arXiv subjects

Junfeng Wu

Publications and source records attributed to Junfeng Wu.

At least 73 records · Page 4Linked to original sources

arXiv2022

SeqFormer: Sequential Transformer for Video Instance Segmentation

In this work, we present SeqFormer for video instance segmentation. SeqFormer follows the principle of vision transformer that models instance relationships among video frames. Nevertheless, we observe that a stand-alone instance query suffices for capturing a time sequence of instances in a video, but attention mechanisms shall be done with each frame independently. To achieve this, SeqFormer locates an instance in each frame and aggregates temporal information to learn a powerful representation of a video-level instance, which is used to predict the mask sequences on each frame dynamically. Instance tracking is achieved naturally without tracking branches or post-processing. On YouTube-VIS, SeqFormer achieves 47.4 AP with a ResNet-50 backbone and 49.0 AP with a ResNet-101 backbone without bells and whistles. Such achievement significantly exceeds the previous state-of-the-art performance by 4.6 and 4.4, respectively. In addition, integrated with the recently-proposed Swin transformer, SeqFormer achieves a much higher AP of 59.3. We hope SeqFormer could be a strong baseline that fosters future research in video instance segmentation, and in the meantime, advances this field with a more robust, accurate, neat model. The code is available at https://github.com/wjf5203/SeqFormer.

cs.CV↗

arXiv2022

In Defense of Online Models for Video Instance Segmentation

In recent years, video instance segmentation (VIS) has been largely advanced by offline models, while online models gradually attracted less attention possibly due to their inferior performance. However, online methods have their inherent advantage in handling long video sequences and ongoing videos while offline models fail due to the limit of computational resources. Therefore, it would be highly desirable if online models can achieve comparable or even better performance than offline models. By dissecting current online models and offline models, we demonstrate that the main cause of the performance gap is the error-prone association between frames caused by the similar appearance among different instances in the feature space. Observing this, we propose an online framework based on contrastive learning that is able to learn more discriminative instance embeddings for association and fully exploit history information for stability. Despite its simplicity, our method outperforms all online and offline methods on three benchmarks. Specifically, we achieve 49.5 AP on YouTube-VIS 2019, a significant improvement of 13.2 AP and 2.1 AP over the prior online and offline art, respectively. Moreover, we achieve 30.2 AP on OVIS, a more challenging dataset with significant crowding and occlusions, surpassing the prior art by 14.8 AP. The proposed method won first place in the video instance segmentation track of the 4th Large-scale Video Object Segmentation Challenge (CVPR2022). We hope the simplicity and effectiveness of our method, as well as our insight into current methods, could shed light on the exploration of VIS models.

cs.CV↗

arXiv2022

Closed-form Error Propagation on the SE_n(3) Group for Invariant Extended Kalman Filtering with Applications to VINS

Pose estimation is important for robotic perception, path planning, etc. Robot poses can be modeled on matrix Lie groups and are usually estimated via filter-based methods. In this paper, we establish the closed-form formula for the error propagation for the Invariant extended Kalman filter (IEKF) in the presence of random noises and apply it to vision-aided inertial navigation. We evaluate our algorithm via numerical simulations and experiments on the OPENVINS platform. Both simulations and the experiments performed on the public EuRoC MAV datasets demonstrate that our algorithm outperforms some state-of-art filter-based methods such as the quaternion-based EKF, first estimates Jacobian EKF, etc.

cs.RO↗

arXiv2022

Super-Resolution Domain Adaptation Networks for Semantic Segmentation via Pixel and Output Level Aligning

Recently, Unsupervised Domain Adaptation (UDA) has attracted increasing attention to address the domain shift problem in the semantic segmentation task. Although previous UDA methods have achieved promising performance, they still suffer from the distribution gaps between source and target domains, especially the resolution discrepany in the remote sensing images. To address this problem, this paper designs a novel end-to-end semantic segmentation network, namely Super-Resolution Domain Adaptation Network (SRDA-Net). SRDA-Net can simultaneously achieve the super-resolution task and the domain adaptation task, thus satisfying the requirement of semantic segmentation for remote sensing images which usually involve various resolution images. The proposed SRDA-Net includes three parts: a Super-Resolution and Segmentation (SRS) model which focuses on recovering high-resolution image and predicting segmentation map, a Pixel-level Domain Classifier (PDC) for determining which domain the pixel belongs to, and an Output-space Domain Classifier (ODC) for distinguishing which domain the pixel contribution is from. By jointly optimizing SRS with two classifiers, the proposed method can not only eliminates the resolution difference between source and target domains, but also improve the performance of the semantic segmentation task. Experimental results on two remote sensing datasets with different resolutions demonstrate that SRDA-Net performs favorably against some state-of-the-art methods in terms of accuracy and visual quality. Code and models are available at https://github.com/tangzhenjie/SRDA-Net.

cs.CV↗

arXiv2021

Event-Triggered Distributed Estimation With Decaying Communication Rate

We study distributed estimation of a high-dimensional static parameter vector through a group of sensors whose communication network is modeled by a fixed directed graph. Different from existing time-triggered communication schemes, an event-triggered asynchronous scheme is investigated in order to reduce communication while preserving estimation convergence. A distributed estimation algorithm with a single step size is first proposed based on an event-triggered communication scheme with a time-dependent decaying threshold. With the event-triggered scheme, each sensor sends its estimate to neighbor sensors only when the difference between the current estimate and the last sent-out estimate is larger than the triggering threshold. We prove that the proposed algorithm has mean-square and almost-sure convergence respectively, under an integrated condition of sensor network topology and sensor measurement matrices. The condition is satisfied if the topology is a balanced digraph containing a spanning tree and the system is collectively observable. Moreover, we provide estimates for the convergence rates, which are related to the step size as well as the triggering threshold. Furthermore, as an essential metric of sensor communication intensity in the event-triggered distributed algorithms, the communication rate is proved to decay to zero with a certain speed almost surely as time goes to infinity. We show that given the step size, adjusting the decay speed of the triggering threshold can lead to a tradeoff between the convergence rate of the estimation error and the decay speed of the communication rate. Specifically, increasing the decay speed of the threshold would make the communication rate decay faster, but reduce the convergence rate of the estimation error. Numerical simulations are provided to illustrate the developed results.

eess.SY↗

arXiv2021

Low-complexity Distributed Detection with One-bit Memory Under Neyman-Pearson Criterion

We consider a multi-stage distributed detection scenario, where $n$ sensors and a fusion center (FC) are deployed to accomplish a binary hypothesis test. At each time stage, local sensors generate binary messages, assumed to be spatially and temporally independent given the hypothesis, and then upload them to the FC for global detection decision making. We suppose a one-bit memory is available at the FC to store its decision history and focus on developing iterative fusion schemes. We first visit the detection problem of performing the Neyman-Pearson (N-P) test at each stage and give an optimal algorithm, called the oracle algorithm, to solve it. Structural properties and limitation of the fusion performance in the asymptotic regime are explored for the oracle algorithm. We notice the computational inefficiency of the oracle fusion and propose a low-complexity alternative, for which the likelihood ratio (LR) test threshold is tuned in connection to the fusion decision history compressed in the one-bit memory. The low-complexity algorithm greatly brings down the computational complexity at each stage from $O(4^n)$ to $O(n)$. We show that the proposed algorithm is capable of converging exponentially to the same detection probability as that of the oracle one. Moreover, the rate of convergence is shown to be asymptotically identical to that of the oracle algorithm. Finally, numerical simulations and real-world experiments demonstrate the effectiveness and efficiency of our distributed algorithm.

eess.SP↗

arXiv2021

Real-time tracking of COVID-19 and coronavirus research updates through text mining

The novel coronavirus (SARS-CoV-2) which causes COVID-19 is an ongoing pandemic. There are ongoing studies with up to hundreds of publications uploaded to databases daily. We are exploring the use-case of artificial intelligence and natural language processing in order to efficiently sort through these publications. We demonstrate that clinical trial information, preclinical studies, and a general topic model can be used as text mining data intelligence tools for scientists all over the world to use as a resource for their own research. To evaluate our method, several metrics are used to measure the information extraction and clustering results. In addition, we demonstrate that our workflow not only have a use-case for COVID-19, but for other disease areas as well. Overall, our system aims to allow scientists to more efficiently research coronavirus. Our automatically updating modules are available on our information portal at https://ghddi-ailab.github.io/Targeting2019-nCoV/ for public viewing.

cs.IR↗

arXiv2021

ParaVS: A Simple, Fast, Efficient and Flexible Graph Neural Network Framework for Structure-Based Virtual Screening

Structure-based virtual screening (SBVS) is a promising in silico technique that integrates computational methods into drug design. An extensively used method in SBVS is molecular docking. However, the docking process can hardly be computationally efficient and accurate simultaneously because classic mechanics scoring function is used to approximate, but hardly reach, the quantum mechanics precision in this method. In order to reduce the computational cost of the protein-ligand scoring process and use data driven approach to boost the scoring function accuracy, we introduce a docking-based SBVS method and, furthermore, a deep learning non-docking-based method that is able to avoid the computational cost of the docking process. Then, we try to integrate these two methods into an easy-to-use framework, ParaVS, that provides both choices for researchers. Graph neural network (GNN) is employed in ParaVS, and we explained how our in-house GNN works and how to model ligands and molecular targets. To verify our approaches, cross validation experiments are done on two datasets, an open dataset Directory of Useful Decoys: Enhanced (DUD.E) and an in-house proprietary dataset without computational generated artificial decoys (NoDecoy). On DUD.E we achieved a state-of-the-art AUC of 0.981 and a state-of-the-art enrichment factor at 2% of 36.2; on NoDecoy we achieved an AUC of 0.974. We further finish inference of an open database, Enamine REAL Database (RDB), that comprises over 1.36 billion molecules in 4050 core-hours using our ParaVS non-docking method (ParaVS-ND). The inference speed of ParaVS-ND is about 3.6e5 molecule / core-hour, while this number of a conventional docking-based method is around 20, which is about 16000 times faster. The experiments indicate that ParaVS is accurate, computationally efficient and can be generalized to different molecular.

q-bio.BM↗

arXiv2021

Multi-Party Dynamic State Estimation that Preserves Data and Model Privacy

In this paper we focus on the dynamic state estimation which harnesses a vast amount of sensing data harvested by multiple parties and recognize that in many applications, to improve collaborations between parties, the estimation procedure must be designed with the awareness of protecting participants' data and model privacy, where the latter refers to the privacy of key parameters of observation models. We develop a state estimation paradigm for the scenario where multiple parties with data and model privacy concerns are involved. Multiple parties monitor a physical dynamic process by deploying their own sensor networks and update the state estimate according to the average state estimate of all the parties calculated by a cloud server and security module. The paradigm taps additively homomorphic encryption which enables the cloud server and security module to jointly fuse parties' data while preserving the data privacy. Meanwhile, all the parties collaboratively develop a stable (or optimal) fusion rule without divulging sensitive model information. For the proposed filtering paradigm, we analyze the stabilization and the optimality. First, to stabilize the multi-party state estimator while preserving observation model privacy, two stabilization design methods are proposed. For special scenarios, the parties directly design their estimator gains by the matrix norm relaxation. For general scenarios, after transforming the original design problem into a convex semi-definite programming problem, the parties collaboratively derive suitable estimator gains based on the ADMM. Second, an optimal collaborative gain design method with model privacy guarantees is provided, which results in the asymptotic MMSE state estimation. Finally, numerical examples are presented to illustrate our design and theoretical findings.

eess.SY↗

arXiv2020

Mean-Field Transmission Power Control in Dense Networks

We consider uplink power control in wireless communication when a large number of users compete over the channel resources. The CDMA protocol, as a supporting technology of 3G networks accommodating signal from different sources over the code domain, represents the orthogonal multiple access (OMA) techniques. With the development of 5G wireless networks, non-orthogonal multiple access (NOMA) is introduced to improve the efficiency of channel allocation. Our goal is to investigate whether the power-domain NOMA protocol can introduce performance improvement when the users interact with each other in a non-cooperative manner. It is compared with the CDMA protocol, where the fierce competition among users jeopardizes the efficiency of channel usage. In this work, we conduct analysis with an aggregative game model, and show the existence and uniqueness of an equilibrium strategy. Next, we adopt the social welfare of the population as the performance metric, which is the average utility achieved by the user population. It is shown that under the corresponding equilibrium strategies, NOMA outperforms CDMA by higher efficiency of channel access for uplink communications.

eess.SY↗

arXiv2020

Learning Quadratic Games on Networks

Individuals, or organizations, cooperate with or compete against one another in a wide range of practical situations. Such strategic interactions are often modeled as games played on networks, where an individual's payoff depends not only on her action but also on that of her neighbors. The current literature has largely focused on analyzing the characteristics of network games in the scenario where the structure of the network, which is represented by a graph, is known beforehand. It is often the case, however, that the actions of the players are readily observable while the underlying interaction network remains hidden. In this paper, we propose two novel frameworks for learning, from the observations on individual actions, network games with linear-quadratic payoffs, and in particular, the structure of the interaction network. Our frameworks are based on the Nash equilibrium of such games and involve solving a joint optimization problem for the graph structure and the individual marginal benefits. Both synthetic and real-world experiments demonstrate the effectiveness of the proposed frameworks, which have theoretical as well as practical implications for understanding strategic interactions in a network environment.

cs.GT↗

arXiv2020

Stochastic Event-based Sensor Schedules for Remote State Estimation in Cognitive Radio Sensor Networks

We consider the problem of communication allocation for remote state estimation in a cognitive radio sensor network~(CRSN). A sensor collects measurements of a physical plant, and transmits the data to a remote estimator as a secondary user (SU) in the shared network. The existence of the primal users (PUs) brings exogenous uncertainties into the transmission scheduling process, and how to design an event-based scheduling scheme considering these uncertainties has not been addressed in the literature. In this work, we start from the formulation of a discrete-time remote estimation process in the CRSN, and then analyze the hidden information contained in the absence of data transmission. In order to achieve a better tradeoff between estimation performance and communication consumption, we propose both open-loop and closed-loop schedules using the hidden information under a Bayesian setting. The open-loop schedule does not rely on any feedback signal but only works for stable plants. For unstable plants, a closed-loop schedule is designed based on feedback signals. The parameter design problems in both schedules are efficiently solved by convex programming. Numerical simulations are included to illustrate the theoretical results.

eess.SY↗

arXiv2020

Mean-Field Transmission Power Control in Dense Networks, Part II -- Social Welfare Evaluation

We consider uplink power control in wireless communication when massive users compete over the channel resources. In Part I, we have formulated massive transmission power control contest in a mean-field game framework. In this part, our goal is to investigate whether the power-domain non-orthogonal multiple access (NOMA) protocol can regulate the non-cooperative channel access behaviors, i.e., steering the competition among the non-cooperative users in a direction with improved efficiency and fairness. It is compared with the CDMA protocol, which drives each user to fiercely compete against the population, hence the efficiency of channel usage is sacrificed. The existence and uniqueness of an equilibrium strategy under CDMA and NOMA have already been characterized in Part I. In this paper, we adopt the social welfare of the population as the performance metric, which is defined as the expectation of utility over the distribution of different types of channel users. It is shown that under the corresponding equilibrium strategies, NOMA outperforms CDMA in the social welfare achieved, which is illustrated through simulation with different unit price for power consumption. Moreover, it can be observed from numerical results that NOMA can improve the fairness of the achieved data rates among different users.

eess.SY↗

arXiv2019

Distributed Least Squares Solver for Network Linear Equations

In this paper, we study the problem of finding the least square solutions of over-determined linear algebraic equations over networks in a distributed manner. Each node has access to one of the linear equations and holds a dynamic state. We first propose a distributed least square solver over connected undirected interaction graphs and establish a necessary and sufficient on the step-size under which the algorithm exponentially converges to the least square solution. Next, we develop a distributed least square solver over strongly connected directed graphs and show that the proposed algorithm exponentially converges to the least square solution provided the step-size is sufficiently small. Moreover, we develop a finite-time least square solver by equipping the proposed algorithms with a finite-time decentralized computation mechanism. The theoretical findings are validated and illustrated by numerical simulation examples.

math.OC↗

arXiv2019

Event-Triggered Control for Consensus of Multi-Agent Systems with Nonlinear Output and Directed Topologies

We propose a distributed event-triggered control law to solve the consensus problem for multi-agent systems with nonlinear output. Under the condition that the underlying digraph is strongly connected, we propose some sufficient conditions related to the nonlinear output function and initial states to guarantee that the event-triggered controller realizes consensus. Then the results are extended to the case where the underlying directed graph contains a directed spanning tree. These theoretical results are illustrated by numerical simulations.

eess.SY↗

arXiv2019

Distributed Parameter Estimation Under Event-triggered Communications

In this paper, we study a distributed parameter estimation problem with an asynchronous communication protocol over multi-agent systems. Different from traditional time-driven communication schemes, in this work, data can be transmitted between agents intermittently rather than in a steady stream. First, we propose a recursive distributed estimator based on an event-triggered communication scheme, through which each agent can decide whether the current estimate is sent out to its neighbors or not. With this scheme, considerable communications between agents can be effectively reduced. Then, under mild conditions including a collective observability, we provide a design principle of triggering thresholds to guarantee the asymptotic unbiasedness and strong consistency. Furthermore, under certain conditions, we prove that, with probability one, for every agent the time interval between two successive triggered instants goes to infinity as time goes to infinity. Finally, we provide a numerical simulation to validate the theoretical results of this paper.

eess.SY↗

arXiv2019

Dynamical Privacy in Distributed Computing -- Part II: PPSC Gossip Algorithms

In the first part of the paper, we have studied the computational privacy risks in distributed computing protocols against local or global dynamics eavesdroppers, and proposed a Privacy-Preserving-Summation-Consistent (PPSC) mechanism as a generic privacy encryption subroutine for consensus-based distributed computations. In this part of this paper, we show that the conventional deterministic and random gossip algorithms can be used to realize the PPSC mechanism over a given network. At each time step, a node is selected to interact with one of its neighbors via deterministic or random gossiping. Such node generates a random number as its new state, and sends the subtraction between its current state and that random number to the neighbor; then the neighbor updates its state by adding the received value to its current state. We establish concrete privacy-preservation conditions by proving the impossibility for the reconstruction of the network input from the output of the gossip-based PPSC mechanism against eavesdroppers with full network knowledge, and by showing that the PPSC mechanism can achieve differential privacy at arbitrary privacy levels. The convergence is characterized explicitly and analytically for both deterministic and randomized gossiping, which is essentially achieved in a finite number of steps. Additionally, we illustrate that the proposed algorithms can be easily made adaptive in real-world applications by making realtime trade-offs between resilience against node dropout or communication failure and privacy preservation capabilities.

eess.SY↗

arXiv2019

Dynamical Privacy in Distributed Computing -- Part I: Privacy Loss and PPSC Mechanism

A distributed computing protocol consists of three components: (i) Data Localization: a network-wide dataset is decomposed into local datasets separately preserved at a network of nodes; (ii) Node Communication: the nodes hold individual dynamical states and communicate with the neighbors about these dynamical states; (iii) Local Computation: state recursions are computed at each individual node. Information about the local datasets enters the computation process through the node-to-node communication and the local computations, which may be leaked to dynamics eavesdroppers having access to global or local node states. In this paper, we systematically investigate this potential computational privacy risks in distributed computing protocols in the form of structured system identification, and then propose and thoroughly analyze a Privacy-Preserving-Summation-Consistent (PPSC) mechanism as a generic privacy encryption subroutine for consensus-based distributed computations. The central idea is that the consensus manifold is where we can both hide node privacy and achieve computational accuracy. In this first part of the paper, we demonstrate the computational privacy risks in distributed algorithms against dynamics eavesdroppers and particularly in distributed linear equation solvers, and then propose the PPSC mechanism and illustrate its usefulness.

eess.SY↗