arXiv subjects

Yang Ni

Publications and source records attributed to Yang Ni.

At least 55 records · Page 3Linked to original sources

arXiv2025

The life cycle of giant molecular clouds in simulated Milky Way-mass galaxies

In this work, we trace the complete life cycle of individual GMCs in high-resolution Milky Way-mass galaxy simulations to determine how different stellar feedback mechanisms and galactic-scale processes govern cloud lifetimes, mass evolution, and local star formation efficiency (SFE). We identify GMCs in simulated galaxies and track their evolution using cloud evolution trees. Via cloud evolution trees, we quantify the lifetimes and SFE of GMCs. We further apply our diagnostics on a suite of simulations with varying star formation and stellar feedback subgrid models and explore their impact together with galactic environments to the GMC life cycles. Our analysis reveals that GMCs undergo dynamic evolution, characterized by continuous gas accretion, gravitational collapse, and star formation, followed by disruption due to stellar feedback. The accretion process sustains the gas content throughout most of the GMC life cycles, resulting in a positive correlation between GMC lifetimes and their maximum masses. The GMC lifetimes range from a few to several tens of Myr, with two distinct dynamical modes: (1) GMCs near the galactic center experience strong tidal disturbances, prolonging their lifetimes when they remain marginally unbound; (2) those in the outer regions are less affected by tides, remain gravitationally bound, and evolve more rapidly. In all model variations, we observe that GMC-scale SFE correlates with the baryonic surface density of GMCs, consistent with previous studies of isolated GMCs. Additionally, we emphasize the critical role of galactic shear in regulating GMC-scale star formation and refine the correlation between local SFE and surface density by including its effects. These findings demonstrate how stellar feedback and galactic-scale dynamics jointly shape GMC-scale star formation in realistic galactic environments.

astro-ph.GA↗

arXiv2025

Truncated Gaussian copula principal component analysis with application to pediatric acute lymphoblastic leukemia patients' gut microbiome

Increasing epidemiologic evidence suggests that the diversity and composition of the gut microbiome can predict infection risk in cancer patients. Infections remain a major cause of morbidity and mortality during chemotherapy. Analyzing microbiome data to identify associations with infection pathogenesis for proactive treatment has become a critical research focus. However, the high-dimensional nature of the data necessitates the use of dimension-reduction methods to facilitate inference and interpretation. Traditional dimension reduction methods, which assume Gaussianity, perform poorly with skewed and zero-inflated microbiome data. To address these challenges, we propose a semiparametric principal component analysis (PCA) method based on a truncated latent Gaussian copula model that accommodates both skewness and zero inflation. Simulation studies demonstrate that the proposed method outperforms existing approaches by providing more accurate estimates of scores and loadings across various copula transformation settings. We apply our method, along with competing approaches, to gut microbiome data from pediatric patients with acute lymphoblastic leukemia. The principal scores derived from the proposed method reveal the strongest associations between pre-chemotherapy microbiome composition and adverse events during subsequent chemotherapy, offering valuable insights for improving patient outcomes.

stat.ME↗

arXiv2025

Enabling Group Fairness in Graph Unlearning via Bi-level Debiasing

Graph unlearning is a crucial approach for protecting user privacy by erasing the influence of user data on trained graph models. Recent developments in graph unlearning methods have primarily focused on maintaining model prediction performance while removing user information. However, we have observed that when user information is deleted from the model, the prediction distribution across different sensitive groups often changes. Furthermore, graph models are shown to be prone to amplifying biases, making the study of fairness in graph unlearning particularly important. This raises the question: Does graph unlearning actually introduce bias? Our findings indicate that the predictions of post-unlearning models become highly correlated with sensitive attributes, confirming the introduction of bias in the graph unlearning process. To address this issue, we propose a fair graph unlearning method, FGU. To guarantee privacy, FGU trains shard models on partitioned subgraphs, unlearns the requested data from the corresponding subgraphs, and retrains the shard models on the modified subgraphs. To ensure fairness, FGU employs a bi-level debiasing process: it first enables shard-level fairness by incorporating a fairness regularizer in the shard model retraining, and then achieves global-level fairness by aligning all shard models to minimize global disparity. Our experiments demonstrate that FGU achieves superior fairness while maintaining privacy and accuracy. Additionally, FGU is robust to diverse unlearning requests, ensuring fairness and utility performance across various data distributions.

cs.LG↗

arXiv2025

The THESAN-ZOOM project: Star formation efficiency from giant molecular clouds to galactic scale in high-redshift starbursts

Star formation in galaxies is inherently complex, involving the interplay of physical processes over a hierarchy of spatial scales. In this work, we investigate the connection between global (galaxy-scale) and local (cloud-scale) star formation efficiencies (SFEs) at high redshifts ($z\gtrsim 3$), using the state-of-the-art cosmological zoom-in simulation suite THESAN-ZOOM. We find that the galaxy-scale average SFE, $\langle ε^{\rm gal}_{\rm ff} \rangle$, scales with $M_{\rm halo}^{1/3}\,(1+z)^{1/2} \sim V_{\rm vir}$, consistent with expectations from feedback-regulated models. On cloud scales, we identify giant molecular clouds (GMCs) in a broad sample of high-redshift starbursts spanning a wide range of halo masses and redshifts. Star formation in these systems is predominantly hosted by filamentary GMCs embedded in a dense and highly turbulent interstellar medium (ISM). GMCs exhibit remarkably universal properties, including mass function, size, turbulence, and surface density, regardless of the environment in which they are identified. The global gas depletion time (and the Kennicutt-Schmidt relation) is determined by the GMC mass fraction in the ISM, while the cloud-scale SFE shows little variation. In particular, we find a nearly constant gas surface density of $Σ_{\rm GMC} \approx 70\,{\rm M}_{\odot}\,{\rm pc}^{-2}$ across different host galaxies. Nevertheless, we identify two regimes where phases with high SFE can arise. First, stars may form efficiently in the shock fronts generated by feedback from a preceding starburst. Second, the increasing background dark matter surface density with redshift may contribute to the gravitational potential of clouds at $z \gtrsim 8$ and confine them in high-SFE phases over extended periods.

astro-ph.GA↗

arXiv2025

Hurdle Network Model With Latent Dynamic Shrinkage For Enhanced Edge Prediction in Zero-Inflated Directed Network Time Series

This article aims to model international trade relationships among 29 countries in the apparel industry between 1994 and 2013. Bilateral trade flows can be represented as a directed network, where nodes correspond to countries and directed edges indicate trade flows (i.e., whether one country exported to another in a given year). Additionally, node (e.g., GDP) and edge-specific (e.g., labor provision) covariates are also available. The study focuses on two key challenges: (1) capturing multiple forms of temporal and network dependence, and dependence on covariates; and (2) accounting for potential trade volume as an important but partially observed edge-specific covariate, which is only available for country pairs that engaged in trade. To address these challenges, we introduce the dynamic hurdle network model (Hurdle-Net) for zero-inflated directed network time series that incorporates several novel features. First, it represents the time series as a paired binary and continuous time series and utilizes a hurdle model that effectively handles sparsity in edge occurrence. Second, the model captures evolving network dependencies using node-specific latent variables governed by a dynamic shrinkage process. Third, it leverages a shared latent structure across the binary and continuous components, reflecting the fact that both networks involve the same nodes. Finally, the model employs a generalized logistic link function to relate edge occurrence to edge weight, allowing for a parsimonious and coherent hierarchical Bayesian framework that jointly models both network components. Compared to static or independent models, Hurdle-Net provides improved model selection, estimation, and prediction performance for analyzing international trade patterns. Its effectiveness is demonstrated through simulation studies and an application to bilateral trade flow data.

stat.ME↗

arXiv2025

LVLM_CSP: Accelerating Large Vision Language Models via Clustering, Scattering, and Pruning for Reasoning Segmentation

Large Vision Language Models (LVLMs) have been widely adopted to guide vision foundation models in performing reasoning segmentation tasks, achieving impressive performance. However, the substantial computational overhead associated with LVLMs presents a new challenge. The primary source of this computational cost arises from processing hundreds of image tokens. Therefore, an effective strategy to mitigate such overhead is to reduce the number of image tokens, a process known as image token pruning. Previous studies on image token pruning for LVLMs have primarily focused on high level visual understanding tasks, such as visual question answering and image captioning. In contrast, guiding vision foundation models to generate accurate visual masks based on textual queries demands precise semantic and spatial reasoning capabilities. Consequently, pruning methods must carefully control individual image tokens throughout the LVLM reasoning process. Our empirical analysis reveals that existing methods struggle to adequately balance reductions in computational overhead with the necessity to maintain high segmentation accuracy. In this work, we propose LVLM_CSP, a novel training free visual token pruning method specifically designed for LVLM based reasoning segmentation tasks. LVLM_CSP consists of three stages: clustering, scattering, and pruning. Initially, the LVLM performs coarse-grained visual reasoning using a subset of selected image tokens. Next, fine grained reasoning is conducted, and finally, most visual tokens are pruned in the last stage. Extensive experiments demonstrate that LVLM_CSP achieves a 65% reduction in image token inference FLOPs with virtually no accuracy degradation, and a 70% reduction with only a minor 1% drop in accuracy on the 7B LVLM.

cs.CV↗

arXiv2025

PacketCLIP: Multi-Modal Embedding of Network Traffic and Language for Cybersecurity Reasoning

Traffic classification is vital for cybersecurity, yet encrypted traffic poses significant challenges. We present PacketCLIP, a multi-modal framework combining packet data with natural language semantics through contrastive pretraining and hierarchical Graph Neural Network (GNN) reasoning. PacketCLIP integrates semantic reasoning with efficient classification, enabling robust detection of anomalies in encrypted network flows. By aligning textual descriptions with packet behaviors, it offers enhanced interpretability, scalability, and practical applicability across diverse security scenarios. PacketCLIP achieves a 95% mean AUC, outperforms baselines by 11.6%, and reduces model size by 92%, making it ideal for real-time anomaly detection. By bridging advanced machine learning techniques and practical cybersecurity needs, PacketCLIP provides a foundation for scalable, efficient, and interpretable solutions to tackle encrypted traffic classification and network intrusion detection challenges in resource-constrained environments.

cs.CR↗

arXiv2024

VLTP: Vision-Language Guided Token Pruning for Task-Oriented Segmentation

Vision Transformers (ViTs) have emerged as the backbone of many segmentation models, consistently achieving state-of-the-art (SOTA) performance. However, their success comes at a significant computational cost. Image token pruning is one of the most effective strategies to address this complexity. However, previous approaches fall short when applied to more complex task-oriented segmentation (TOS), where the class of each image patch is not predefined but dependent on the specific input task. This work introduces the Vision Language Guided Token Pruning (VLTP), a novel token pruning mechanism that can accelerate ViT-based segmentation models, particularly for TOS guided by multi-modal large language model (MLLM). We argue that ViT does not need to process every image token through all of its layers -- only the tokens related to reasoning tasks are necessary. We design a new pruning decoder to take both image tokens and vision-language guidance as input to predict the relevance of each image token to the task. Only image tokens with high relevance are passed to deeper layers of the ViT. Experiments show that the VLTP framework reduces the computational costs of ViT by approximately 25% without performance degradation and by around 40% with only a 1% performance drop. The code associated with this study can be found at this URL.

cs.CV↗

arXiv2024

MR.RGM: An R Package for Fitting Bayesian Multivariate Bidirectional Mendelian Randomization Networks

Motivation: Mendelian randomization (MR) infers causal relationships between exposures and outcomes using genetic variants as instrumental variables. Typically, MR considers only a pair of exposure and outcome at a time, limiting its capability of capturing the entire causal network. We overcome this limitation by developing 'MR.RGM' (Mendelian randomization via reciprocal graphical model), a fast R-package that implements the Bayesian reciprocal graphical model and enables practitioners to construct holistic causal networks with possibly cyclic/reciprocal causation and proper uncertainty quantifications, offering a comprehensive understanding of complex biological systems and their interconnections. We developed 'MR.RGM', an open-source R package that applies bidirectional MR using a network-based strategy, enabling the exploration of causal relationships among multiple variables in complex biological systems. 'MR.RGM' holds the promise of unveiling intricate interactions and advancing our understanding of genetic networks, disease risks, and phenotypic complexities.

stat.AP↗

arXiv2024

TaskCLIP: Extend Large Vision-Language Model for Task Oriented Object Detection

Task-oriented object detection aims to find objects suitable for accomplishing specific tasks. As a challenging task, it requires simultaneous visual data processing and reasoning under ambiguous semantics. Recent solutions are mainly all-in-one models. However, the object detection backbones are pre-trained without text supervision. Thus, to incorporate task requirements, their intricate models undergo extensive learning on a highly imbalanced and scarce dataset, resulting in capped performance, laborious training, and poor generalizability. In contrast, we propose TaskCLIP, a more natural two-stage design composed of general object detection and task-guided object selection. Particularly for the latter, we resort to the recently successful large Vision-Language Models (VLMs) as our backbone, which provides rich semantic knowledge and a uniform embedding space for images and texts. Nevertheless, the naive application of VLMs leads to sub-optimal quality, due to the misalignment between embeddings of object images and their visual attributes, which are mainly adjective phrases. To this end, we design a transformer-based aligner after the pre-trained VLMs to re-calibrate both embeddings. Finally, we employ a trainable score function to post-process the VLM matching results for object selection. Experimental results demonstrate that our TaskCLIP outperforms the state-of-the-art DETR-based model TOIST by 3.5% and only requires a single NVIDIA RTX 4090 for both training and inference.

cs.CV↗

arXiv2024

Recoverable Anonymization for Pose Estimation: A Privacy-Enhancing Approach

Human pose estimation (HPE) is crucial for various applications. However, deploying HPE algorithms in surveillance contexts raises significant privacy concerns due to the potential leakage of sensitive personal information (SPI) such as facial features, and ethnicity. Existing privacy-enhancing methods often compromise either privacy or performance, or they require costly additional modalities. We propose a novel privacy-enhancing system that generates privacy-enhanced portraits while maintaining high HPE performance. Our key innovations include the reversible recovery of SPI for authorized personnel and the preservation of contextual information. By jointly optimizing a privacy-enhancing module, a privacy recovery module, and a pose estimator, our system ensures robust privacy protection, efficient SPI recovery, and high-performance HPE. Experimental results demonstrate the system's robust performance in privacy enhancement, SPI recovery, and HPE.

cs.CV↗

arXiv2024

Blocked Gibbs sampler for hierarchical Dirichlet processes

Posterior computation in hierarchical Dirichlet process (HDP) mixture models is an active area of research in nonparametric Bayes inference of grouped data. Existing literature almost exclusively focuses on the Chinese restaurant franchise (CRF) analogy of the marginal distribution of the parameters, which can mix poorly and has a quadratic complexity with the sample size. A recently developed slice sampler allows for efficient blocked updates of the parameters, but is shown to be statistically unstable in our article. We develop a blocked Gibbs sampler that employs a truncated approximation of the underlying random measures to sample from the posterior distribution of HDP, which produces statistically stable results, is highly scalable with respect to sample size, and is shown to have good mixing. The heart of the construction is to endow the shared concentration parameter with an appropriately chosen gamma prior that allows us to break the dependence of the shared mixing proportions and permits independent updates of certain log-concave random variables in a block. En route, we develop an efficient rejection sampler for these random variables leveraging piece-wise tangent-line approximations.

stat.CO↗

arXiv2024

EcoSense: Energy-Efficient Intelligent Sensing for In-Shore Ship Detection through Edge-Cloud Collaboration

Detecting marine objects inshore presents challenges owing to algorithmic intricacies and complexities in system deployment. We propose a difficulty-aware edge-cloud collaborative sensing system that splits the task into object localization and fine-grained classification. Objects are classified either at the edge or within the cloud, based on their estimated difficulty. The framework comprises a low-power device-tailored front-end model for object localization, classification, and difficulty estimation, along with a transformer-graph convolutional network-based back-end model for fine-grained classification. Our system demonstrates superior performance (mAP@0.5 +4.3%}) on widely used marine object detection datasets, significantly reducing both data transmission volume (by 95.43%) and energy consumption (by 72.7%}) at the system level. We validate the proposed system across various embedded system platforms and in real-world scenarios involving drone deployment.

cs.CV↗

arXiv2024

Generalized Criterion for Identifiability of Additive Noise Models Using Majorization

The discovery of causal relationships from observational data is very challenging. Many recent approaches rely on complexity or uncertainty concepts to impose constraints on probability distributions, aiming to identify specific classes of directed acyclic graph (DAG) models. In this paper, we introduce a novel identifiability criterion for DAGs that places constraints on the conditional variances of additive noise models. We demonstrate that this criterion extends and generalizes existing identifiability criteria in the literature that employ (conditional) variances as measures of uncertainty in (conditional) distributions. For linear Structural Equation Models, we present a new algorithm that leverages the concept of weak majorization applied to the diagonal elements of the Cholesky factor of the covariance matrix to learn a topological ordering of variables. Through extensive simulations and the analysis of bank connectivity data, we provide evidence of the effectiveness of our approach in successfully recovering DAGs. The code for reproducing the results in this paper is available in Supplementary Materials.

stat.ME↗

arXiv2024

Bayesian segmented Gaussian copula factor model for single-cell sequencing data

Single-cell sequencing technologies have significantly advanced molecular and cellular biology, offering unprecedented insights into cellular heterogeneity by allowing for the measurement of gene expression at an individual cell level. However, the analysis of such data is challenged by the prevalence of low counts due to dropout events and the skewed nature of the data distribution, which conventional Gaussian factor models struggle to handle effectively. To address these challenges, we propose a novel Bayesian segmented Gaussian copula model to explicitly account for inflation of zero and near-zero counts, and to address the high skewness in the data. By employing a Dirichlet-Laplace prior for each column of the factor loadings matrix, we shrink the loadings of unnecessary factors towards zero, which leads to a simple approach to automatically determine the number of latent factors, and resolve the identifiability issue inherent in factor models due to the rotational invariance of the factor loadings matrix. Through simulation studies, we demonstrate the superior performance of our method over existing approaches in conducting factor analysis on data exhibiting the characteristics of single-cell data, such as excessive low counts and high skewness. Furthermore, we apply the proposed method to a real single-cell RNA-sequencing dataset from a lymphoblastoid cell line, successfully identifying biologically meaningful latent factors and detecting previously uncharacterized cell subtypes.

stat.ME↗

arXiv2024

HDReason: Algorithm-Hardware Codesign for Hyperdimensional Knowledge Graph Reasoning

In recent times, a plethora of hardware accelerators have been put forth for graph learning applications such as vertex classification and graph classification. However, previous works have paid little attention to Knowledge Graph Completion (KGC), a task that is well-known for its significantly higher algorithm complexity. The state-of-the-art KGC solutions based on graph convolution neural network (GCN) involve extensive vertex/relation embedding updates and complicated score functions, which are inherently cumbersome for acceleration. As a result, existing accelerator designs are no longer optimal, and a novel algorithm-hardware co-design for KG reasoning is needed. Recently, brain-inspired HyperDimensional Computing (HDC) has been introduced as a promising solution for lightweight machine learning, particularly for graph learning applications. In this paper, we leverage HDC for an intrinsically more efficient and acceleration-friendly KGC algorithm. We also co-design an acceleration framework named HDReason targeting FPGA platforms. On the algorithm level, HDReason achieves a balance between high reasoning accuracy, strong model interpretability, and less computation complexity. In terms of architecture, HDReason offers reconfigurability, high training throughput, and low energy consumption. When compared with NVIDIA RTX 4090 GPU, the proposed accelerator achieves an average 10.6x speedup and 65x energy efficiency improvement. When conducting cross-models and cross-platforms comparison, HDReason yields an average 4.2x higher performance and 3.4x better energy efficiency with similar accuracy versus the state-of-the-art FPGA-based GCN training platform.

cs.AR↗

arXiv2024

HEAL: Brain-inspired Hyperdimensional Efficient Active Learning

Drawing inspiration from the outstanding learning capability of our human brains, Hyperdimensional Computing (HDC) emerges as a novel computing paradigm, and it leverages high-dimensional vector presentation and operations for brain-like lightweight Machine Learning (ML). Practical deployments of HDC have significantly enhanced the learning efficiency compared to current deep ML methods on a broad spectrum of applications. However, boosting the data efficiency of HDC classifiers in supervised learning remains an open question. In this paper, we introduce Hyperdimensional Efficient Active Learning (HEAL), a novel Active Learning (AL) framework tailored for HDC classification. HEAL proactively annotates unlabeled data points via uncertainty and diversity-guided acquisition, leading to a more efficient dataset annotation and lowering labor costs. Unlike conventional AL methods that only support classifiers built upon deep neural networks (DNN), HEAL operates without the need for gradient or probabilistic computations. This allows it to be effortlessly integrated with any existing HDC classifier architecture. The key design of HEAL is a novel approach for uncertainty estimation in HDC classifiers through a lightweight HDC ensemble with prior hypervectors. Additionally, by exploiting hypervectors as prototypes (i.e., compact representations), we develop an extra metric for HEAL to select diverse samples within each batch for annotation. Our evaluation shows that HEAL surpasses a diverse set of baselines in AL quality and achieves notably faster acquisition than many BNN-powered or diversity-guided AL methods, recording 11 times to 40,000 times speedup in acquisition runtime per batch.

cs.LG↗

arXiv2024

A Plug-in Tiny AI Module for Intelligent and Selective Sensor Data Transmission

Applications in the Internet of Things (IoT) utilize machine learning to analyze sensor-generated data. However, a major challenge lies in the lack of targeted intelligence in current sensing systems, leading to vast data generation and increased computational and communication costs. To address this challenge, we propose a novel sensing module to equip sensing frameworks with intelligent data transmission capabilities by integrating a highly efficient machine learning model placed near the sensor. This model provides prompt feedback for the sensing system to transmit only valuable data while discarding irrelevant information by regulating the frequency of data transmission. The near-sensor model is quantized and optimized for real-time sensor control. To enhance the framework's performance, the training process is customized and a "lazy" sensor deactivation strategy utilizing temporal information is introduced. The suggested method is orthogonal to other IoT frameworks and can be considered as a plugin for selective data transmission. The framework is implemented, encompassing both software and hardware components. The experiments demonstrate that the framework utilizing the suggested module achieves over 85% system efficiency in terms of energy consumption and storage, with negligible impact on performance. This methodology has the potential to significantly reduce data output from sensors, benefiting a wide range of IoT applications.

cs.LG↗