arXiv subjects

Zeyu Gao

Publications and source records attributed to Zeyu Gao.

At least 55 records · Page 3Linked to original sources

arXiv2025

PH2ST:ST-Prompt Guided Histological Hypergraph Learning for Spatial Gene Expression Prediction

Spatial Transcriptomics (ST) reveals the spatial distribution of gene expression in tissues, offering critical insights into biological processes and disease mechanisms. However, the high cost, limited coverage, and technical complexity of current ST technologies restrict their widespread use in clinical and research settings, making obtaining high-resolution transcriptomic profiles across large tissue areas challenging. Predicting ST from H\&E-stained histology images has emerged as a promising alternative to address these limitations but remains challenging due to the heterogeneous relationship between histomorphology and gene expression, which is affected by substantial variability across patients and tissue sections. In response, we propose PH2ST, a prompt-guided hypergraph learning framework, which leverages limited ST signals to guide multi-scale histological representation learning for accurate and robust spatial gene expression prediction. Extensive evaluations on two public ST datasets and multiple prompt sampling strategies simulating real-world scenarios demonstrate that PH2ST not only outperforms existing state-of-the-art methods, but also shows strong potential for practical applications such as imputing missing spots, ST super-resolution, and local-to-global prediction, highlighting its value for scalable and cost-effective spatial gene expression mapping in biomedical contexts.

cs.CV↗

arXiv2025

RoboTwin: Dual-Arm Robot Benchmark with Generative Digital Twins

In the rapidly advancing field of robotics, dual-arm coordination and complex object manipulation are essential capabilities for developing advanced autonomous systems. However, the scarcity of diverse, high-quality demonstration data and real-world-aligned evaluation benchmarks severely limits such development. To address this, we introduce RoboTwin, a generative digital twin framework that uses 3D generative foundation models and large language models to produce diverse expert datasets and provide a real-world-aligned evaluation platform for dual-arm robotic tasks. Specifically, RoboTwin creates varied digital twins of objects from single 2D images, generating realistic and interactive scenarios. It also introduces a spatial relation-aware code generation framework that combines object annotations with large language models to break down tasks, determine spatial constraints, and generate precise robotic movement code. Our framework offers a comprehensive benchmark with both simulated and real-world data, enabling standardized evaluation and better alignment between simulated training and real-world performance. We validated our approach using the open-source COBOT Magic Robot platform. Policies pre-trained on RoboTwin-generated data and fine-tuned with limited real-world samples demonstrate significant potential for enhancing dual-arm robotic manipulation systems by improving success rates by over 70% for single-arm tasks and over 40% for dual-arm tasks compared to models trained solely on real-world data.

cs.RO↗

arXiv2025

RoboTwin: Dual-Arm Robot Benchmark with Generative Digital Twins (early version)

In the rapidly advancing field of robotics, dual-arm coordination and complex object manipulation are essential capabilities for developing advanced autonomous systems. However, the scarcity of diverse, high-quality demonstration data and real-world-aligned evaluation benchmarks severely limits such development. To address this, we introduce RoboTwin, a generative digital twin framework that uses 3D generative foundation models and large language models to produce diverse expert datasets and provide a real-world-aligned evaluation platform for dual-arm robotic tasks. Specifically, RoboTwin creates varied digital twins of objects from single 2D images, generating realistic and interactive scenarios. It also introduces a spatial relation-aware code generation framework that combines object annotations with large language models to break down tasks, determine spatial constraints, and generate precise robotic movement code. Our framework offers a comprehensive benchmark with both simulated and real-world data, enabling standardized evaluation and better alignment between simulated training and real-world performance. We validated our approach using the open-source COBOT Magic Robot platform. Policies pre-trained on RoboTwin-generated data and fine-tuned with limited real-world samples improve the success rate of over 70% for single-arm tasks and over 40% for dual-arm tasks compared to models trained solely on real-world data. This significant improvement demonstrates RoboTwin's potential to enhance the development and evaluation of dual-arm robotic manipulation systems. Project Page: https://robotwin-benchmark.github.io/early-version/.

cs.RO↗

arXiv2025

Formation and Environmental Context of Giant Bulgeless Disk Galaxies in the Early Universe: Insights from Cosmological Simulations

Giant bulgeless disk galaxies, theoretically expected to be rare in the early Universe, have been confirmed by the James Webb Space Telescope (JWST) to exist as early as 2 billion years after the Big Bang. These morphologically extreme systems offer valuable insights into the physics of disk formation and the interplay between galaxies and their dark-matter halos. Using cosmological simulations, we identify analogs of such galaxies with stellar masses around $10^{11} M_\odot$ and half-light radii up to 6 kpc at $z \sim 3$ and characterize the factors that contribute to their formation. These galaxies form in young cosmic knots, populating host halos of high spin, low concentration, and spherical shapes. They feature dynamically coherent circum-galactic medium, as well as gas-rich, coherent mergers, which preserve their disk morphology and drive their large sizes. Interestingly, all the simulated giant disks harbor a compact, aligned inner disk, marginally resolvable in JWST images with a Sérsic index near unity. These findings highlight the environmental and structural conditions necessary for forming and sustaining giant bulgeless disks and provide a theoretical framework for interpreting JWST observations of extreme disk morphologies in the early Universe.

astro-ph.GA↗

arXiv2025

Coherence Properties of Rare-Earth Spins in Micrometer-Thin Films

Rare-earth ions in bulk crystals are excellent solid-state quantum systems in quantum information science, owing to the exceptional optical and spin coherence properties. However, the weak fluorescence of single rare-earth ions present a significant challenge for scalability, necessitating the integration into micro-cavities. Thin films serve as a promising material platform for the integration, yet the fabrication without compromising the properties of the materials and rare-earth ions remains challenging. In this work, we fabricate micrometer-thin yttrium aluminum garnet (YAG) films from bulk crystals using ion implantation techniques. The resulting films preserve the single-crystalline structure of the original bulk crystal. Notably, the embedded rare-earth ions are photo-stable and exhibit bulk-like spin coherence properties. Our results demonstrate the compatibility of bulk-like spin properties with the thin-film fabrication technique, facilitating the efficient integration of rare-earth ions into on-chip photonic devices and advancing the applications of rare-earth ions systems in quantum technologies.

quant-ph↗

arXiv2025

The critical role of dark matter halos in driving star formation

Understanding the physical mechanisms that drive star formation is crucial for advancing our knowledge of galaxy evolution. We explore the interrelationships between key galaxy properties associated with star formation, with a particular focus on the impact of dark matter halos. Given the sensitivity of atomic hydrogen (HI) to external processes, we concentrate exclusively on central spiral galaxies. We find that the molecular-to-atomic gas mass ratio ($M_{\rm H_2}/M_{\rm HI}$) strongly depends on stellar mass and specific star formation rate (sSFR). In the star formation efficiency (SFE)-sSFR plane, most galaxies fall below the H$_2$ fundamental formation relation (FFR), with SFE$_{\rm HI}$ being consistently lower than SFE$_{\rm H_2}$. Using the improved halo masses derived by Zhao et al. (2025), for star-forming galaxies, both SFE$_{\rm HI}$ and $M_{\rm H_2}/M_{\rm HI}$ increase rapidly and monotonically with halo mass, indicating a higher efficiency in converting HI to H$_2$ in more massive halos. This trend ultimately leads to the unsustainable state where SFE$_{\rm HI}$ exceeds SFE$_{\rm H_2}$ at halo mass around $10^{12} \hbox{$M_{\odot}$}$. For halos with masses exceeding $10^{12} \hbox{$M_{\odot}$}$, galaxies predominantly experience quenching. We propose a plausible evolutionary scenario in which the growth of halo mass regulates the conversion of HI to H$_2$, star formation, and the eventual quenching of galaxies. The disk size, primarily regulated by the mass, spin and concentration of the dark matter halo, also significantly influences HI to H$_2$ conversion and star formation. These findings underscore the critical role of dark matter halos as a global regulator of galaxy-wide star formation, a key factor that has been largely underappreciated in previous studies.

astro-ph.GA↗

arXiv2025

From Halos to Galaxies. X: Decoding Galaxy SEDs with Physical Priors and Accurate Star Formation History Reconstruction

The spectral energy distribution (SED) of galaxies is essential for deriving fundamental properties like stellar mass and star formation history (SFH). However, conventional methods, including both parametric and non-parametric approaches, often fail to accurately recover the observed cosmic star formation rate (SFR) density due to oversimplified or unrealistic assumptions about SFH and their inability to account for the complex SFH variations across different galaxy populations. To address this issue, we introduce a novel approach that improves galaxy broadband SED analysis by incorporating physical priors derived from hydrodynamical simulations. Tests using IllustrisTNG simulations demonstrate that our method can reliably determine galaxy physical properties from broadband photometry, including stellar mass within 0.05 dex, current SFR within 0.3 dex, and fractional stellar formation time within 0.2 dex, with a negligible fraction of catastrophic failures. When applied to the Sloan Digital Sky Survey (SDSS) main photometric galaxy sample with spectroscopic redshift, our estimates of stellar mass and SFR are consistent with the widely used MPA-JHU and GSWLC catalogs. Notably, using the derived SFHs of individual SDSS galaxies, we estimate the cosmic SFR density and stellar mass density with remarkable consistency to direct observations up to $z \sim 6$. This demonstrates a significant advancement in deriving SFHs from SEDs that closely align with observational data. Consequently, our method can reliably recover observed spectral indices such as $\rm D_{\rm n}(4000)$ and $\rm Hδ_{\rm A}$ by synthesizing the full spectra of galaxies using the estimated SFHs and metal enrichment histories, relying solely on broadband photometry as input. Furthermore, this method is extremely computationally efficient compared to conventional approaches.

astro-ph.GA↗

arXiv2025

From Halos to Galaxies. VI. Improved halo mass estimation for SDSS groups and measurement of the halo mass function

In $Λ$CDM cosmology, galaxies form and evolve in their host dark matter (DM) halos. Halo mass is crucial for understanding the halo-galaxy connection. The abundance matching (AM) technique has been widely used to derive the halo masses of galaxy groups. However, quenching of the central galaxy can decouple the coevolution of its stellar mass and DM halo mass. Different halo assembly histories can also result in significantly different final stellar mass of the central galaxies. These processes can introduce substantial uncertainties in the halo masses derived from the AM method, particularly leading to a systematic bias between groups with star-forming centrals (blue groups) and passive centrals (red groups). To improve, we developed a new machine learning (ML) algorithm that accounts for these effects and is trained on simulations. Our results show that the ML method eliminates the systematic bias in the derived halo masses for blue and red groups and is, on average, $\sim1/3$ more accurate than the AM method. With careful calibration of observable quantities from simulations and observations from SDSS, we apply our ML model to the SDSS Yang et al. groups to derive their halo masses down to $10^{11.5}\mathrm{M_\odot}$ or even lower. The derived SDSS group halo mass function agrees well with the theoretical predictions, and the derived stellar-to-halo mass relations for both red and blue groups matches well with those obtained from direct weak lensing measurements. These new halo mass estimates enable more accurate investigation of the galaxy-halo connection and the role of the halos in galaxy evolution.

astro-ph.GA↗

arXiv2024

Abundant Molecular Gas in the Central Region of Lenticular Galaxy PGC 39535

Lenticular galaxies (S0s) in the local universe are generally absent of recent star formation and lack molecular gas. In this paper, we investigate one massive ($M_*$$\sim$5$\times10^{10}$ M$_\odot$) star-forming S0, PGC 39535, with the Northern Extended Millimeter Array (NOEMA). Using optical data from SDSS-IV MaNGA survey, we find star formation mainly concentrates in the central region of PGC 39535. The total star formation rate estimated using extinction-corrected H$α$ flux is 1.57 M$_\odot$ yr$^{-1}$. Results of NOEMA observation suggest that the molecular gas mainly concentrates in the central regions as a gaseous bar and a ring-like structure, and shows similar kinematics as the stellar and ionized gas components. The total molecular gas mass estimated from CO(1-0) is (5.42$\pm$1.52)$\times$10$^{9}$ M$_{\odot}$. We find PGC 39535 lies on the star-forming main sequence, but falls below Kennicutt-Schmidt relation of spiral galaxies, suggesting that the star formation efficiency may be suppressed by the massive bulge. The existence of a second Gaussian component in the CO spectrum of the central region indicates possible gas flows. Furthermore, our analyses suggest that PGC 39535 resides in the center of a massive group and the derived star formation history indicates it may experience a series of gas-rich mergers over the past 2$\sim$7 Gyr.

astro-ph.GA↗

arXiv2024

ViC: Virtual Compiler Is All You Need For Assembly Code Search

Assembly code search is vital for reducing the burden on reverse engineers, allowing them to quickly identify specific functions using natural language within vast binary programs. Despite its significance, this critical task is impeded by the complexities involved in building high-quality datasets. This paper explores training a Large Language Model (LLM) to emulate a general compiler. By leveraging Ubuntu packages to compile a dataset of 20 billion tokens, we further continue pre-train CodeLlama as a Virtual Compiler (ViC), capable of compiling any source code of any language to assembly code. This approach allows for virtual compilation across a wide range of programming languages without the need for a real compiler, preserving semantic equivalency and expanding the possibilities for assembly code dataset construction. Furthermore, we use ViC to construct a sufficiently large dataset for assembly code search. Employing this extensive dataset, we achieve a substantial improvement in assembly code search performance, with our model surpassing the leading baseline by 26%.

cs.SE↗

arXiv2024

A Self-Supervised Image Registration Approach for Measuring Local Response Patterns in Metastatic Ovarian Cancer

High-grade serous ovarian carcinoma (HGSOC) is characterised by significant spatial and temporal heterogeneity, typically manifesting at an advanced metastatic stage. A major challenge in treating advanced HGSOC is effectively monitoring localised change in tumour burden across multiple sites during neoadjuvant chemotherapy (NACT) and predicting long-term pathological response and overall patient survival. In this work, we propose a self-supervised deformable image registration algorithm that utilises a general-purpose image encoder for image feature extraction to co-register contrast-enhanced computerised tomography scan images acquired before and after neoadjuvant chemotherapy. This approach addresses challenges posed by highly complex tumour deformations and longitudinal lesion matching during treatment. Localised tumour changes are calculated using the Jacobian determinant maps of the registration deformation at multiple disease sites and their macroscopic areas, including hypo-dense (i.e., cystic/necrotic), hyper-dense (i.e., calcified), and intermediate density (i.e., soft tissue) portions. A series of experiments is conducted to understand the role of a general-purpose image encoder and its application in quantifying change in tumour burden during neoadjuvant chemotherapy in HGSOC. This work is the first to demonstrate the feasibility of a self-supervised image registration approach in quantifying NACT-induced localised tumour changes across the whole disease burden of patients with complex multi-site HGSOC, which could be used as a potential marker for ovarian cancer patient's long-term pathological response and survival.

cs.CV↗

arXiv2024

Enhance Sample Efficiency and Robustness of End-to-end Urban Autonomous Driving via Semantic Masked World Model

End-to-end autonomous driving provides a feasible way to automatically maximize overall driving system performance by directly mapping the raw pixels from a front-facing camera to control signals. Recent advanced methods construct a latent world model to map the high dimensional observations into compact latent space. However, the latent states embedded by the world model proposed in previous works may contain a large amount of task-irrelevant information, resulting in low sampling efficiency and poor robustness to input perturbations. Meanwhile, the training data distribution is usually unbalanced, and the learned policy is challenging to cope with the corner cases during the driving process. To solve the above challenges, we present a SEMantic Masked recurrent world model (SEM2), which introduces a semantic filter to extract key driving-relevant features and make decisions via the filtered features, and is trained with a multi-source data sampler, which aggregates common data and multiple corner case data in a single batch, to balance the data distribution. Extensive experiments on CARLA show our method outperforms the state-of-the-art approaches in terms of sample efficiency and robustness to input permutations.

cs.LG↗

arXiv2024

CEBin: A Cost-Effective Framework for Large-Scale Binary Code Similarity Detection

Binary code similarity detection (BCSD) is a fundamental technique for various application. Many BCSD solutions have been proposed recently, which mostly are embedding-based, but have shown limited accuracy and efficiency especially when the volume of target binaries to search is large. To address this issue, we propose a cost-effective BCSD framework, CEBin, which fuses embedding-based and comparison-based approaches to significantly improve accuracy while minimizing overheads. Specifically, CEBin utilizes a refined embedding-based approach to extract features of target code, which efficiently narrows down the scope of candidate similar code and boosts performance. Then, it utilizes a comparison-based approach that performs a pairwise comparison on the candidates to capture more nuanced and complex relationships, which greatly improves the accuracy of similarity detection. By bridging the gap between embedding-based and comparison-based approaches, CEBin is able to provide an effective and efficient solution for detecting similar code (including vulnerable ones) in large-scale software ecosystems. Experimental results on three well-known datasets demonstrate the superiority of CEBin over existing state-of-the-art (SOTA) baselines. To further evaluate the usefulness of BCSD in real world, we construct a large-scale benchmark of vulnerability, offering the first precise evaluation scheme to assess BCSD methods for the 1-day vulnerability detection task. CEBin could identify the similar function from millions of candidate functions in just a few seconds and achieves an impressive recall rate of $85.46\%$ on this more practical but challenging task, which are several order of magnitudes faster and $4.07\times$ better than the best SOTA baseline. Our code is available at https://github.com/Hustcw/CEBin.

cs.SE↗

arXiv2024

CLAP: Learning Transferable Binary Code Representations with Natural Language Supervision

Binary code representation learning has shown significant performance in binary analysis tasks. But existing solutions often have poor transferability, particularly in few-shot and zero-shot scenarios where few or no training samples are available for the tasks. To address this problem, we present CLAP (Contrastive Language-Assembly Pre-training), which employs natural language supervision to learn better representations of binary code (i.e., assembly code) and get better transferability. At the core, our approach boosts superior transfer learning capabilities by effectively aligning binary code with their semantics explanations (in natural language), resulting a model able to generate better embeddings for binary code. To enable this alignment training, we then propose an efficient dataset engine that could automatically generate a large and diverse dataset comprising of binary code and corresponding natural language explanations. We have generated 195 million pairs of binary code and explanations and trained a prototype of CLAP. The evaluations of CLAP across various downstream tasks in binary analysis all demonstrate exceptional performance. Notably, without any task-specific training, CLAP is often competitive with a fully supervised baseline, showing excellent transferability. We release our pre-trained model and code at https://github.com/Hustcw/CLAP.

cs.SE↗

arXiv2023

How Far Have We Gone in Vulnerability Detection Using Large Language Models

As software becomes increasingly complex and prone to vulnerabilities, automated vulnerability detection is critically important, yet challenging. Given the significant successes of large language models (LLMs) in various tasks, there is growing anticipation of their efficacy in vulnerability detection. However, a quantitative understanding of their potential in vulnerability detection is still missing. To bridge this gap, we introduce a comprehensive vulnerability benchmark VulBench. This benchmark aggregates high-quality data from a wide range of CTF (Capture-the-Flag) challenges and real-world applications, with annotations for each vulnerable function detailing the vulnerability type and its root cause. Through our experiments encompassing 16 LLMs and 6 state-of-the-art (SOTA) deep learning-based models and static analyzers, we find that several LLMs outperform traditional deep learning approaches in vulnerability detection, revealing an untapped potential in LLMs. This work contributes to the understanding and utilization of LLMs for enhanced software security.

cs.AI↗

arXiv2023

Automated Small Kidney Cancer Detection in Non-Contrast Computed Tomography

This study introduces an automated pipeline for renal cancer (RC) detection in non-contrast computed tomography (NCCT). In the development of our pipeline, we test three detections models: a shape model, a 2D-, and a 3D axial-sample model. Training (n=1348) and testing (n=64) data were gathered from open sources (KiTS23, Abdomen1k, CT-ORG) and Cambridge University Hospital (CUH). Results from cross-validation and testing revealed that the 2D axial sample model had the highest small ($\leq$40mm diameter) RC detection area under the curve (AUC) of 0.804. Our pipeline achieves 61.9\% sensitivity and 92.7\% specificity for small kidney cancers on unseen test data. Our results are much more accurate than previous attempts to automatically detect small renal cancers in NCCT, the most likely imaging modality for RC screening. This pipeline offers a promising advance that may enable screening for kidney cancers.

eess.IV↗

arXiv2023

kTrans: Knowledge-Aware Transformer for Binary Code Embedding

Binary Code Embedding (BCE) has important applications in various reverse engineering tasks such as binary code similarity detection, type recovery, control-flow recovery and data-flow analysis. Recent studies have shown that the Transformer model can comprehend the semantics of binary code to support downstream tasks. However, existing models overlooked the prior knowledge of assembly language. In this paper, we propose a novel Transformer-based approach, namely kTrans, to generate knowledge-aware binary code embedding. By feeding explicit knowledge as additional inputs to the Transformer, and fusing implicit knowledge with a novel pre-training task, kTrans provides a new perspective to incorporating domain knowledge into a Transformer framework. We inspect the generated embeddings with outlier detection and visualization, and also apply kTrans to 3 downstream tasks: Binary Code Similarity Detection (BCSD), Function Type Recovery (FTR) and Indirect Call Recognition (ICR). Evaluation results show that kTrans can generate high-quality binary code embeddings, and outperforms state-of-the-art (SOTA) approaches on downstream tasks by 5.2%, 6.8%, and 12.6% respectively. kTrans is publicly available at: https://github.com/Learner0x5a/kTrans-release

cs.SE↗

arXiv2023

Spiking Neural Network for Ultra-low-latency and High-accurate Object Detection

Spiking Neural Networks (SNNs) have garnered widespread interest for their energy efficiency and brain-inspired event-driven properties. While recent methods like Spiking-YOLO have expanded the SNNs to more challenging object detection tasks, they often suffer from high latency and low detection accuracy, making them difficult to deploy on latency sensitive mobile platforms. Furthermore, the conversion method from Artificial Neural Networks (ANNs) to SNNs is hard to maintain the complete structure of the ANNs, resulting in poor feature representation and high conversion errors. To address these challenges, we propose two methods: timesteps compression and spike-time-dependent integrated (STDI) coding. The former reduces the timesteps required in ANN-SNN conversion by compressing information, while the latter sets a time-varying threshold to expand the information holding capacity. We also present a SNN-based ultra-low latency and high accurate object detection model (SUHD) that achieves state-of-the-art performance on nontrivial datasets like PASCAL VOC and MS COCO, with about remarkable 750x fewer timesteps and 30% mean average precision (mAP) improvement, compared to the Spiking-YOLO on MS COCO datasets. To the best of our knowledge, SUHD is the deepest spike-based object detection model to date that achieves ultra low timesteps to complete the lossless conversion.

cs.CV↗