arXiv subjects

Qiong Zhang

Publications and source records attributed to Qiong Zhang.

At least 55 records · Page 3Linked to original sources

arXiv2024

Just a Few Glances: Open-Set Visual Perception with Image Prompt Paradigm

To break through the limitations of pre-training models on fixed categories, Open-Set Object Detection (OSOD) and Open-Set Segmentation (OSS) have attracted a surge of interest from researchers. Inspired by large language models, mainstream OSOD and OSS methods generally utilize text as a prompt, achieving remarkable performance. Following SAM paradigm, some researchers use visual prompts, such as points, boxes, and masks that cover detection or segmentation targets. Despite these two prompt paradigms exhibit excellent performance, they also reveal inherent limitations. On the one hand, it is difficult to accurately describe characteristics of specialized category using textual description. On the other hand, existing visual prompt paradigms heavily rely on multi-round human interaction, which hinders them being applied to fully automated pipeline. To address the above issues, we propose a novel prompt paradigm in OSOD and OSS, that is, \textbf{Image Prompt Paradigm}. This brand new prompt paradigm enables to detect or segment specialized categories without multi-round human intervention. To achieve this goal, the proposed image prompt paradigm uses just a few image instances as prompts, and we propose a novel framework named \textbf{MI Grounding} for this new paradigm. In this framework, high-quality image prompts are automatically encoded, selected and fused, achieving the single-stage and non-interactive inference. We conduct extensive experiments on public datasets, showing that MI Grounding achieves competitive performance on OSOD and OSS benchmarks compared to text prompt paradigm methods and visual prompt paradigm methods. Moreover, MI Grounding can greatly outperform existing method on our constructed specialized ADR50K dataset.

cs.CV↗

arXiv2024

Collaborative Design of Controlled Experiments in the Presence of Subject Covariates

We consider the optimal experimental design problem of allocating subjects to treatment or control when subjects participate in multiple, separate controlled experiments within a short time-frame and subject covariate information is available. Here, in addition to subject covariates, we consider the dependence among the responses coming from the subject's random effect across experiments. In this setting, the goal of the allocation is to provide precise estimates of treatment effects for each experiment. Deriving the precision matrix of the treatment effects and using D-optimality as our allocation criterion, we demonstrate the advantage of collaboratively designing and analyzing multiple experiments over traditional independent design and analysis, and propose two randomized algorithms to provide solutions to the D-optimality problem for collaborative design. The first algorithm decomposes the D-optimality problem into a sequence of subproblems, where each subproblem is a quadratic binary program that can be solved through a semi-definite relaxation based randomized algorithm with performance guarantees. The second algorithm involves solving a single semi-definite program, and randomly generating allocations for each experiment from the solution of this program. We showcase the performance of these algorithms through a simulation study, finding that our algorithms outperform covariate-agnostic methods when there are a large number of covariates.

stat.ME↗

arXiv2024

Lift Force on a Moving Intruder in Granular Shear Flow

Lift and drag forces on moving intruders in granular materials are of fundamental interest. While the drag force on an intruder in granular flow has been studied, the few studies characterizing the lift force explore a relatively limited range of parameters. Here we use discrete element method (DEM) simulations to measure the lift force, $F_\mathrm{L}$, on a spherical intruder in a uniformly sheared bed of smaller spheres for a range of intruder slip velocities, $u_\mathrm{s}$, relative to the unperturbed flow. In what at first appears as a puzzling result, $F_\mathrm{L}$ in granular shear flow acts in the opposite direction to the Saffman lift force on a sphere in a sheared fluid at low $u_\mathrm{s}$, reaches a maximum value, and then decreases, eventually reversing direction and becoming comparable to $F_\mathrm{L}$ for a fluid. This non-monotonic response holds over a range of flow conditions, and the $F_\mathrm{L}$ versus $u_\mathrm{s}$ data can be collapsed by scaling both quantities using the particle sizes, shear rate, and overburden pressure. Analogous fluid simulations demonstrate that the flow field around the intruder particle is similar in the granular and fluid cases. However, the shear stress acting on the intruder in a granular shear flow is much less than that in a fluid shear flow. This difference, combined with a void region behind the intruder in granular flow, which alters the pressure and shear stress on the trailing side of the intruder, significantly changes the lift-force inducing stresses acting on the intruder between the granular and fluid cases.

cond-mat.soft↗

arXiv2024

Stability and Optimal Decay Rates for Abstract Systems with Thermal Damping of Cattaneo's Type

This paper studies the stability of an abstract thermoelastic system with Cattaneo's law, which describes finite heat propagation speed in a medium. We introduce a region of parameters containing coupling, thermal dissipation, and possible inertial characteristics. The region is partitioned into distinct subregions based on the spectral properties of the generator of the corresponding semigroup. By a careful estimation of the resolvent operator on the imaginary axis, we obtain distinct polynomial decay rates for systems with parameters located in different subregions. Furthermore, the optimality of these decay rates is proved. Finally, we apply our results to several coupled systems of partial differential equations.

math.AP↗

arXiv2024

Stability of the Abstract Thermoelastic System with Singularity

In this paper, we analyze an abstract thermoelastic system, where the heat conduction follows the Cattaneo law. Zero becomes a spectrum point of the system operator when the coupling and thermal damping parameters of system satisfy specific conditions. We obtain the decay rates of solutions to the system with or without the inertial term. Furthermore, the decay rate of the system without inertial terms is shown to be optimal.

math.AP↗

arXiv2024

Collaborative Analysis for Paired A/B Testing Experiments

With the extensive use of digital devices, online experimental platforms are commonly used to conduct experiments to collect data for evaluating different variations of products, algorithms, and interface designs, a.k.a., A/B tests. In practice, multiple A/B testing experiments are often carried out based on a common user population on the same platform. The same user's responses to different experiments can be correlated to some extent due to the individual effect of the user. In this paper, we propose a novel framework that collaboratively analyzes the data from paired A/B tests, namely, a pair of A/B testing experiments conducted on the same set of experimental subjects. The proposed analysis approach for paired A/B tests can lead to more accurate estimates than the traditional separate analysis of each experiment. We obtain the asymptotic distribution of the proposed estimators and demonstrate that the proposed estimators are asymptotically the best linear unbiased estimators under certain assumptions. Moreover, the proposed analysis approach is computationally efficient, easy to implement, and robust to different types of responses. Both numerical simulations and numerical studies based on a real case are used to examine the performance of the proposed method.

stat.ME↗

arXiv2024

Reconciling Shared versus Context-Specific Information in a Neural Network Model of Latent Causes

It has been proposed that, when processing a stream of events, humans divide their experiences in terms of inferred latent causes (LCs) to support context-dependent learning. However, when shared structure is present across contexts, it is still unclear how the "splitting" of LCs and learning of shared structure can be simultaneously achieved. Here, we present the Latent Cause Network (LCNet), a neural network model of LC inference. Through learning, it naturally stores structure that is shared across tasks in the network weights. Additionally, it represents context-specific structure using a context module, controlled by a Bayesian nonparametric inference algorithm, which assigns a unique context vector for each inferred LC. Across three simulations, we found that LCNet could 1) extract shared structure across LCs in a function learning task while avoiding catastrophic interference, 2) capture human data on curriculum effects in schema learning, and 3) infer the underlying event structure when processing naturalistic videos of daily events. Overall, these results demonstrate a computationally feasible approach to reconciling shared structure and context-specific structure in a model of LCs that is scalable from laboratory experiment settings to naturalistic settings.

q-bio.NC↗

arXiv2024

FedMT: Federated Learning with Mixed-type Labels

In federated learning (FL), classifiers (e.g., deep networks) are trained on datasets from multiple data centers without exchanging data across them, which improves the sample efficiency. However, the conventional FL setting assumes the same labeling criterion in all data centers involved, thus limiting its practical utility. This limitation becomes particularly notable in domains like disease diagnosis, where different clinical centers may adhere to different standards, making traditional FL methods unsuitable. This paper addresses this important yet under-explored setting of FL, namely FL with mixed-type labels, where the allowance of different labeling criteria introduces inter-center label space differences. To address this challenge effectively and efficiently, we introduce a model-agnostic approach called FedMT, which estimates label space correspondences and projects classification scores to construct loss functions. The proposed FedMT is versatile and integrates seamlessly with various FL methods, such as FedAvg. Experimental results on benchmark and medical datasets highlight the substantial improvement in classification accuracy achieved by FedMT in the presence of mixed-type labels.

cs.LG↗

arXiv2023

How should the advent of large language models affect the practice of science?

Large language models (LLMs) are being increasingly incorporated into scientific workflows. However, we have yet to fully grasp the implications of this integration. How should the advent of large language models affect the practice of science? For this opinion piece, we have invited four diverse groups of scientists to reflect on this query, sharing their perspectives and engaging in debate. Schulz et al. make the argument that working with LLMs is not fundamentally different from working with human collaborators, while Bender et al. argue that LLMs are often misused and over-hyped, and that their limitations warrant a focus on more specialized, easily interpretable tools. Marelli et al. emphasize the importance of transparent attribution and responsible use of LLMs. Finally, Botvinick and Gershman advocate that humans should retain responsibility for determining the scientific roadmap. To facilitate the discussion, the four perspectives are complemented with a response from each group. By putting these different perspectives in conversation, we aim to bring attention to important considerations within the academic community regarding the adoption of LLMs and their impact on both current and future scientific practices.

cs.CL↗

arXiv2023

Gaussian Mixture Reduction with Composite Transportation Divergence

Gaussian mixtures are widely used for approximating density functions in various applications such as density estimation, belief propagation, and Bayesian filtering. These applications often utilize Gaussian mixtures as initial approximations that are updated recursively. A key challenge in these recursive processes stems from the exponential increase in the mixture's order, resulting in intractable inference. To overcome the difficulty, the Gaussian mixture reduction (GMR), which approximates a high order Gaussian mixture by one with a lower order, can be used. Although existing clustering-based methods are known for their satisfactory performance and computational efficiency, their convergence properties and optimal targets remain unknown. In this paper, we propose a novel optimization-based GMR method based on composite transportation divergence (CTD). We develop a majorization-minimization algorithm for computing the reduced mixture and establish its theoretical convergence under general conditions. Furthermore, we demonstrate that many existing clustering-based methods are special cases of ours, effectively bridging the gap between optimization-based and clustering-based techniques. Our unified framework empowers users to select the most appropriate cost function in CTD to achieve superior performance in their specific applications. Through extensive empirical experiments, we demonstrate the efficiency and effectiveness of our proposed method, showcasing its potential in various domains.

stat.ML↗

arXiv2023

On the Asymptotics of Graph Cut Objectives for Experimental Designs of Network A/B Testing

A/B testing is an effective way to assess the potential impacts of two treatments. For A/B tests conducted by IT companies, the test users of A/B testing are often connected and form a social network. The responses of A/B testing can be related to the network connection of test users. This paper discusses the relationship between the design criteria of network A/B testing and graph cut objectives. We develop asymptotic distributions of graph cut objectives to enable rerandomization algorithms for the design of network A/B testing under two scenarios.

stat.ME↗

arXiv2022

Reinforced Structured State-Evolution for Vision-Language Navigation

Vision-and-language Navigation (VLN) task requires an embodied agent to navigate to a remote location following a natural language instruction. Previous methods usually adopt a sequence model (e.g., Transformer and LSTM) as the navigator. In such a paradigm, the sequence model predicts action at each step through a maintained navigation state, which is generally represented as a one-dimensional vector. However, the crucial navigation clues (i.e., object-level environment layout) for embodied navigation task is discarded since the maintained vector is essentially unstructured. In this paper, we propose a novel Structured state-Evolution (SEvol) model to effectively maintain the environment layout clues for VLN. Specifically, we utilise the graph-based feature to represent the navigation state instead of the vector-based state. Accordingly, we devise a Reinforced Layout clues Miner (RLM) to mine and detect the most crucial layout graph for long-term navigation via a customised reinforcement learning strategy. Moreover, the Structured Evolving Module (SEM) is proposed to maintain the structured graph-based state during navigation, where the state is gradually evolved to learn the object-level spatial-temporal relationship. The experiments on the R2R and R4R datasets show that the proposed SEvol model improves VLN models' performance by large margins, e.g., +3% absolute SPL accuracy for NvEM and +8% for EnvDrop on the R2R test set.

cs.CV↗

arXiv2022

Locally Optimal Design for A/B Testing in the Presence of Covariates and Network Connection

A/B test, a simple type of controlled experiment, refers to the statistical procedure of experimenting to compare two treatments applied to test subjects. For example, many IT companies frequently conduct A/B tests on their users who are connected and form social networks. Often, the users' responses could be related to the network connection. In this paper, we assume that the users, or the test subjects of the experiments, are connected on an undirected network, and the responses of two connected users are correlated. We include the treatment assignment, covariate features, and network connection in a conditional autoregressive model. Based on this model, we propose a design criterion that measures the variance of the estimated treatment effect and allocate the treatment settings to the test subjects by minimizing the criterion. Since the design criterion depends on an unknown network correlation parameter, we adopt the locally optimal design method and develop a hybrid optimization approach to obtain the optimal design. Through synthetic and real social network examples, we demonstrate the value of including network dependence in designing A/B experiments and validate that the proposed locally optimal design is robust to the choices of parameters.

stat.ME↗

arXiv2021

A growth model for water distribution networks with loops

Water distribution networks (WDNs) expand their service areas over time. These growth dynamics are poorly understood. One facet of WDNs is that they have loops in general, and closing loops may be a functionally important process for enhancing their robustness and efficiency. We propose a growth model for WDNs that generates networks with loops and is applicable to networks with multiple water sources. We apply the proposed model to four empirical WDNs to show that it produces networks whose structure is similar to that of the empirical WDNs. The comparison between the empirical and modelled WDNs suggests that the empirical WDNs may realize a reasonable balance between cost, efficiency and robustness in terms of the network structure. We also study the design of pipe diameters based on a biological positive feedback mechanism. Specifically, we apply a model inspired by Physarum polycephalum to find moderate positive correlations between the empirical and modelled pipe diameters. The difference between the empirical and modelled pipe diameters suggests that we may be able to improve the performance of WDNs by following organizing principles of biological flow networks.

physics.soc-ph↗

arXiv2021

Sharp Stability of a String with Local Degenerate Kelvin-Voigt Damping

This paper is on the asymptotic behavior of the elastic string equation with localized degenerate Kelvin--Voigt damping $$ u_{tt}(x,t)-[u_{x}(x,t)+b(x)u_{x,t}(x,t)]_{x}=0,\; x\in(-1,1),\; t>0,$$ where $b(x)=0$ on $x\in (-1,0]$, and $b(x)=x^α>0$ on $x\in (0,1)$ for $α\in(0,1)$. It is known that the optimal decay rate of solution is $t^{-2}$ in the limit case $α=0$, and exponential decay rate for $α\ge 1$. When $α\in (0,1)$, the damping coefficient $b(x)$ is continuous, but its derivative has a singularity at the interface $x=0$. In this case, the best known decay rate is $t^{-\frac{3-α}{2(1-α)}}$. Although this rate is consistent with the exponential one at $α=1$, it failed to match the optimal one at $α=0$. In this paper, we obtain a sharper polynomial decay rate $t^{-\frac{2-α}{1-α}}$. More significantly, it is consistent with the optimal polynomial decay rate at $α=0$ and the exponential decay rate at $α= 1$.This is a big step toward the goal of obtaining eventually the optimal decay rate.

math.OC↗

arXiv2021

Distributed Learning of Finite Gaussian Mixtures

Advances in information technology have led to extremely large datasets that are often kept in different storage centers. Existing statistical methods must be adapted to overcome the resulting computational obstacles while retaining statistical validity and efficiency. Split-and-conquer approaches have been applied in many areas, including quantile processes, regression analysis, principal eigenspaces, and exponential families. We study split-and-conquer approaches for the distributed learning of finite Gaussian mixtures. We recommend a reduction strategy and develop an effective MM algorithm. The new estimator is shown to be consistent and retains root-n consistency under some general conditions. Experiments based on simulated and real-world data show that the proposed split-and-conquer approach has comparable statistical performance with the global estimator based on the full dataset, if the latter is feasible. It can even slightly outperform the global estimator if the model assumption does not match the real-world data. It also has better statistical and computational performance than some existing methods.

stat.ME↗

arXiv2021

Boosting the Generalization Capability in Cross-Domain Few-shot Learning via Noise-enhanced Supervised Autoencoder

State of the art (SOTA) few-shot learning (FSL) methods suffer significant performance drop in the presence of domain differences between source and target datasets. The strong discrimination ability on the source dataset does not necessarily translate to high classification accuracy on the target dataset. In this work, we address this cross-domain few-shot learning (CDFSL) problem by boosting the generalization capability of the model. Specifically, we teach the model to capture broader variations of the feature distributions with a novel noise-enhanced supervised autoencoder (NSAE). NSAE trains the model by jointly reconstructing inputs and predicting the labels of inputs as well as their reconstructed pairs. Theoretical analysis based on intra-class correlation (ICC) shows that the feature embeddings learned from NSAE have stronger discrimination and generalization abilities in the target domain. We also take advantage of NSAE structure and propose a two-step fine-tuning procedure that achieves better adaption and improves classification performance in the target domain. Extensive experiments and ablation studies are conducted to demonstrate the effectiveness of the proposed method. Experimental results show that our proposed method consistently outperforms SOTA methods under various conditions.

cs.CV↗

arXiv2021

Minimum Wasserstein Distance Estimator under Finite Location-scale Mixtures

When a population exhibits heterogeneity, we often model it via a finite mixture: decompose it into several different but homogeneous subpopulations. Contemporary practice favors learning the mixtures by maximizing the likelihood for statistical efficiency and the convenient EM-algorithm for numerical computation. Yet the maximum likelihood estimate (MLE) is not well defined for the most widely used finite normal mixture in particular and for finite location-scale mixture in general. We hence investigate feasible alternatives to MLE such as minimum distance estimators. Recently, the Wasserstein distance has drawn increased attention in the machine learning community. It has intuitive geometric interpretation and is successfully employed in many new applications. Do we gain anything by learning finite location-scale mixtures via a minimum Wasserstein distance estimator (MWDE)? This paper investigates this possibility in several respects. We find that the MWDE is consistent and derive a numerical solution under finite location-scale mixtures. We study its robustness against outliers and mild model mis-specifications. Our moderate scaled simulation study shows the MWDE suffers some efficiency loss against a penalized version of MLE in general without noticeable gain in robustness. We reaffirm the general superiority of the likelihood based learning strategies even for the non-regular finite location-scale mixtures.

stat.ML↗