arXiv subjects

Wen Dong

Publications and source records attributed to Wen Dong.

At least 19 recordsLinked to original sources

arXiv2026

TD-VAD: Breaking Visual Dependence in Video Anomaly Detection with Text-Driven Learning

Visual data is typically a prerequisite for training existing video anomaly detection (VAD) methods. However, obtaining sufficient annotated anomaly data for training is challenging and not scalable due to the rarity of anomaly data and the wide variety of abnormal events. In this work, we advocate that the effectiveness of treating texts as video sequences for the VAD model and propose a novel Text-Driven Video Anomaly Detection (TD-VAD) approach to break visual dependence. In contrast to the anomaly video data, text descriptions of abnormal events are easy to collect, and their class labels can be directly derived. Specifically, our method utilizes video-like text descriptions with temporal characteristics generated by LLM to train a VAD model, without any reliance on target-domain anomaly data. To capture the long- and short-range temporal logic of events, we design the event evolution causal attention module to model contextual dependencies across time. During inference, considering the domain gap between the texts and video sequences, we use the frozen CLIP encoder to extract embeddings of video frames to align the text modality while retaining crucial visual information. Comprehensive experiments on two large-scale VAD datasets, XD-Violence and UCF-Crime, demonstrate that our method outperforms prior one-class and unsupervised VAD methods by a large margin.

cs.CV↗

arXiv2026

CL-Anomaly: Layer-Adaptive Mixture-of-Experts with Multimodal Large Language Model for Continual Learning in Anomaly Detection

Multimodal Large Language Models (MLLMs) excel in diverse vision tasks, but full-parameter retraining is computationally expensive as real-world knowledge evolves. Existing continual learning methods often suffer from semantic entanglement in parameter spaces across tasks, impeding the continuous deployment of models. This challenge is especially pronounced in Anomaly Detection (AD), which exhibits triple heterogeneity across modalities, domains, and defect scale variability, significantly complicating multi-task knowledge transfer. In this paper, we propose CL-Anomaly, a parameter-efficient fine-tuning framework based on an isolation-sharing collaboration to enable continual learning for anomaly detection with MLLMs. We introduce the task-private expert PrivLoRA, which physically isolates task-specific subspaces in the parameter space to prevent semantic entanglement of anomaly knowledge in diverse scenarios. The Layer-Adaptive Shared Experts maintain cross-task representations within a unified feature space, enabling knowledge sharing between previous and new tasks. Furthermore, we propose a Layer-Adaptive Knowledge Transfer strategy that automatically selects and dynamically updates the layer-wise key shared experts of each task via a momentum-based mechanism, promoting effective knowledge transfer across related anomaly detection tasks. Extensive experiments across three continual learning scenarios for anomaly detection, including class-incremental, cross-domain, and cross-modal, demonstrate that CL-Anomaly outperforms state-of-the-art methods. Code is available at https://github.com/WenDongyp/CL-Anomaly.

cs.CV↗

arXiv2024

Mittag-Leffler stability of complete monotonicity-preserving schemes for time-dependent coefficients sub-diffusion equations

A key characteristic of the anomalous sub-solution equation is that the solution exhibits algebraic decay rate over long time intervals, which is often refered to the Mittag-Leffler type stability. For a class of power nonlinear sub-diffusion models with variable coefficients, we prove that their solutions have Mittag-Leffler stability when the source functions satisfy natural decay assumptions. That is the solutions have the decay rate $\|u(t)\|_{L^{s}(Ω)}=O\left( t^{-(α+β)/γ} \right)$ as $t\rightarrow\infty$, where $α$, $γ$ are positive constants, $β\in(-α,\infty)$ and $s\in (1,\infty)$. Then we develop the structure preserving algorithm for this type of model. For the complete monotonicity-preserving ($\mathcal{CM}$-preserving) schemes developed by Li and Wang (Commun. Math. Sci., 19(5):1301-1336, 2021), we prove that they satisfy the discrete comparison principle for time fractional differential equations with variable coefficients. Then, by carefully constructing the fine the discrete supsolution and subsolution, we obtain the long time optimal decay rate of the numerical solution $\|u_{n}\|_{L^{s}(Ω)}=O\left( t_n^{-(α+β)/γ} \right)$ as $t_{n}\rightarrow\infty$, which is fully agree with the theoretical solution. Finally, we validated the analysis results through numerical experiments.

math.NA↗

arXiv2024

Learning Individual Interactions from Population Dynamics with Discrete-Event Simulation Model

The abundance of data affords researchers to pursue more powerful computational tools to learn the dynamics of complex system, such as neural networks, engineered systems and social networks. Traditional machine learning approaches capture complex system dynamics either with dynamic Bayesian networks and state space models, which is hard to scale because it is non-trivial to prescribe the dynamics with a sparse graph or a system of differential equations; or a deep neural networks, where the distributed representation of the learned dynamics is hard to interpret. In this paper, we will explore the possibility of learning a discrete-event simulation representation of complex system dynamics assuming multivariate normal distribution of the state variables, based on the observation that many complex system dynamics can be decomposed into a sequence of local interactions, which individually change the system state only minimally but in sequence generate complex and diverse dynamics. Our results show that the algorithm can data-efficiently capture complex network dynamics in several fields with meaningful events.

cs.LG↗

arXiv2024

Exploiting Polarized Material Cues for Robust Car Detection

Car detection is an important task that serves as a crucial prerequisite for many automated driving functions. The large variations in lighting/weather conditions and vehicle densities of the scenes pose significant challenges to existing car detection algorithms to meet the highly accurate perception demand for safety, due to the unstable/limited color information, which impedes the extraction of meaningful/discriminative features of cars. In this work, we present a novel learning-based car detection method that leverages trichromatic linear polarization as an additional cue to disambiguate such challenging cases. A key observation is that polarization, characteristic of the light wave, can robustly describe intrinsic physical properties of the scene objects in various imaging conditions and is strongly linked to the nature of materials for cars (e.g., metal and glass) and their surrounding environment (e.g., soil and trees), thereby providing reliable and discriminative features for robust car detection in challenging scenes. To exploit polarization cues, we first construct a pixel-aligned RGB-Polarization car detection dataset, which we subsequently employ to train a novel multimodal fusion network. Our car detection network dynamically integrates RGB and polarization features in a request-and-complement manner and can explore the intrinsic material properties of cars across all learning samples. We extensively validate our method and demonstrate that it outperforms state-of-the-art detection methods. Experimental results show that polarization is a powerful cue for car detection.

cs.CV↗

arXiv2022

Progressive Glass Segmentation

Glass is very common in the real world. Influenced by the uncertainty about the glass region and the varying complex scenes behind the glass, the existence of glass poses severe challenges to many computer vision tasks, making glass segmentation as an important computer vision task. Glass does not have its own visual appearances but only transmit/reflect the appearances of its surroundings, making it fundamentally different from other common objects. To address such a challenging task, existing methods typically explore and combine useful cues from different levels of features in the deep network. As there exists a characteristic gap between level-different features, i.e., deep layer features embed more high-level semantics and are better at locating the target objects while shallow layer features have larger spatial sizes and keep richer and more detailed low-level information, fusing these features naively thus would lead to a sub-optimal solution. In this paper, we approach the effective features fusion towards accurate glass segmentation in two steps. First, we attempt to bridge the characteristic gap between different levels of features by developing a Discriminability Enhancement (DE) module which enables level-specific features to be a more discriminative representation, alleviating the features incompatibility for fusion. Second, we design a Focus-and-Exploration Based Fusion (FEBF) module to richly excavate useful information in the fusion process by highlighting the common and exploring the difference between level-different features.

cs.CV↗

arXiv2022

Unsupervised learning of ferroic variants from atomically resolved STEM images

An approach for the analysis of atomically resolved scanning transmission electron microscopy data with multiple ferroic variants in the presence of imaging non-idealities and chemical variabilities based on a rotationally invariant variational autoencoder (rVAE) is presented. We show that an optimal local descriptor for the analysis is a sub-image centered at specific atomic units, since materials and microscope distortions preclude the use of an ideal lattice as a reference point. The applicability of unsupervised clustering and dimensionality reduction methods is explored and are shown to produce clusters dominated by chemical and microscope effects, with a large number of classes required to establish the presence of rotational variants. Comparatively, the rVAE allows extraction of the angle corresponding to the orientation of ferroic variants explicitly, enabling straightforward identification of the ferroic variants as regions with constant or smoothly changing latent variables and sharp orientational changes. This approach allows further exploration of the chemical variability by separating the rotational degrees of freedom via rVAE and searching for remaining variability in the system. The code used in the manuscript is available at https://github.com/saimani5/ferroelectric_domains_rVAE.

cond-mat.mtrl-sci↗

arXiv2021

Stable and Efficient Policy Evaluation

Policy evaluation algorithms are essential to reinforcement learning due to their ability to predict the performance of a policy. However, there are two long-standing issues lying in this prediction problem that need to be tackled: off-policy stability and on-policy efficiency. The conventional temporal difference (TD) algorithm is known to perform very well in the on-policy setting, yet is not off-policy stable. On the other hand, the gradient TD and emphatic TD algorithms are off-policy stable, but are not on-policy efficient. This paper introduces novel algorithms that are both off-policy stable and on-policy efficient by using the oblique projection method. The empirical experimental results on various domains validate the effectiveness of the proposed approach.

cs.LG↗

arXiv2021

TDM: Trustworthy Decision-Making via Interpretability Enhancement

Human-robot interactive decision-making is increasingly becoming ubiquitous, and trust is an influential factor in determining the reliance on autonomy. However, it is not reasonable to trust systems that are beyond our comprehension, and typical machine learning and data-driven decision-making are black-box paradigms that impede interpretability. Therefore, it is critical to establish computational trustworthy decision-making mechanisms enhanced by interpretability-aware strategies. To this end, we propose a Trustworthy Decision-Making (TDM) framework, which integrates symbolic planning into sequential decision-making. The framework learns interpretable subtasks that result in a complex, higher-level composite task that can be formally evaluated using the proposed trust metric. TDM enables the subtask-level interpretability by design and converges to an optimal symbolic plan from the learned subtasks. Moreover, a TDM-based algorithm is introduced to demonstrate the unification of symbolic planning with other sequential-decision making algorithms, reaping the benefits of both. Experimental results validate the effectiveness of trust-score-based planning while improving the interpretability of subtasks.

cs.LG↗

arXiv2021

Optimizing city-scale traffic through modeling observations of vehicle movements

The capability of traffic-information systems to sense the movement of millions of users and offer trip plans through mobile phones has enabled a new way of optimizing city traffic dynamics, turning transportation big data into insights and actions in a closed-loop and evaluating this approach in the real world. Existing research has applied dynamic Bayesian networks and deep neural networks to make traffic predictions from floating car data, utilized dynamic programming and simulation approaches to identify how people normally travel with dynamic traffic assignment for policy research, and introduced Markov decision processes and reinforcement learning to optimally control traffic signals. However, none of these works utilized floating car data to suggest departure times and route choices in order to optimize city traffic dynamics. In this paper, we present a study showing that floating car data can lead to lower average trip time, higher on-time arrival ratio, and higher Charypar-Nagel score compared with how people normally travel. The study is based on optimizing a partially observable discrete-time decision process and is evaluated in one synthesized scenario, one partly synthesized scenario, and three real-world scenarios. This study points to the potential of a "living lab" approach where we learn, predict, and optimize behaviors in the real world.

cs.SI↗

arXiv2021

Transformer-based Conditional Variational Autoencoder for Controllable Story Generation

We investigate large-scale latent variable models (LVMs) for neural story generation -- an under-explored application for open-domain long text -- with objectives in two threads: generation effectiveness and controllability. LVMs, especially the variational autoencoder (VAE), have achieved both effective and controllable generation through exploiting flexible distributional latent representations. Recently, Transformers and its variants have achieved remarkable effectiveness without explicit latent representation learning, thus lack satisfying controllability in generation. In this paper, we advocate to revive latent variable modeling, essentially the power of representation learning, in the era of Transformers to enhance controllability without hurting state-of-the-art generation effectiveness. Specifically, we integrate latent representation vectors with a Transformer-based pre-trained architecture to build conditional variational autoencoder (CVAE). Model components such as encoder, decoder and the variational posterior are all built on top of pre-trained language models -- GPT2 specifically in this paper. Experiments demonstrate state-of-the-art conditional generation ability of our model, as well as its excellent representation learning capability and controllability.

cs.CL↗

arXiv2021

Outline to Story: Fine-grained Controllable Story Generation from Cascaded Events

Large-scale pretrained language models have shown thrilling generation capabilities, especially when they generate consistent long text in thousands of words with ease. However, users of these models can only control the prefix of sentences or certain global aspects of generated text. It is challenging to simultaneously achieve fine-grained controllability and preserve the state-of-the-art unconditional text generation capability. In this paper, we first propose a new task named "Outline to Story" (O2S) as a test bed for fine-grained controllable generation of long text, which generates a multi-paragraph story from cascaded events, i.e. a sequence of outline events that guide subsequent paragraph generation. We then create dedicate datasets for future benchmarks, built by state-of-the-art keyword extraction techniques. Finally, we propose an extremely simple yet strong baseline method for the O2S task, which fine tunes pre-trained language models on augmented sequences of outline-story pairs with simple language modeling objective. Our method does not introduce any new parameters or perform any architecture modification, except several special tokens as delimiters to build augmented sequences. Extensive experiments on various datasets demonstrate state-of-the-art conditional story generation performance with our model, achieving better fine-grained controllability and user flexibility. Our paper is among the first ones by our knowledge to propose a model and to create datasets for the task of "outline to story". Our work also instantiates research interest of fine-grained controllable generation of open-domain long text, where controlling inputs are represented by short text.

cs.CL↗

arXiv2020

Unsupervised Community Detection with a Potts Model Hamiltonian, an Efficient Algorithmic Solution, and Applications in Digital Pathology

Unsupervised segmentation of large images using a Potts model Hamiltonian is unique in that segmentation is governed by a resolution parameter which scales the sensitivity to small clusters. Here, the input image is first modeled as a graph, which is then segmented by minimizing a Hamiltonian cost function defined on the graph and the respective segments. However, there exists no closed form solution of this optimization, and using previous iterative algorithmic solution techniques, the problem scales quadratically in the Input Length. Therefore, while Potts model segmentation gives accurate segmentation, it is grossly underutilized as an unsupervised learning technique. We propose a fast statistical down-sampling of input image pixels based on the respective color features, and a new iterative method to minimize the Potts model energy considering pixel to segment relationship. This method is generalizable and can be extended for image pixel texture features as well as spatial features. We demonstrate that this new method is highly efficient, and outperforms existing methods for Potts model based image segmentation. We demonstrate the application of our method in medical microscopy image segmentation; particularly, in segmenting renal glomerular micro-environment in renal pathology. Our method is not limited to image segmentation, and can be extended to any image/data segmentation/clustering task for arbitrary datasets with discrete features.

cs.CV↗

arXiv2019

Implicit Deep Latent Variable Models for Text Generation

Deep latent variable models (LVM) such as variational auto-encoder (VAE) have recently played an important role in text generation. One key factor is the exploitation of smooth latent structures to guide the generation. However, the representation power of VAEs is limited due to two reasons: (1) the Gaussian assumption is often made on the variational posteriors; and meanwhile (2) a notorious "posterior collapse" issue occurs. In this paper, we advocate sample-based representations of variational distributions for natural language, leading to implicit latent features, which can provide flexible representation power compared with Gaussian-based posteriors. We further develop an LVM to directly match the aggregated posterior to the prior. It can be viewed as a natural extension of VAEs with a regularization of maximizing mutual information, mitigating the "posterior collapse" issue. We demonstrate the effectiveness and versatility of our models in various text generation scenarios, including language modeling, unaligned style transfer, and dialog response generation. The source code to reproduce our experimental results is available on GitHub.

cs.LG↗

arXiv2019

Optimal Control of Complex Systems through Variational Inference with a Discrete Event Decision Process

Complex social systems are composed of interconnected individuals whose interactions result in group behaviors. Optimal control of a real-world complex system has many applications, including road traffic management, epidemic prevention, and information dissemination. However, such real-world complex system control is difficult to achieve because of high-dimensional and non-linear system dynamics, and the exploding state and action spaces for the decision maker. Prior methods can be divided into two categories: simulation-based and analytical approaches. Existing simulation approaches have high-variance in Monte Carlo integration, and the analytical approaches suffer from modeling inaccuracy. We adopted simulation modeling in specifying the complex dynamics of a complex system, and developed analytical solutions for searching optimal strategies in a complex network with high-dimensional state-action space. To capture the complex system dynamics, we formulate the complex social network decision making problem as a discrete event decision process. To address the curse of dimensionality and search in high-dimensional state action spaces in complex systems, we reduce control of a complex system to variational inference and parameter learning, introduce Bethe entropy approximation, and develop an expectation propagation algorithm. Our proposed algorithm leads to higher system expected rewards, faster convergence, and lower variance of value function in a real-world transportation scenario than state-of-the-art analytical and sampling approaches.

eess.SY↗

arXiv2019

PocketCare: Tracking the Flu with Mobile Phones using Partial Observations of Proximity and Symptoms

Mobile phones provide a powerful sensing platform that researchers may adopt to understand proximity interactions among people and the diffusion, through these interactions, of diseases, behaviors, and opinions. However, it remains a challenge to track the proximity-based interactions of a whole community and then model the social diffusion of diseases and behaviors starting from the observations of a small fraction of the volunteer population. In this paper, we propose a novel approach that tries to connect together these sparse observations using a model of how individuals interact with each other and how social interactions happen in terms of a sequence of proximity interactions. We apply our approach to track the spreading of flu in the spatial-proximity network of a 3000-people university campus by mobilizing 300 volunteers from this population to monitor nearby mobile phones through Bluetooth scanning and to daily report flu symptoms about and around them. Our aim is to predict the likelihood for an individual to get flu based on how often her/his daily routine intersects with those of the volunteers. Thus, we use the daily routines of the volunteers to build a model of the volunteers as well as of the non-volunteers. Our results show that we can predict flu infection two weeks ahead of time with an average precision from 0.24 to 0.35 depending on the amount of information. This precision is six to nine times higher than with a random guess model. At the population level, we can predict infectious population in a two-week window with an r-squared value of 0.95 (a random-guess model obtains an r-squared value of 0.2). These results point to an innovative approach for tracking individuals who have interacted with people showing symptoms, allowing us to warn those in danger of infection and to inform health researchers about the progression of contact-induced diseases.

cs.SI↗

arXiv2017

O$^2$TD: (Near)-Optimal Off-Policy TD Learning

Temporal difference learning and Residual Gradient methods are the most widely used temporal difference based learning algorithms; however, it has been shown that none of their objective functions is optimal w.r.t approximating the true value function $V$. Two novel algorithms are proposed to approximate the true value function $V$. This paper makes the following contributions: (1) A batch algorithm that can help find the approximate optimal off-policy prediction of the true value function $V$. (2) A linear computational cost (per step) near-optimal algorithm that can learn from a collection of off-policy samples. (3) A new perspective of the emphatic temporal difference learning which bridges the gap between off-policy optimality and off-policy stability.

cs.LG↗

arXiv2016

Using Social Dynamics to Make Individual Predictions: Variational Inference with a Stochastic Kinetic Model

Social dynamics is concerned primarily with interactions among individuals and the resulting group behaviors, modeling the temporal evolution of social systems via the interactions of individuals within these systems. In particular, the availability of large-scale data from social networks and sensor networks offers an unprecedented opportunity to predict state-changing events at the individual level. Examples of such events include disease transmission, opinion transition in elections, and rumor propagation. Unlike previous research focusing on the collective effects of social systems, this study makes efficient inferences at the individual level. In order to cope with dynamic interactions among a large number of individuals, we introduce the stochastic kinetic model to capture adaptive transition probabilities and propose an efficient variational inference algorithm the complexity of which grows linearly --- rather than exponentially --- with the number of individuals. To validate this method, we have performed epidemic-dynamics experiments on wireless sensor network data collected from more than ten thousand people over three years. The proposed algorithm was used to track disease transmission and predict the probability of infection for each individual. Our results demonstrate that this method is more efficient than sampling while nonetheless achieving high accuracy.

stat.ML↗