Search arXiv⌕ Search

arXiv subjects

Jiaqi Liu

Publications and source records attributed to Jiaqi Liu.

At least 199 records · Page 11Linked to original sources
arXiv2023

Critical drift estimates for the frog model on trees

Place an active particle at the root of a $d$-ary tree and a single dormant particle at each non-root site. In discrete time, active particles move towards the root with probability $p$ and, otherwise, away from the root to a uniformly sampled child vertex. When an active particle moves to a site containing a dormant particle, the dormant particle becomes active. The critical drift $p_d$ is the infimum over all $p$ for which infinitely many particles visit the root almost surely. Guo, Tang, and Wei proved that $\sup_{d\geq 3} p_d \leq 1/3$. We improve this bound to $5/17$ with a shorter argument that generalizes to give bounds on $\sup_{d \geq m} p_d$. We additionally prove that $\limsup p_d \leq 1/6$ by finding the limiting critical drift for a non-backtracking variant.

math.PR↗
arXiv2023

Unsupervised Domain Generalization for Person Re-identification: A Domain-specific Adaptive Framework

Domain generalization (DG) has attracted much attention in person re-identification (ReID) recently. It aims to make a model trained on multiple source domains generalize to an unseen target domain. Although achieving promising progress, existing methods usually need the source domains to be labeled, which could be a significant burden for practical ReID tasks. In this paper, we turn to investigate unsupervised domain generalization for ReID, by assuming that no label is available for any source domains. To address this challenging setting, we propose a simple and efficient domain-specific adaptive framework, and realize it with an adaptive normalization module designed upon the batch and instance normalization techniques. In doing so, we successfully yield reliable pseudo-labels to implement training and also enhance the domain generalization capability of the model as required. In addition, we show that our framework can even be applied to improve person ReID under the settings of supervised domain generalization and unsupervised domain adaptation, demonstrating competitive performance with respect to relevant methods. Extensive experimental study on benchmark datasets is conducted to validate the proposed framework. A significance of our work lies in that it shows the potential of unsupervised domain generalization for person ReID and sets a strong baseline for the further research on this topic.

cs.CV↗
arXiv2022

Label Distribution Learning for Generalizable Multi-source Person Re-identification

Person re-identification (Re-ID) is a critical technique in the video surveillance system, which has achieved significant success in the supervised setting. However, it is difficult to directly apply the supervised model to arbitrary unseen domains due to the domain gap between the available source domains and unseen target domains. In this paper, we propose a novel label distribution learning (LDL) method to address the generalizable multi-source person Re-ID task (i.e., there are multiple available source domains, and the testing domain is unseen during training), which aims to explore the relation of different classes and mitigate the domain-shift across different domains so as to improve the discrimination of the model and learn the domain-invariant feature, simultaneously. Specifically, during the training process, we produce the label distribution via the online manner to mine the relation information of different classes, thus it is beneficial for extracting the discriminative feature. Besides, for the label distribution of each class, we further revise it to give more and equal attention to the other domains that the class does not belong to, which can effectively reduce the domain gap across different domains and obtain the domain-invariant feature. Furthermore, we also give the theoretical analysis to demonstrate that the proposed method can effectively deal with the domain-shift issue. Extensive experiments on multiple benchmark datasets validate the effectiveness of the proposed method and show that the proposed method can outperform the state-of-the-art methods. Besides, further analysis also reveals the superiority of the proposed method.

cs.CV↗
arXiv2022

Boosting Distributed Training Performance of the Unpadded BERT Model

Pre-training models are an important tool in Natural Language Processing (NLP), while the BERT model is a classic pre-training model whose structure has been widely adopted by followers. It was even chosen as the reference model for the MLPerf training benchmark. The distributed training performance optimization of BERT models plays an important role in accelerating the solutions of most NLP tasks. BERT model often uses padding tensors as its inputs, leading to excessive redundant computations. Thus, removing these redundant computations is essential to improve the distributed training performance. This paper designs a new approach to train BERT models with variable-length inputs efficiently. Firstly, we propose a general structure for the variable-length BERT models, and accelerate the encoder layer via our grouped multi-stream FMHA (Fused Multi-Head Attention) method. Secondly, through data exchange, we address the unbalanced workload problem caused by the variable-length inputs, which overlaps highly with the training process. Finally, we optimize the overall performance of the BERT model, such as kernel fusion, and operator optimization. Our experimental results show that our highly optimized BERT model achieves state-of-the-art throughput and ranks first in MLPerf Training v2.0 within the same GPU configuration. The optimizations in this paper can be applied to more BERT-like models in our future works.

cs.DC↗
arXiv2022

Egocentric Human Trajectory Forecasting with a Wearable Camera and Multi-Modal Fusion

In this paper, we address the problem of forecasting the trajectory of an egocentric camera wearer (ego-person) in crowded spaces. The trajectory forecasting ability learned from the data of different camera wearers walking around in the real world can be transferred to assist visually impaired people in navigation, as well as to instill human navigation behaviours in mobile robots, enabling better human-robot interactions. To this end, a novel egocentric human trajectory forecasting dataset was constructed, containing real trajectories of people navigating in crowded spaces wearing a camera, as well as extracted rich contextual data. We extract and utilize three different modalities to forecast the trajectory of the camera wearer, i.e., his/her past trajectory, the past trajectories of nearby people, and the environment such as the scene semantics or the depth of the scene. A Transformer-based encoder-decoder neural network model, integrated with a novel cascaded cross-attention mechanism that fuses multiple modalities, has been designed to predict the future trajectory of the camera wearer. Extensive experiments have been conducted, with results showing that our model outperforms the state-of-the-art methods in egocentric human trajectory forecasting.

cs.CV↗
arXiv2022

Long-time asymptotics of the modified KdV equation in weighted Sobolev spaces

The long time behavior of solutions to the defocusing modified Korteweg-de vries (MKdV) equation is established for initial conditions in some weighted Sobolev spaces. Our approach is based on the nonlinear steepest descent method of Deift and Zhou and its reformulation by Dieng and McLaughlin through $\overline{\partial}$-derivatives. To extend the asymptotics to solutions with initial data in lower regularity spaces, we apply a global approximation via PDE techniques.

math.AP↗
arXiv2022

Long-time asymptotics and stability for the sine-Gordon equation

In this paper, we study the long-time dynamics and stability properties of the sine-Gordon equation $$f_{tt}-f_{xx}+\sin f=0.$$ Firstly, we use the nonlinear steepest descent for Riemann-Hilbert problems to compute the long-time asymptotics of the solutions to the sine-Gordon equation whose initial condition belongs to some weighted Sobolev spaces. Secondly, we study the asymptotic stability of the sine-Gordon equation. It is known that the obstruction to the asymptotic stability of the sine-Gordon equation in the energy space is the existence of small breathers which is also closely related to the emergence of wobbling kinks. Combining the long-time asymptotics and a refined approximation argument, we analyze the asymptotic stability properties of the sine-Gordon equation in weighted energy spaces. Our stability analysis gives a criterion for the weight which is sharp up to the endpoint so that the asymptotic stability holds.

math.AP↗
arXiv2022

Particle configurations for branching Brownian motion with an inhomogeneous branching rate

Aiming to understand the distribution of fitness levels of individuals in a large population undergoing selection, we study the particle configurations of branching Brownian motion where each particle independently moves as Brownian motion with negative drift, particles can die or undergo dyadic fission, and the difference between the birth rate and the death rate is proportional to the particle's location. Under some assumptions, we obtain the limit in probability of the number of particles in any given interval and an explicit formula for the asymptotic empirical density of the fitness distribution. We show that after a sufficiently long time, the fitness distribution from the lowest to the highest fitness levels approximately evolves as a traveling wave with a profile which is asymptotically related the the Airy function. Our work complements the results in Roberts and Schweinsberg (2021), giving a fuller picture of the fitness distribution.

math.PR↗
arXiv2022

A Yaglom type asymptotic result for subcritical branching Brownian motion with absorption

We consider a slightly subcritical branching Brownian motion with absorption, where particles move as Brownian motions with drift $-\sqrt{2+2\varepsilon}$, undergo dyadic fission at rate $1$, and are killed when they reach the origin. We obtain a Yaglom type asymptotic result, showing that the long run expected number of particles conditioned on survival grows exponentially as $1/\sqrt{\varepsilon}$ as the process approaches criticality.

math.PR↗
arXiv2021

AskMe: Joint Individual-level and Community-level Behavior Interaction for Question Recommendation

Questions in Community Question Answering (CQA) sites are recommended to users, mainly based on users' interest extracted from questions that users have answered or have asked. However, there is a general phenomenon that users answer fewer questions while pay more attention to follow questions and vote answers. This can impact the performance when recommending questions to users (for obtaining their answers) by using their historical answering behaviors on existing studies. To address the data sparsity issue, we propose AskMe, which aims to leverage the rich, hybrid behavior interactions in CQA to improve the question recommendation performance. On the one hand, we model the rich correlations between the users' diverse behaviors (e.g., answer, follow, vote) to obtain the individual-level behavior interaction. On the other hand, we model the sophisticated behavioral associations between similar users to obtain the community-level behavior interaction. Finally, we propose the way of element-level fusion to mix these two kinds of interactions together to predict the ranking scores. A dataset collected from Zhihu (1126 users, 219434 questions) is utilized to evaluate the performance of the proposed model, and the experimental results show that our model has gained the best performance compared to baseline methods, especially when the historical answering behaviors data is scarce.

cs.IR↗
arXiv2021

DeepExpress: Heterogeneous and Coupled Sequence Modeling for Express Delivery Prediction

The prediction of express delivery sequence, i.e., modeling and estimating the volumes of daily incoming and outgoing parcels for delivery, is critical for online business, logistics, and positive customer experience, and specifically for resource allocation optimization and promotional activity arrangement. A precise estimate of consumer delivery requests has to involve sequential factors such as shopping behaviors, weather conditions, events, business campaigns, and their couplings. Besides, conventional sequence prediction assumes a stable sequence evolution, failing to address complex nonlinear sequences and various feature effects in the above multi-source data. Although deep networks and attention mechanisms demonstrate the potential of complex sequence modeling, extant networks ignore the heterogeneous and coupling situation between features and sequences, resulting in weak prediction accuracy. To address these issues, we propose DeepExpress - a deep-learning based express delivery sequence prediction model, which extends the classic seq2seq framework to learning complex coupling between sequence and features. DeepExpress leverages an express delivery seq2seq learning, a carefully-designed heterogeneous feature representation, and a novel joint training attention mechanism to adaptively map heterogeneous data, and capture sequence-feature coupling for precise estimation. Experimental results on real-world data demonstrate that the proposed method outperforms both shallow and deep baseline models.

cs.LG↗
arXiv2021

Decentralized Multi-AGV Task Allocation based on Multi-Agent Reinforcement Learning with Information Potential Field Rewards

Automated Guided Vehicles (AGVs) have been widely used for material handling in flexible shop floors. Each product requires various raw materials to complete the assembly in production process. AGVs are used to realize the automatic handling of raw materials in different locations. Efficient AGVs task allocation strategy can reduce transportation costs and improve distribution efficiency. However, the traditional centralized approaches make high demands on the control center's computing power and real-time capability. In this paper, we present decentralized solutions to achieve flexible and self-organized AGVs task allocation. In particular, we propose two improved multi-agent reinforcement learning algorithms, MADDPG-IPF (Information Potential Field) and BiCNet-IPF, to realize the coordination among AGVs adapting to different scenarios. To address the reward-sparsity issue, we propose a reward shaping strategy based on information potential field, which provides stepwise rewards and implicitly guides the AGVs to different material targets. We conduct experiments under different settings (3 AGVs and 6 AGVs), and the experiment results indicate that, compared with baseline methods, our work obtains up to 47\% task response improvement and 22\% training iterations reduction.

cs.RO↗
arXiv2021

TL-SDD: A Transfer Learning-Based Method for Surface Defect Detection with Few Samples

Surface defect detection plays an increasingly important role in manufacturing industry to guarantee the product quality. Many deep learning methods have been widely used in surface defect detection tasks, and have been proven to perform well in defects classification and location. However, deep learning-based detection methods often require plenty of data for training, which fail to apply to the real industrial scenarios since the distribution of defect categories is often imbalanced. In other words, common defect classes have many samples but rare defect classes have extremely few samples, and it is difficult for these methods to well detect rare defect classes. To solve the imbalanced distribution problem, in this paper we propose TL-SDD: a novel Transfer Learning-based method for Surface Defect Detection. First, we adopt a two-phase training scheme to transfer the knowledge from common defect classes to rare defect classes. Second, we propose a novel Metric-based Surface Defect Detection (M-SDD) model. We design three modules for this model: (1) feature extraction module: containing feature fusion which combines high-level semantic information with low-level structural information. (2) feature reweighting module: transforming examples to a reweighting vector that indicates the importance of features. (3) distance metric module: learning a metric space in which defects are classified by computing distances to representations of each category. Finally, we validate the performance of our proposed method on a real dataset including surface defects of aluminum profiles. Compared to the baseline methods, the performance of our proposed method has improved by up to 11.98% for rare defect classes.

cs.CV↗
arXiv2021

An Efficient Learning Framework For Federated XGBoost Using Secret Sharing And Distributed Optimization

XGBoost is one of the most widely used machine learning models in the industry due to its superior learning accuracy and efficiency. Targeting at data isolation issues in the big data problems, it is crucial to deploy a secure and efficient federated XGBoost (FedXGB) model. Existing FedXGB models either have data leakage issues or are only applicable to the two-party setting with heavy communication and computation overheads. In this paper, a lossless multi-party federated XGB learning framework is proposed with a security guarantee, which reshapes the XGBoost's split criterion calculation process under a secret sharing setting and solves the leaf weight calculation problem by leveraging distributed optimization. Remarkably, a thorough analysis of model security is provided as well, and multiple numerical results showcase the superiority of the proposed FedXGB compared with the state-of-the-art models on benchmark datasets.

cs.LG↗
arXiv2020

Combine Convolution with Recurrent Networks for Text Classification

Convolutional neural network (CNN) and recurrent neural network (RNN) are two popular architectures used in text classification. Traditional methods to combine the strengths of the two networks rely on streamlining them or concatenating features extracted from them. In this paper, we propose a novel method to keep the strengths of the two networks to a great extent. In the proposed model, a convolutional neural network is applied to learn a 2D weight matrix where each row reflects the importance of each word from different aspects. Meanwhile, we use a bi-directional RNN to process each word and employ a neural tensor layer that fuses forward and backward hidden states to get word representations. In the end, the weight matrix and word representations are combined to obtain the representation in a 2D matrix form for the text. We carry out experiments on a number of datasets for text classification. The experimental results confirm the effectiveness of the proposed method.

cs.CL↗
arXiv2020

Modeling, Analysis, and Optimization of Grant-Free NOMA in Massive MTC via Stochastic Geometry

Massive machine-type communications (mMTC) is a crucial scenario to support booming Internet of Things (IoTs) applications. In mMTC, although a large number of devices are registered to an access point (AP), very few of them are active with uplink short packet transmission at the same time, which requires novel design of protocols and receivers to enable efficient data transmission and accurate multi-user detection (MUD). Aiming at this problem, grant-free non-orthogonal multiple access (GF-NOMA) protocol is proposed. In GF-NOMA, active devices can directly transmit their preambles and data symbols altogether within one time frame, without grant from the AP. Compressive sensing (CS)-based receivers are adopted for non-orthogonal preambles (NOP)-based MUD, and successive interference cancellation is exploited to decode the superimposed data signals. In this paper, we model, analyze, and optimize the CS-based GF-MONA mMTC system via stochastic geometry (SG), from an aspect of network deployment. Based on the SG network model, we first analyze the success probability as well as the channel estimation error of the CS-based MUD in the preamble phase and then analyze the average aggregate data rate in the data phase. As IoT applications highly demands low energy consumption, low infrastructure cost, and flexible deployment, we optimize the energy efficiency and AP coverage efficiency of GF-NOMA via numerical methods. The validity of our analysis is verified via Monte Carlo simulations. Simulation results also show that CS-based GF-NOMA with NOP yields better MUD and data rate performances than contention-based GF-NOMA with orthogonal preambles and CS-based grant-free orthogonal multiple access.

cs.IT↗
arXiv2019

Soliton resolution for the modified KdV equation

The soliton resolution for the focusing modified Korteweg-de vries (mKdV) equation is established for initial conditions in some weighted Sobolev spaces. Our approach is based on the nonlinear steepest descent method and its reformulation through $\bar{\partial}$-derivatives. From the view of stationary points, we give precise asymptotic formulas along trajectory $x=\textrm{v}t$ for any fixed $ \textrm{v} $. To extend the asymptotics to solutions with initial data in low regularity spaces, we apply a global approximation via PDE techniques. As byproducts of our long-time asymptotics, we also obtain the asymptotic stability of nonlinear structures involving solitons and breathers.

math.AP↗