Wei Yi

Suqi Li, Wei Yi, Reza Hoseinnezhad et al.

This paper presents an exact Bayesian filtering solution for the multi-object tracking problem with the generic observation model. The proposed solution is designed in the labeled random finite set framework, using the product styled representation of labeled multi-object densities, with the standard multi-object transition kernel and no particular simplifying assumptions on the multi-object likelihood. Computationally tractable solutions are also devised by applying a principled approximation involving the replacement of the full multi-object density with a labeled multi-Bernoulli density that minimizes the Kullback-Leibler divergence and preserves the first-order moment. To achieve the fast performance, a dynamic grouping procedure based implementation is presented with a step-by-step algorithm. The performance of the proposed filter and its tractable implementations are verified and compared with the state-of-the-art in numerical experiments.

Meng Jiang, Wei Yi, Reza Hoseinnezhad et al.

The paper addresses distributed multi-target tracking in the framework of generalized Covariance Intersection (GCI) over multistatic radar system. The proposed method is based on the unlabeled version of generalized labeled multi-Bernoulli (GLMB) family by discarding the labels, referred as generalized multi-Bernoulli (GMB) family. However, it doesn't permit closed form solution for GCI fusion with GMB family. To solve this challenging problem, firstly, we propose an efficient approximation to the GMB family which preserves both the probability hypothesis density (PHD) and cardinality distribution, named as second-order approximation of GMB (SO-GMB) density. Then, we derive explicit expression for the GCI fusion with SO-GMB density. Finally, we compare the first-order approximation of GMB (FO-GMB) density with SO-GMB density in two scenarios and make a concrete analysis of the advantages of the second-order approximation. Simulation results are presented to verify the proposed approach.

Meng Jiang, Wei Yi, Lingjiang Kong

The paper addresses the problem of multi-sensor control for multi-target tracking via labelled random finite sets (RFS) in the sensor network systems. Based on an information theoretic divergence measure, namely Cauchy-Schwarz (CS) divergence which admits a closed form solution for GLMB densities, we propose two novel multi-sensor control approaches in the framework of generalized Covariance Intersection (GCI). The first joint decision making (JDM) method is optimal and can achieve overall good performance, while the second independent decision making (IDM) method is suboptimal as a fast realization with smaller amount of computations. Simulation in challenging situation is presented to verify the effectiveness of the two proposed approaches.

Yi Wei, Zixin Zhong, Vincent Y. F. Tan

The beam alignment (BA) problem consists in accurately aligning the transmitter and receiver beams to establish a reliable communication link in wireless communication systems. Existing BA methods search the entire beam space to identify the optimal transmit-receive beam pair. This incurs a significant latency when the number of antennas is large. In this work, we develop a bandit-based fast BA algorithm to reduce BA latency for millimeter-wave (mmWave) communications. Our algorithm is named Two-Phase Heteroscedastic Track-and-Stop (2PHT\&S). We first formulate the BA problem as a pure exploration problem in multi-armed bandits in which the objective is to minimize the required number of time steps given a certain fixed confidence level. By taking advantage of the correlation structure among beams that the information from nearby beams is similar and the heteroscedastic property that the variance of the reward of an arm (beam) is related to its mean, the proposed algorithm groups all beams into several beam sets such that the optimal beam set is first selected and the optimal beam is identified in this set after that. Theoretical analysis and simulation results on synthetic and semi-practical channel data demonstrate the clear superiority of the proposed algorithm vis-à-vis other baseline competitors.

Bailu Wang, Wei Yi, Suqi Li et al.

In this paper, we address the problem of the distributed multi-target tracking with labeled set filters in the framework of Generalized Covariance Intersection (GCI). Our analyses show that the label space mismatching (LS-DM) phenomenon, which means the same realization drawn from label spaces of different sensors does not have the same implication, is quite common in practical scenarios and may bring serious problems. Our contributions are two-fold. Firstly, we provide a principled mathematical definition of "label spaces matching (LS-DM)" based on information divergence, which is also referred to as LS-M criterion. Then, to handle the LS-DM, we propose a novel two-step distributed fusion algorithm, named as GCI fusion via label spaces matching (GCI-LSM). The first step is to match the label spaces from different sensors. To this end, we build a ranked assignment problem and design a cost function consistent with LS-M criterion to seek the optimal solution of matching correspondence between label spaces of different sensors. The second step is to perform the GCI fusion on the matched label space. We also derive the GCI fusion with generic labeled multi-object (LMO) densities based on LS-M, which is the foundation of labeled distributed fusion algorithms. Simulation results for Gaussian mixture implementation highlight the performance of the proposed GCI-LSM algorithm in two different tracking scenarios.

Suqi Li, Wei Yi, Bailu Wang et al.

As a fundamental piece of multi-object Bayesian inference, multi-object density has the ability to describe the uncertainty of the number and values of objects, as well as the statistical correlation between objects, thus perfectly matches the behavior of multi-object system. However, it also makes the set integral suffer from the curse of dimensionality and the inherently combinatorial nature of the problem. In this paper, we study the approximations for the universal labeled multi-object (LMO) density and derive several principled approximations including labeled multi-Bernoulli, labeled Poisson and labeled independent identically clustering process based approximations. Also, a detailed analysis on the characteristics (e.g., approximation error and computational complexity) of the proposed approximations is provided. Then some practical suggestions are made for the applications of these approximations based on the preceding analysis and discussion. Finally, an numerical example is given to support our study.

Wei Yi, Suqi Li

Multi-object density is a fundamental descriptor of a point process and has ability to describe the randomness of number and values of objects, as well as the statistical correlation between objects. Due to its comprehensive nature, it usually has a complicate mathematical structure making the set integral suffering from the curse of dimension and the combinatorial nature of the problem. Hence, the approximation of multi-object density is a key research theme in point process theory or finite set statistics (FISST). Conventional approaches usually discard part or all of statistical correlation mechanically in return for computational efficiency, without considering the real situation of correlation between objects. In this paper, we propose an enhanced approximation of labeled multi-object (LMO) density which evaluates the correlation between objects adaptively and factorizes the LMO density into densities of several independent subsets according to the correlation analysis. Besides, to get a tractable factorization of LMO density, we derive the set marginal density of any subset of the universal labeled RFS, the generalized labeled multi-Bernoulli (GLMB) RFS family and its subclasses. The proposed method takes into account the simplification of the complicate structure of LMO density and the reservation of necessary correlation at the same time.

Haiyi Mao, Cong Peng, Yue Liu et al.

A variety of filters with track-before-detect (TBD) strategies have been developed and applied to low signal-to-noise ratio (SNR) scenarios, including the probability hypothesis density (PHD) filter. Assumptions of the standard point measurement model based on detect-before-track (DBT) strategies are not suitable for the amplitude echo model based on TBD strategies. However, based on different models and unmatched assumptions, the measurement update formulas for DBT-PHD filter are just mechanically applied to existing TBD-PHD filters. In this paper, based on the Kullback-Leibler divergence minimization criterion, finite set statistics theory and rigorous Bayes rule, a principled closed-form solution of TBD-PHD filter is derived. Furthermore, we emphasize that PHD filter is conjugated to the Poisson prior based on TBD strategies. Next, a capping operation is devised to handle the divergence of target number estimation as SNR increases. Moreover, the sequential Monte Carlo implementations of dynamic and amplitude echo models are proposed for the radar system. Finally, Monte Carlo experiments exhibit good performance in Rayleigh noise and low SNR scenarios.

Zhenyu Wu, Ziwei Wang, Zibu Wei et al.

Recognizing objects in dense clutter accurately plays an important role to a wide variety of robotic manipulation tasks including grasping, packing, rearranging and many others. However, conventional visual recognition models usually miss objects because of the significant occlusion among instances and causes incorrect prediction due to the visual ambiguity with the high object crowdedness. In this paper, we propose an interactive exploration framework called Smart Explorer for recognizing all objects in dense clutters. Our Smart Explorer physically interacts with the clutter to maximize the recognition performance while minimize the number of motions, where the false positives and negatives can be alleviated effectively with the optimal accuracy-efficiency trade-offs. Specifically, we first collect the multi-view RGB-D images of the clutter and reconstruct the corresponding point cloud. By aggregating the instance segmentation of RGB images across views, we acquire the instance-wise point cloud partition of the clutter through which the existed classes and the number of objects for each class are predicted. The pushing actions for effective physical interaction are generated to sizably reduce the recognition uncertainty that consists of the instance segmentation entropy and multi-view object disagreement. Therefore, the optimal accuracy-efficiency trade-off of object recognition in dense clutter is achieved via iterative instance prediction and physical interaction. Extensive experiments demonstrate that our Smart Explorer acquires promising recognition accuracy with only a few actions, which also outperforms the random pushing by a large margin.

Suhan Guo, Jiahong Deng, Yi Wei et al.

Attention-based architectures have become ubiquitous in time series forecasting tasks, including spatio-temporal (STF) and long-term time series forecasting (LTSF). Yet, our understanding of the reasons for their effectiveness remains limited. In this work, we propose a novel pruning strategy, $\textbf{R}$eplace $\textbf{A}$ttention with $\textbf{M}$LP (RAM), that approximates the attention mechanism using only feedforward layers, residual connections, and layer normalization for temporal and/or spatial modeling in multivariate time series forecasting. Specifically, the Q, K, and V projections, the attention score calculation, the dot-product between the attention score and the V, and the final projection can be removed from the attention-based networks without significantly degrading the performance, so that the given network remains the top-tier compared to other SOTA methods. RAM achieves a $62.579\%$ reduction in FLOPs for spatio-temporal models with less than $2.5\%$ performance drop, and a $42.233\%$ FLOPs reduction for LTSF models with less than $2\%$ performance drop.

Chenyu Zhang, Yuanhang Wu, Xiaoxi Ma et al.

Detection and tracking are the basic tasks of radar systems. Current joint detection tracking methods, which focus on dynamically adjusting detection thresholds from tracking results, still present challenges in fully utilizing the potential of radar signals. These are mainly reflected in the limited capacity of the constant false-alarm rate model to accurately represent information, the insufficient depiction of complex scenes, and the limited information acquired by the tracker. We introduce the Integrated Detection and Tracking based on radar feature (InDT) method, which comprises a network architecture for radar signal detection and a tracker that leverages detection assistance. The InDT detector extracts feature information from each Range-Doppler (RD) matrix and then returns the target position through the feature enhancement module and the detection head. The InDT tracker adaptively updates the measurement noise covariance of the Kalman filter based on detection confidence. The similarity of target RD features is measured by cosine distance, which enhances the data association process by combining location and feature information. Finally, the efficacy of the proposed method was validated through testing on both simulated data and publicly available datasets.

Fuwei Zhang, Zhao Zhang, Xiang Ao et al.

Cross-Lingual Information Retrieval (CLIR) aims to rank the documents written in a language different from the user's query. The intrinsic gap between different languages is an essential challenge for CLIR. In this paper, we introduce the multilingual knowledge graph (KG) to the CLIR task due to the sufficient information of entities in multiple languages. It is regarded as a "silver bullet" to simultaneously perform explicit alignment between queries and documents and also broaden the representations of queries. And we propose a model named CLIR with hierarchical knowledge enhancement (HIKE) for our task. The proposed model encodes the textual information in queries, documents and the KG with multilingual BERT, and incorporates the KG information in the query-document matching process with a hierarchical information fusion mechanism. Particularly, HIKE first integrates the entities and their neighborhood in KG into query representations with a knowledge-level fusion, then combines the knowledge from both source and target languages to further mitigate the linguistic gap with a language-level fusion. Finally, experimental results demonstrate that HIKE achieves substantial improvements over state-of-the-art competitors.

Yongming Rao, Benlin Liu, Yi Wei et al.

3D point cloud understanding has made great progress in recent years. However, one major bottleneck is the scarcity of annotated real datasets, especially compared to 2D object detection tasks, since a large amount of labor is involved in annotating the real scans of a scene. A promising solution to this problem is to make better use of the synthetic dataset, which consists of CAD object models, to boost the learning on real datasets. This can be achieved by the pre-training and fine-tuning procedure. However, recent work on 3D pre-training exhibits failure when transfer features learned on synthetic objects to other real-world applications. In this work, we put forward a new method called RandomRooms to accomplish this objective. In particular, we propose to generate random layouts of a scene by making use of the objects in the synthetic CAD dataset and learn the 3D scene representation by applying object-level contrastive learning on two random scenes generated from the same set of synthetic objects. The model pre-trained in this way can serve as a better initialization when later fine-tuning on the 3D object detection task. Empirically, we show consistent improvement in downstream 3D detection tasks on several base models, especially when less training data are used, which strongly demonstrates the effectiveness and generalization of our method. Benefiting from the rich semantic knowledge and diverse objects from synthetic data, our method establishes the new state-of-the-art on widely-used 3D detection benchmarks ScanNetV2 and SUN RGB-D. We expect our attempt to provide a new perspective for bridging object and scene-level 3D understanding.

Ángel F. García-Fernández, Wei Yi

This paper derives the optimal Bayesian processing of an out-of-sequence (OOS) set of measurements in continuous-time for multiple target tracking. We consider a multi-target system modelled in continuous time that is discretised at the time steps when we receive the measurements, which are distributed according to the standard point target model. All information about this system at the sampled time steps is provided by the posterior density on the set of all trajectories. This density can be computed via the continuous-discrete trajectory Poisson multi-Bernoulli mixture (TPMBM) filter. When we receive an OOS measurement, the optimal Bayesian processing performs a retrodiction step that adds trajectory information at the OOS measurement time stamp followed by an update step. After the OOS measurement update, the posterior remains in TPMBM form. We also provide a computationally lighter alternative based on a trajectory Poisson multi-Bernoulli filter. The effectiveness of the two approaches to handle OOS measurements is evaluated via simulations.

Hansheng Xue, Luwei Yang, Vaibhav Rajan et al.

A bipartite network is a graph structure where nodes are from two distinct domains and only inter-domain interactions exist as edges. A large number of network embedding methods exist to learn vectorial node representations from general graphs with both homogeneous and heterogeneous node and edge types, including some that can specifically model the distinct properties of bipartite networks. However, these methods are inadequate to model multiplex bipartite networks (e.g., in e-commerce), that have multiple types of interactions (e.g., click, inquiry, and buy) and node attributes. Most real-world multiplex bipartite networks are also sparse and have imbalanced node distributions that are challenging to model. In this paper, we develop an unsupervised Dual HyperGraph Convolutional Network (DualHGCN) model that scalably transforms the multiplex bipartite network into two sets of homogeneous hypergraphs and uses spectral hypergraph convolutional operators, along with intra- and inter-message passing strategies to promote information exchange within and across domains, to learn effective node embedding. We benchmark DualHGCN using four real-world datasets on link prediction and node classification tasks. Our extensive experiments demonstrate that DualHGCN significantly outperforms state-of-the-art methods, and is robust to varying sparsity levels and imbalanced node distributions.

Dehong Gao, Wenjing Yang, Huiling Zhou et al.

The Multi-Task Learning (MTL) technique has been widely studied by word-wide researchers. The majority of current MTL studies adopt the hard parameter sharing structure, where hard layers tend to learn general representations over all tasks and specific layers are prone to learn specific representations for each task. Since the specific layers directly follow the hard layers, the MTL model needs to estimate this direct change (from general to specific) as well. To alleviate this problem, we introduce the novel cluster layer, which groups tasks into clusters during training procedures. In a cluster layer, the tasks in the same cluster are further required to share the same network. By this way, the cluster layer produces the general presentation for the same cluster, while produces relatively specific presentations for different clusters. As transitions the cluster layers are used between the hard layers and the specific layers. The MTL model thus learns general representations to specific representations gradually. We evaluate our model with MTL document classification and the results demonstrate the cluster layer is quite efficient in MTL.

Hansheng Xue, Luwei Yang, Wen Jiang et al.

Network embedding aims to learn low-dimensional representations of nodes while capturing structure information of networks. It has achieved great success on many tasks of network analysis such as link prediction and node classification. Most of existing network embedding algorithms focus on how to learn static homogeneous networks effectively. However, networks in the real world are more complex, e.g., networks may consist of several types of nodes and edges (called heterogeneous information) and may vary over time in terms of dynamic nodes and edges (called evolutionary patterns). Limited work has been done for network embedding of dynamic heterogeneous networks as it is challenging to learn both evolutionary and heterogeneous information simultaneously. In this paper, we propose a novel dynamic heterogeneous network embedding method, termed as DyHATR, which uses hierarchical attention to learn heterogeneous information and incorporates recurrent neural networks with temporal attention to capture evolutionary patterns. We benchmark our method on four real-world datasets for the task of link prediction. Experimental results show that DyHATR significantly outperforms several state-of-the-art baselines.