arXiv subjects

Fan Zhou

Publications and source records attributed to Fan Zhou.

At least 145 records · Page 8Linked to original sources

arXiv2022

Forgetting Prevention for Cross-regional Fraud Detection with Heterogeneous Trade Graph

With the booming growth of e-commerce, detecting financial fraud has become an urgent task to avoid transaction risks. Despite the successful applications of Graph Neural Networks (GNNs) in fraud detection, the existing solutions are only suitable for a narrow scope due to the limitation in data collection. Especially when expanding a business into new territory, e.g., new cities or new countries, developing a totally new model will bring the cost issue and result in forgetting previous knowledge. Moreover, recent works strive to devise GNNs to expose the implicit interactions behind financial transactions. However, most existing GNNs-based solutions concentrate on either homogeneous graphs or decomposing heterogeneous interactions into several homogeneous connections for convenience. To this end, this study proposes a novel solution based on heterogeneous trade graphs, namely HTG-CFD, to prevent knowledge forgetting of cross-regional fraud detection. In particular, the heterogeneous trade graph (HTG) is meticulously constructed from original transaction records to explore the complex semantics among different types of entities and relationships. And motivated by recent continual learning, we present a practical and task-oriented forgetting prevention method to alleviate knowledge forgetting in the context of cross-regional detection. Extensive experiments demonstrate that the proposed HTG-CFD not only promotes the performance in cross-regional scenarios but also significantly contributes to single-regional fraud detection.

cs.CE↗

arXiv2022

Table Pre-training: A Survey on Model Architectures, Pre-training Objectives, and Downstream Tasks

Since a vast number of tables can be easily collected from web pages, spreadsheets, PDFs, and various other document types, a flurry of table pre-training frameworks have been proposed following the success of text and images, and they have achieved new state-of-the-arts on various tasks such as table question answering, table type recognition, column relation classification, table search, formula prediction, etc. To fully use the supervision signals in unlabeled tables, a variety of pre-training objectives have been designed and evaluated, for example, denoising cell values, predicting numerical relationships, and implicitly executing SQLs. And to best leverage the characteristics of (semi-)structured tables, various tabular language models, particularly with specially-designed attention mechanisms, have been explored. Since tables usually appear and interact with free-form text, table pre-training usually takes the form of table-text joint pre-training, which attracts significant research interests from multiple domains. This survey aims to provide a comprehensive review of different model designs, pre-training objectives, and downstream tasks for table pre-training, and we further share our thoughts and vision on existing challenges and future opportunities.

cs.CL↗

arXiv2022

Domain Generalization via Optimal Transport with Metric Similarity Learning

Generalizing knowledge to unseen domains, where data and labels are unavailable, is crucial for machine learning models. We tackle the domain generalization problem to learn from multiple source domains and generalize to a target domain with unknown statistics. The crucial idea is to extract the underlying invariant features across all the domains. Previous domain generalization approaches mainly focused on learning invariant features and stacking the learned features from each source domain to generalize to a new target domain while ignoring the label information, which will lead to indistinguishable features with an ambiguous classification boundary. For this, one possible solution is to constrain the label-similarity when extracting the invariant features and to take advantage of the label similarities for class-specific cohesion and separation of features across domains. Therefore we adopt optimal transport with Wasserstein distance, which could constrain the class label similarity, for adversarial training and also further deploy a metric learning objective to leverage the label information for achieving distinguishable classification boundary. Empirical results show that our proposed method could outperform most of the baselines. Furthermore, ablation studies also demonstrate the effectiveness of each component of our method.

cs.CV↗

arXiv2022

Automatic Registration of Images with Inconsistent Content Through Line-Support Region Segmentation and Geometrical Outlier Removal

The implementation of automatic image registration is still difficult in various applications. In this paper, an automatic image registration approach through line-support region segmentation and geometrical outlier removal (ALRS-GOR) is proposed. This new approach is designed to address the problems associated with the registration of images with affine deformations and inconsistent content, such as remote sensing images with different spectral content or noise interference, or map images with inconsistent annotations. To begin with, line-support regions, namely a straight region whose points share roughly the same image gradient angle, are extracted to address the issues of inconsistent content existing in images. To alleviate the incompleteness of line segments, an iterative strategy with multi-resolution is employed to preserve global structures that are masked at full resolution by image details or noise. Then, Geometrical Outlier Removal (GOR) is developed to provide reliable feature point matching, which is based on affineinvariant geometrical classifications for corresponding matches initialized by SIFT. The candidate outliers are selected by comparing the disparity of accumulated classifications among all matches, instead of conventional methods which only rely on local geometrical relations. Various image sets have been considered in this paper for the evaluation of the proposed approach, including aerial images with simulated affine deformations, remote sensing optical and synthetic aperture radar images taken at different situations (multispectral, multisensor, and multitemporal), and map images with inconsistent annotations. Experimental results demonstrate the superior performance of the proposed method over the existing approaches for the whole data set.

eess.IV↗

arXiv2022

CCGL: Contrastive Cascade Graph Learning

Supervised learning, while prevalent for information cascade modeling, often requires abundant labeled data in training, and the trained model is not easy to generalize across tasks and datasets. It often learns task-specific representations, which can easily result in overfitting for downstream tasks. Recently, self-supervised learning is designed to alleviate these two fundamental issues in linguistic and visual tasks. However, its direct applicability for information cascade modeling, especially graph cascade related tasks, remains underexplored. In this work, we present Contrastive Cascade Graph Learning (CCGL), a novel framework for information cascade graph learning in a contrastive, self-supervised, and task-agnostic way. In particular, CCGL first designs an effective data augmentation strategy to capture variation and uncertainty by simulating the information diffusion in graphs. Second, it learns a generic model for graph cascade tasks via self-supervised contrastive pre-training using both unlabeled and labeled data. Third, CCGL learns a task-specific cascade model via fine-tuning using labeled data. Finally, to make the model transferable across datasets and cascade applications, CCGL further enhances the model via distillation using a teacher-student architecture. We demonstrate that CCGL significantly outperforms its supervised and semi-supervised counterparts for several downstream tasks.

cs.SI↗

arXiv2021

Construction D' Lattices for Power-Constrained Communications

Designs and methods for nested lattice codes using Construction D' lattices for coding and convolutional code lattices for shaping are described. Two encoding methods and a decoding algorithm for Construction D' coding lattices that can be used with shaping lattices for power-constrained channels are given. We construct nested lattice codes with good coding properties, a high shaping gain, and low-complexity encoding and decoding. Convolutional code generator polynomials for Construction A lattices with the greatest shaping gain are given, as a result of an extensive search. It is shown that rate 1/3 convolutional codes provide a more favorable performance-complexity trade-off than rate 1/2 convolutional codes. Tail-biting convolutional codes have higher shaping gain than that of zero-tailed convolutional codes. A design for quasi-cyclic low-density parity-check (QC-LDPC) codes to form Construction D' lattices which have efficient encoding and indexing is presented. The resulting QC-LDPC Construction D' lattices are evaluated using four shaping lattices: the $E_8$ lattice, the $BW_{16}$ lattice, the Leech lattice and our best-found convolutional code lattice, showing a shaping gain of approximately 0.65 dB, 0.86 dB, 1.03 dB and 1.25 dB at dimension 2304.

cs.IT↗

arXiv2021

High-Order Statistical Functional Expansion and Its Application To Some Nonsmooth Problems

Let $\bx_j = \btheta +\bep_j, j=1,...,n$, be observations of an unknown parameter $\btheta$ in a Euclidean or separable Hilbert space $\scrH$, where $\bep_j$ are noises as random elements in $\scrH$ from a general distribution. We study the estimation of $f(\btheta)$ for a given functional $f:\scrH\rightarrow \RR$ based on $\bx_j$'s. The key element of our approach is a new method which we call High-Order Degenerate Statistical Expansion. It leverages the use of classical multivariate Taylor expansion and degenerate $U$-statistic and yields an elegant explicit formula. In the univariate case of $\scrH=\R$, the formula expresses the error of the proposed estimator as a sum of order $k$ degenerate $U$-products of the noises with coefficient $f^{(k)}(\btheta)/k!$ and an explicit remainder term in the form of the Riemann-Liouville integral as in the Taylor expansion around the true $\btheta$. For general $\scrH$, the formula expresses the estimation error in terms of the inner product of $f^{(k)}(\btheta)/k!$ and the average of the tensor products of $k$ noises with distinct indices and a parallel extension of the remainder term from the univariate case. This makes the proposed method a natural statistical version of the classical Taylor expansion. The proposed estimator can be viewed as a jackknife estimator of an ideal degenerate expansion of $f(\cdot)$ around the true $\btheta$ with the degenerate $U$-product of the noises, and can be approximated by bootstrap. Thus, the jackknife, bootstrap and Taylor expansion approaches all converge to the proposed estimator. We develop risk bounds for the proposed estimator and a central limit theorem under a second moment condition (even in expansions of higher than the second order). We apply this new method to generalize several existing results with smooth and nonsmooth $f$ to universal $\bep_j$'s with only minimum moment constraints.

math.ST↗

arXiv2021

Non-decreasing Quantile Function Network with Efficient Exploration for Distributional Reinforcement Learning

Although distributional reinforcement learning (DRL) has been widely examined in the past few years, there are two open questions people are still trying to address. One is how to ensure the validity of the learned quantile function, the other is how to efficiently utilize the distribution information. This paper attempts to provide some new perspectives to encourage the future in-depth studies in these two fields. We first propose a non-decreasing quantile function network (NDQFN) to guarantee the monotonicity of the obtained quantile estimates and then design a general exploration framework called distributional prediction error (DPE) for DRL which utilizes the entire distribution of the quantile function. In this paper, we not only discuss the theoretical necessity of our method but also show the performance gain it achieves in practice by comparing with some competitors on Atari 2600 Games especially in some hard-explored games.

cs.LG↗

arXiv2021

A Survey of Information Cascade Analysis: Models, Predictions, and Recent Advances

The deluge of digital information in our daily life -- from user-generated content, such as microblogs and scientific papers, to online business, such as viral marketing and advertising -- offers unprecedented opportunities to explore and exploit the trajectories and structures of the evolution of information cascades. Abundant research efforts, both academic and industrial, have aimed to reach a better understanding of the mechanisms driving the spread of information and quantifying the outcome of information diffusion. This article presents a comprehensive review and categorization of information popularity prediction methods, from feature engineering and stochastic processes, through graph representation, to deep learning-based approaches. Specifically, we first formally define different types of information cascades and summarize the perspectives of existing studies. We then present a taxonomy that categorizes existing works into the aforementioned three main groups as well as the main subclasses in each group, and we systematically review cutting-edge research work. Finally, we summarize the pros and cons of existing research efforts and outline the open challenges and opportunities in this field.

cs.SI↗

arXiv2021

Graph-Based Equilibrium Metrics for Dynamic Supply-Demand Systems with Applications to Ride-sourcing Platforms

How to dynamically measure the local-to-global spatio-temporal coherence between demand and supply networks is a fundamental task for ride-sourcing platforms, such as DiDi. Such coherence measurement is critically important for the quantification of the market efficiency and the comparison of different platform policies, such as dispatching. The aim of this paper is to introduce a graph-based equilibrium metric (GEM) to quantify the distance between demand and supply networks based on a weighted graph structure. We formulate GEM as the optimal objective value of an unbalanced transport problem, which can be efficiently solved by optimizing an equivalent linear programming. We examine how the GEM can help solve three operational tasks of ride-sourcing platforms. The first one is that GEM achieves up to 70.6% reduction in root-mean-square error over the second-best distance measurement for the prediction accuracy. The second one is that the use of GEM for designing order dispatching policy increases answer rate and drivers' revenue for more than 1%, representing a huge improvement in number. The third one is that GEM is to serve as an endpoint for comparing different platform policies in AB test.

math.OC↗

arXiv2021

Multi-task Learning by Leveraging the Semantic Information

One crucial objective of multi-task learning is to align distributions across tasks so that the information between them can be transferred and shared. However, existing approaches only focused on matching the marginal feature distribution while ignoring the semantic information, which may hinder the learning performance. To address this issue, we propose to leverage the label information in multi-task learning by exploring the semantic conditional relations among tasks. We first theoretically analyze the generalization bound of multi-task learning based on the notion of Jensen-Shannon divergence, which provides new insights into the value of label information in multi-task learning. Our analysis also leads to a concrete algorithm that jointly matches the semantic distribution and controls label distribution divergence. To confirm the effectiveness of the proposed method, we first compare the algorithm with several baselines on some benchmarks and then test the algorithms under label space shift conditions. Empirical results demonstrate that the proposed method could outperform most baselines and achieve state-of-the-art performance, particularly showing the benefits under the label shift conditions.

cs.LG↗

arXiv2021

Encoding and Decoding Construction D' Lattices for Power-Constrained Communications

This paper focuses on the encoding and decoding of Construction D' coding lattices that can be used with shaping lattices for power-constrained channels. Two encoding methods and a decoding algorithm for Construction D' lattices are given. A design of quasi-cyclic low-density parity-check (QC-LDPC) codes to form Construction D' lattices is presented. This allows construction of nested lattice codes which are good for coding, good for shaping, and have low complexity encoding and decoding. Numerical results of using $E_8$, $BW_{16}$ and Leech lattices for shaping a Construction D' lattice indicate that the shaping gains 0.65 dB, 0.86 dB and 1.03 dB are preserved, respectively.

cs.IT↗

arXiv2021

Overcoming Catastrophic Forgetting in Graph Neural Networks with Experience Replay

Graph Neural Networks (GNNs) have recently received significant research attention due to their superior performance on a variety of graph-related learning tasks. Most of the current works focus on either static or dynamic graph settings, addressing a single particular task, e.g., node/graph classification, link prediction. In this work, we investigate the question: can GNNs be applied to continuously learning a sequence of tasks? Towards that, we explore the Continual Graph Learning (CGL) paradigm and present the Experience Replay based framework ER-GNN for CGL to alleviate the catastrophic forgetting problem in existing GNNs. ER-GNN stores knowledge from previous tasks as experiences and replays them when learning new tasks to mitigate the catastrophic forgetting issue. We propose three experience node selection strategies: mean of feature, coverage maximization, and influence maximization, to guide the process of selecting experience nodes. Extensive experiments on three benchmark datasets demonstrate the effectiveness of our ER-GNN and shed light on the incremental graph (non-Euclidean) structure learning.

cs.LG↗

arXiv2021

Embedding Symbolic Temporal Knowledge into Deep Sequential Models

Sequences and time-series often arise in robot tasks, e.g., in activity recognition and imitation learning. In recent years, deep neural networks (DNNs) have emerged as an effective data-driven methodology for processing sequences given sufficient training data and compute resources. However, when data is limited, simpler models such as logic/rule-based methods work surprisingly well, especially when relevant prior knowledge is applied in their construction. However, unlike DNNs, these "structured" models can be difficult to extend, and do not work well with raw unstructured data. In this work, we seek to learn flexible DNNs, yet leverage prior temporal knowledge when available. Our approach is to embed symbolic knowledge expressed as linear temporal logic (LTL) and use these embeddings to guide the training of deep models. Specifically, we construct semantic-based embeddings of automata generated from LTL formula via a Graph Neural Network. Experiments show that these learnt embeddings can lead to improvements in downstream robot tasks such as sequential action recognition and imitation learning.

cs.AI↗

arXiv2020

A Fourier Analytical Approach to Estimation of Smooth Functions in Gaussian Shift Model

Let $\mathbf{x}_j = \mathbfθ + \mathbfε_j$, $j=1,\dots,n$ be i.i.d. copies of a Gaussian random vector $\mathbf{x}\sim\mathcal{N}(\mathbfθ,\mathbfΣ)$ with unknown mean $\mathbfθ \in \mathbb{R}^d$ and unknown covariance matrix $\mathbfΣ\in \mathbb{R}^{d\times d}$. The goal of this article is to study the estimation of $f(\mathbfθ)$ where $f$ is a given smooth function of which smoothness is characterized by a Besov-type norm. The problem of interest resides in the high dimensional regime where the intrinsic dimension can grow with the sample size $n$. Inspired by the classical work of A. N. Kolmogorov on unbiased estimation and Littlewood-Paley theory, we develop a new estimator based on a Fourier analytical approach that achieves effective bias reduction. Asymptotic normality and efficiency are proved when the smoothness index of $f$ is above certain threshold which was discovered recently by Koltchinskii et. al. (2018) for a Hölder type class. Numerical simulations are presented to validate our analysis. The simplicity of implementation and its superiority over the plug-in approach indicate the new estimator can be applied to a broad range of real world applications.

math.ST↗

arXiv2020

On Eventually Periodic Sets as Minimal Additive Complements

We say a subset $C$ of an abelian group $G$ \textit{arises as a minimal additive complement} if there is some other subset $W$ of $G$ such that $C+W=\{c+w:c\in C,\ w\in W\}=G$ and such that there is no proper subset $C'\supset C$ such that $C'+W=G$. In their recent paper, Burcroff and Luntzlara studied, among many other things, the conditions under which "eventually periodic sets", which are finite unions of infinite (in the positive direction) arithmetic progressions and singletons, arise as minimal additive complements in $\mathbb Z$. In the present paper we shall study this question further. We give, in the form of bounds on the period $m$, some sufficient conditions for an eventually periodic set to be a minimal additive complement; in particular we show that "all eventually periodic sets are eventually minimal additive complements". Moreover, we generalize this to a framework in which "patterns" of points are projected down to $\mathbb Z$, and we show that all sets which arise this way are eventually minimal additive complements. We also introduce a formalism of formal power series, which serves purely as a bookkeeper in writing down proofs. Through our work we are able to answer a question of Burcroff and Luntzlara in a large class of cases.

math.CO↗

arXiv2020

Beyond $\mathcal{H}$-Divergence: Domain Adaptation Theory With Jensen-Shannon Divergence

We reveal the incoherence between the widely-adopted empirical domain adversarial training and its generally-assumed theoretical counterpart based on $\mathcal{H}$-divergence. Concretely, we find that $\mathcal{H}$-divergence is not equivalent to Jensen-Shannon divergence, the optimization objective in domain adversarial training. To this end, we establish a new theoretical framework by directly proving the upper and lower target risk bounds based on joint distributional Jensen-Shannon divergence. We further derive bi-directional upper bounds for marginal and conditional shifts. Our framework exhibits inherent flexibilities for different transfer learning problems, which is usable for various scenarios where $\mathcal{H}$-divergence-based theory fails to adapt. From an algorithmic perspective, our theory enables a generic guideline unifying principles of semantic conditional matching, feature marginal matching, and label marginal shift correction. We employ algorithms for each principle and empirically validate the benefits of our framework on real datasets.

cs.LG↗

arXiv2020

Discriminative Active Learning for Domain Adaptation

Domain Adaptation aiming to learn a transferable feature between different but related domains has been well investigated and has shown excellent empirical performances. Previous works mainly focused on matching the marginal feature distributions using the adversarial training methods while assuming the conditional relations between the source and target domain remained unchanged, $i.e.$, ignoring the conditional shift problem. However, recent works have shown that such a conditional shift problem exists and can hinder the adaptation process. To address this issue, we have to leverage labelled data from the target domain, but collecting labelled data can be quite expensive and time-consuming. To this end, we introduce a discriminative active learning approach for domain adaptation to reduce the efforts of data annotation. Specifically, we propose three-stage active adversarial training of neural networks: invariant feature space learning (first stage), uncertainty and diversity criteria and their trade-off for query strategy (second stage) and re-training with queried target labels (third stage). Empirical comparisons with existing domain adaptation methods using four benchmark datasets demonstrate the effectiveness of the proposed approach.

cs.LG↗