem. We evaluate the efficiency of our search framework on five real-world data sources, and the experimental results show that our framework can significantly reduce running time and communication costs compared with baselines.

cs.DB↗

arXiv2024

Thermodynamics of heavy quarkonium in the spinning black hole background

In this paper, we examine the thermodynamics of heavy quarkonium in the spinning black hole background. Specifically, we investigate the effect of angular momentum on the interquark distance, free energy, binding energy, entropy, entropic force, and internal energy of heavy quarkonium from the thermodynamic relationship. Our findings indicate that the angular momentum reduces the maximum value of interquark distance, suggesting that it promotes the dissociation of quarkonium. Additionally, we observe that the angular momentum suppresses free energy. From the results of binding energy, the angular momentum favors the melting of meson into a free quark and antiquark. Moreover, the results show that angular momentum increases the entropy and entropic force, thus accelerates the dissociation of quarkonium. The angular momentum increases the internal energy at large interquark distance. Finally, we find that the angular momentum has a more pronounced effect on quarkonium when the axis of quark pair $Q\overline{Q}$ is transverse to the direction of angular momentum.

hep-ph↗

arXiv2024

On Simplifying Large-Scale Spatial Vectors: Fast, Memory-Efficient, and Cost-Predictable k-means

The k-means algorithm can simplify large-scale spatial vectors, such as 2D geo-locations and 3D point clouds, to support fast analytics and learning. However, when processing large-scale datasets, existing k-means algorithms have been developed to achieve high performance with significant computational resources, such as memory and CPU usage time. These algorithms, though effective, are not well-suited for resource-constrained devices. In this paper, we propose a fast, memory-efficient, and cost-predictable k-means called Dask-means. We first accelerate k-means by designing a memory-efficient accelerator, which utilizes an optimized nearest neighbor search over a memory-tunable index to assign spatial vectors to clusters in batches. We then design a lightweight cost estimator to predict the memory cost and runtime of the k-means task, allowing it to request appropriate memory from devices or adjust the accelerator's required space to meet memory constraints, and ensure sufficient CPU time for running k-means. Experiments show that when simplifying datasets with scale such as $10^6$, Dask-means uses less than $30$MB of memory, achieves over $168$ times speedup compared to the widely-used Lloyd's algorithm. We also validate Dask-means on mobile devices, where it demonstrates significant speedup and low memory cost compared to other state-of-the-art (SOTA) k-means algorithms. Our cost estimator estimates the memory cost with a difference of less than $3\%$ from the actual ones and predicts runtime with an MSE up to $33.3\%$ lower than SOTA methods.

cs.LG↗

arXiv2024

LHPF: Look back the History and Plan for the Future in Autonomous Driving

Decision-making and planning in autonomous driving critically reflect the safety of the system, making effective planning imperative. Current imitation learning-based planning algorithms often merge historical trajectories with present observations to predict future candidate paths. However, these algorithms typically assess the current and historical plans independently, leading to discontinuities in driving intentions and an accumulation of errors with each step in a discontinuous plan. To tackle this challenge, this paper introduces LHPF, an imitation learning planner that integrates historical planning information. Our approach employs a historical intention aggregation module that pools historical planning intentions, which are then combined with a spatial query vector to decode the final planning trajectory. Furthermore, we incorporate a comfort auxiliary task to enhance the human-like quality of the driving behavior. Extensive experiments using both real-world and synthetic data demonstrate that LHPF not only surpasses existing advanced learning-based planners in planning performance but also marks the first instance of a purely learning-based planner outperforming the expert. Additionally, the application of the historical intention aggregation module across various backbones highlights the considerable potential of the proposed method. The code will be made publicly available.

cs.RO↗

arXiv2024

Monocular 3D lane detection for Autonomous Driving: Recent Achievements, Challenges, and Outlooks

3D lane detection is essential in autonomous driving as it extracts structural and traffic information from the road in three-dimensional space, aiding self-driving cars in logical, safe, and comfortable path planning and motion control. Given the cost of sensors and the advantages of visual data in color information, 3D lane detection based on monocular vision is an important research direction in the realm of autonomous driving, increasingly gaining attention in both industry and academia. Regrettably, recent advancements in visual perception seem inadequate for the development of fully reliable 3D lane detection algorithms, which also hampers the progress of vision-based fully autonomous vehicles. We believe that there is still considerable room for improvement in 3D lane detection algorithms for autonomous vehicles using visual sensors, and significant enhancements are needed. This review looks back and analyzes the current state of achievements in the field of 3D lane detection research. It covers all current monocular-based 3D lane detection processes, discusses the performance of these cutting-edge algorithms, analyzes the time complexity of various algorithms, and highlights the main achievements and limitations of ongoing research efforts. The survey also includes a comprehensive discussion of available 3D lane detection datasets and the challenges that researchers face but have not yet resolved. Finally, our work outlines future research directions and invites researchers and practitioners to join this exciting field.

cs.CV↗

arXiv2024

Forewarned is Forearmed: Leveraging LLMs for Data Synthesis through Failure-Inducing Exploration

Large language models (LLMs) have significantly benefited from training on diverse, high-quality task-specific data, leading to impressive performance across a range of downstream applications. Current methods often rely on human-annotated data or predefined task templates to direct powerful LLMs in synthesizing task-relevant data for effective model training. However, this dependence on manually designed components may constrain the scope of generated data, potentially overlooking critical edge cases or novel scenarios that could challenge the model. In this paper, we present a novel approach, ReverseGen, designed to automatically generate effective training samples that expose the weaknesses of LLMs. Specifically, we introduce a dedicated proposer trained to produce queries that lead target models to generate unsatisfactory responses. These failure-inducing queries are then used to construct training data, helping to address the models' shortcomings and improve overall performance. Our approach is flexible and can be applied to models of various scales (3B, 7B, and 8B). We evaluate ReverseGen on three key applications (safety, honesty, and math), demonstrating that our generated data is both highly effective and diverse. Models fine-tuned with ReverseGen-generated data consistently outperform those trained on human-annotated or general model-generated data, offering a new perspective on data synthesis for task-specific LLM enhancement.

cs.CL↗

arXiv2024

Visualization of Skyrmion-Superconducting Vortex Pairs in a Chiral-Magnet-Superconductor Heterostructure

Magnetic skyrmions, the topological states possessing chiral magnetic structure with nontrivial topology, have been widely investigated as a promising candidate for spintronic devices. They can also couple with superconducting vortices to form skyrmion-vortex pairs, hosting Majorana zero mode, which is a potential candidate for topological quantum computing. Many theoretical proposals have been put forward on constructing skyrmion-vortex pairs in heterostructures of chiral magnets and superconductors. Nevertheless, how to generate skyrmion-vortex pairs in a controllable way experimentally remains a significant challenge. We have designed a heterostructure of a chiral magnet and superconductor [Ta/Ir/CoFeB/MgO]7/Nb in which zero field Néel-type skyrmions can be stabilized and the superconducting vortices can couple with the skyrmions when Nb is in the superconducting state. We have directly observed the formation of skyrmion-superconducting vortex pairs that is dependent on the direction of the applied magnetic field. Our results provide an effective method to manipulate the quantum states of skyrmions with the help of superconducting vortices, which can be used to explore new routines to control the skyrmions for spintronics devices.

cond-mat.supr-con↗

arXiv2024

MlingConf: A Comprehensive Study of Multilingual Confidence Estimation on Large Language Models

The tendency of Large Language Models (LLMs) to generate hallucinations raises concerns regarding their reliability. Therefore, confidence estimations indicating the extent of trustworthiness of the generations become essential. However, current LLM confidence estimations in languages other than English remain underexplored. This paper addresses this gap by introducing a comprehensive investigation of Multilingual Confidence estimation (MlingConf) on LLMs, focusing on both language-agnostic (LA) and language-specific (LS) tasks to explore the performance and language dominance effects of multilingual confidence estimations on different tasks. The benchmark comprises four meticulously checked and human-evaluate high-quality multilingual datasets for LA tasks and one for the LS task tailored to specific social, cultural, and geographical contexts of a language. Our experiments reveal that on LA tasks English exhibits notable linguistic dominance in confidence estimations than other languages, while on LS tasks, using question-related language to prompt LLMs demonstrates better linguistic dominance in multilingual confidence estimations. The phenomena inspire a simple yet effective native-tone prompting strategy by employing language-specific prompts for LS tasks, effectively improving LLMs' reliability and accuracy on LS tasks.

cs.CL↗

arXiv2024

RiskMap: A Unified Driving Context Representation for Autonomous Motion Planning in Urban Driving Environment

Motion planning is a complicated task that requires the combination of perception, map information integration and prediction, particularly when driving in heavy traffic. Developing an extensible and efficient representation that visualizes sensor noise and provides basis to real-time planning tasks is desirable. We aim to develop an interpretable map representation, which offers prior of driving cost in planning tasks. In this way, we can simplify the planning process for dealing with complex driving scenarios and visualize sensor noise. Specifically, we propose a unified context representation empowered by deep neural networks. The unified representation is a differentiable risk field, which is an analytical representation of statistical cognition regarding traffic participants for downstream planning tasks. This representation method is nominated as RiskMap. A sampling-based planner is adopted to train and compare RiskMap generation methods. In this paper, the RiskMap generation tools and model structures are explored, the results illustrate that our method can improve driving safety and smoothness, and the limitation of our method is also discussed.

cs.RO↗

arXiv2024

DragTraffic: Interactive and Controllable Traffic Scene Generation for Autonomous Driving

Evaluating and training autonomous driving systems require diverse and scalable corner cases. However, most existing scene generation methods lack controllability, accuracy, and versatility, resulting in unsatisfactory generation results. Inspired by DragGAN in image generation, we propose DragTraffic, a generalized, interactive, and controllable traffic scene generation framework based on conditional diffusion. DragTraffic enables non-experts to generate a variety of realistic driving scenarios for different types of traffic agents through an adaptive mixture expert architecture. We employ a regression model to provide a general initial solution and a refinement process based on the conditional diffusion model to ensure diversity. User-customized context is introduced through cross-attention to ensure high controllability. Experiments on a real-world driving dataset show that DragTraffic outperforms existing methods in terms of authenticity, diversity, and freedom. Demo videos and code are available at https://chantsss.github.io/Dragtraffic/.

cs.RO↗

arXiv2024

A Comprehensive Survey of Scientific Large Language Models and Their Applications in Scientific Discovery

In many scientific fields, large language models (LLMs) have revolutionized the way text and other modalities of data (e.g., molecules and proteins) are handled, achieving superior performance in various applications and augmenting the scientific discovery process. Nevertheless, previous surveys on scientific LLMs often concentrate on one or two fields or a single modality. In this paper, we aim to provide a more holistic view of the research landscape by unveiling cross-field and cross-modal connections between scientific LLMs regarding their architectures and pre-training techniques. To this end, we comprehensively survey over 260 scientific LLMs, discuss their commonalities and differences, as well as summarize pre-training datasets and evaluation tasks for each field and modality. Moreover, we investigate how LLMs have been deployed to benefit scientific discovery. Resources related to this survey are available at https://github.com/yuzhimanhua/Awesome-Scientific-Language-Models.

cs.CL↗

arXiv2024

Velocity Field: An Informative Traveling Cost Representation for Trajectory Planning

Trajectory planning involves generating a series of space points to be followed in the near future. However, due to the complex and uncertain nature of the driving environment, it is impractical for autonomous vehicles~(AVs) to exhaustively design planning rules for optimizing future trajectories. To address this issue, we propose a local map representation method called Velocity Field. This approach provides heading and velocity priors for trajectory planning tasks, simplifying the planning process in complex urban driving. The heading and velocity priors can be learned from demonstrations of human drivers using our proposed loss. Additionally, we developed an iterative sampling-based planner to train and compare the differences between local map representations. We investigated local map representation forms for planning performance on a real-world dataset. Compared to learned rasterized cost maps, our method demonstrated greater reliability and computational efficiency.

cs.RO↗

arXiv2024

A Generic Trajectory Planning Method for Constrained All-Wheel-Steering Robots

This paper presents a generic trajectory planning method for wheeled robots with fixed steering axes while the steering angle of each wheel is constrained. In the existing literatures, All-Wheel-Steering (AWS) robots, incorporating modes such as rotation-free translation maneuvers, in-situ rotational maneuvers, and proportional steering, exhibit inefficient performance due to time-consuming mode switches. This inefficiency arises from wheel rotation constraints and inter-wheel cooperation requirements. The direct application of a holonomic moving strategy can lead to significant slip angles or even structural failure. Additionally, the limited steering range of AWS wheeled robots exacerbates non-linearity characteristics, thereby complicating control processes. To address these challenges, we developed a novel planning method termed Constrained AWS (C-AWS), which integrates second-order discrete search with predictive control techniques. Experimental results demonstrate that our method adeptly generates feasible and smooth trajectories for C-AWS while adhering to steering angle constraints.

cs.RO↗