arXiv subjects

Jian Sun

Publications and source records attributed to Jian Sun.

At least 235 records · Page 13Linked to original sources

arXiv2022

Model-Based and Data-Driven Control of Event- and Self-Triggered Discrete-Time LTI Systems

The present paper considers the model-based and data-driven control of unknown linear time-invariant discrete-time systems under event-triggering and self-triggering transmission schemes. To this end, we begin by presenting a dynamic event-triggering scheme (ETS) based on periodic sampling, and a discrete-time looped-functional approach, through which a model-based stability condition is derived. Combining the model-based condition with a recent data-based system representation, a data-driven stability criterion in the form of linear matrix inequalities (LMIs) is established, which also offers a way of co-designing the ETS matrix and the controller. To further alleviate the sampling burden of ETS due to its continuous/periodic detection, a self-triggering scheme (STS) is developed. Leveraging pre-collected input-state data, an algorithm for predicting the next transmission instant is given, while achieving system stability. Finally, numerical simulations showcase the efficacy of ETS and STS in reducing data transmissions as well as of the proposed co-design methods.

eess.SY↗

arXiv2022

SPACE-3: Unified Dialog Model Pre-training for Task-Oriented Dialog Understanding and Generation

Recently, pre-training methods have shown remarkable success in task-oriented dialog (TOD) systems. However, most existing pre-trained models for TOD focus on either dialog understanding or dialog generation, but not both. In this paper, we propose SPACE-3, a novel unified semi-supervised pre-trained conversation model learning from large-scale dialog corpora with limited annotations, which can be effectively fine-tuned on a wide range of downstream dialog tasks. Specifically, SPACE-3 consists of four successive components in a single transformer to maintain a task-flow in TOD systems: (i) a dialog encoding module to encode dialog history, (ii) a dialog understanding module to extract semantic vectors from either user queries or system responses, (iii) a dialog policy module to generate a policy vector that contains high-level semantics of the response, and (iv) a dialog generation module to produce appropriate responses. We design a dedicated pre-training objective for each component. Concretely, we pre-train the dialog encoding module with span mask language modeling to learn contextualized dialog information. To capture the structured dialog semantics, we pre-train the dialog understanding module via a novel tree-induced semi-supervised contrastive learning objective with the help of extra dialog annotations. In addition, we pre-train the dialog policy module by minimizing the L2 distance between its output policy vector and the semantic vector of the response for policy optimization. Finally, the dialog generation model is pre-trained by language modeling. Results show that SPACE-3 achieves state-of-the-art performance on eight downstream dialog benchmarks, including intent prediction, dialog state tracking, and end-to-end dialog modeling. We also show that SPACE-3 has a stronger few-shot ability than existing models under the low-resource setting.

cs.CL↗

arXiv2022

Implicit Full Waveform Inversion with Deep Neural Representation

Full waveform inversion (FWI) commonly stands for the state-of-the-art approach for imaging subsurface structures and physical parameters, however, its implementation usually faces great challenges, such as building a good initial model to escape from local minima, and evaluating the uncertainty of inversion results. In this paper, we propose the implicit full waveform inversion (IFWI) algorithm using continuously and implicitly defined deep neural representations. Compared to FWI, which is sensitive to the initial model, IFWI benefits from the increased degrees of freedom with deep learning optimization, thus allowing to start from a random initialization, which greatly reduces the risk of non-uniqueness and being trapped in local minima. Both theoretical and experimental analyses indicates that, given a random initial model, IFWI is able to converge to the global minimum and produce a high-resolution image of subsurface with fine structures. In addition, uncertainty analysis of IFWI can be easily performed by approximating Bayesian inference with various deep learning approaches, which is analyzed in this paper by adding dropout neurons. Furthermore, IFWI has a certain degree of robustness and strong generalization ability that are exemplified in the experiments of various 2D geological models. With proper setup, IFWI can also be well suited for multi-scale joint geophysical inversion.

physics.geo-ph↗

arXiv2022

A Survey on Text-to-SQL Parsing: Concepts, Methods, and Future Directions

Text-to-SQL parsing is an essential and challenging task. The goal of text-to-SQL parsing is to convert a natural language (NL) question to its corresponding structured query language (SQL) based on the evidences provided by relational databases. Early text-to-SQL parsing systems from the database community achieved a noticeable progress with the cost of heavy human engineering and user interactions with the systems. In recent years, deep neural networks have significantly advanced this task by neural generation models, which automatically learn a mapping function from an input NL question to an output SQL query. Subsequently, the large pre-trained language models have taken the state-of-the-art of the text-to-SQL parsing task to a new level. In this survey, we present a comprehensive review on deep learning approaches for text-to-SQL parsing. First, we introduce the text-to-SQL parsing corpora which can be categorized as single-turn and multi-turn. Second, we provide a systematical overview of pre-trained language models and existing methods for text-to-SQL parsing. Third, we present readers with the challenges faced by text-to-SQL parsing and explore some potential future directions in this field.

cs.CL↗

arXiv2022

Data-Driven Control of Distributed Event-Triggered Network Systems

The present paper deals with data-driven event-triggered control of a class of unknown discrete-time interconnected systems (a.k.a. network systems). To this end, we start by putting forth a novel distributed event-triggering transmission strategy based on periodic sampling, under which a model-based stability criterion for the closed-loop network system is derived, by leveraging a discrete-time looped-functional approach. Marrying the model-based criterion with a data-driven system representation recently developed in the literature, a purely data-driven stability criterion expressed in the form of linear matrix inequalities (LMIs) is established. Meanwhile, the data-driven stability criterion suggests a means for co-designing the event-triggering coefficient matrix and the feedback control gain matrix using only some offline collected state-input data. Finally, numerical results corroborate the efficacy of the proposed distributed data-driven ETS in cutting off data transmissions and the co-design procedure.

eess.SY↗

arXiv2022

Differentiable Architecture Search with Random Features

Differentiable architecture search (DARTS) has significantly promoted the development of NAS techniques because of its high search efficiency and effectiveness but suffers from performance collapse. In this paper, we make efforts to alleviate the performance collapse problem for DARTS from two aspects. First, we investigate the expressive power of the supernet in DARTS and then derive a new setup of DARTS paradigm with only training BatchNorm. Second, we theoretically find that random features dilute the auxiliary connection role of skip-connection in supernet optimization and enable search algorithm focus on fairer operation selection, thereby solving the performance collapse problem. We instantiate DARTS and PC-DARTS with random features to build an improved version for each named RF-DARTS and RF-PCDARTS respectively. Experimental results show that RF-DARTS obtains \textbf{94.36\%} test accuracy on CIFAR-10 (which is the nearest optimal result in NAS-Bench-201), and achieves the newest state-of-the-art top-1 test error of \textbf{24.0\%} on ImageNet when transferring from CIFAR-10. Moreover, RF-DARTS performs robustly across three datasets (CIFAR-10, CIFAR-100, and SVHN) and four search spaces (S1-S4). Besides, RF-PCDARTS achieves even better results on ImageNet, that is, \textbf{23.9\%} top-1 and \textbf{7.1\%} top-5 test error, surpassing representative methods like single-path, training-free, and partial-channel paradigms directly searched on ImageNet.

cs.CV↗

arXiv2022

Distributed Momentum-based Frank-Wolfe Algorithm for Stochastic Optimization

This paper considers distributed stochastic optimization, in which a number of agents cooperate to optimize a global objective function through local computations and information exchanges with neighbors over a network. Stochastic optimization problems are usually tackled by variants of projected stochastic gradient descent. However, projecting a point onto a feasible set is often expensive. The Frank-Wolfe (FW) method has well-documented merits in handling convex constraints, but existing stochastic FW algorithms are basically developed for centralized settings. In this context, the present work puts forth a distributed stochastic Frank-Wolfe solver, by judiciously combining Nesterov's momentum and gradient tracking techniques for stochastic convex and nonconvex optimization over networks. It is shown that the convergence rate of the proposed algorithm is $\mathcal{O}(k^{-\frac{1}{2}})$ for convex optimization, and $\mathcal{O}(1/\mathrm{log}_2(k))$ for nonconvex optimization. The efficacy of the algorithm is demonstrated by numerical simulations against a number of competing alternatives.

math.OC↗

arXiv2022

Simple Baselines for Image Restoration

Although there have been significant advances in the field of image restoration recently, the system complexity of the state-of-the-art (SOTA) methods is increasing as well, which may hinder the convenient analysis and comparison of methods. In this paper, we propose a simple baseline that exceeds the SOTA methods and is computationally efficient. To further simplify the baseline, we reveal that the nonlinear activation functions, e.g. Sigmoid, ReLU, GELU, Softmax, etc. are not necessary: they could be replaced by multiplication or removed. Thus, we derive a Nonlinear Activation Free Network, namely NAFNet, from the baseline. SOTA results are achieved on various challenging benchmarks, e.g. 33.69 dB PSNR on GoPro (for image deblurring), exceeding the previous SOTA 0.38 dB with only 8.4% of its computational costs; 40.30 dB PSNR on SIDD (for image denoising), exceeding the previous SOTA 0.28 dB with less than half of its computational costs. The code and the pre-trained models are released at https://github.com/megvii-research/NAFNet.

cs.CV↗

arXiv2022

Event-triggered Consensus Control of Heterogeneous Multi-agent Systems: Model- and Data-based Analysis

This article deals with model- and data-based consensus control of heterogenous leader-following multi-agent systems (MASs) under an event-triggering transmission scheme. A dynamic periodic transmission protocol is developed to significantly alleviate the transmission frequency and computational burden, where the followers can interact locally with each other approaching the dynamics of the leader. Capitalizing on a discrete-time looped-functional, a model-based consensus condition for the closed-loop MASs is derived in form of linear matrix inequalities (LMIs), as well as a design method for obtaining the distributed controllers and event-triggering parameters. Upon collecting noise-corrupted state-input measurements during open-loop operation, a data-driven leader-following MAS representation is presented, and employed to solve the data-driven consensus control problem without requiring any knowledge of the agents' models. This result is then extended to the case of guaranteeing an $\mathcal{H}_{\infty}$ performance. A simulation example is finally given to corroborate the efficacy of the proposed distributed event-triggering scheme in cutting off data transmissions and the data-driven design method.

eess.SY↗

arXiv2022

StreamYOLO: Real-time Object Detection for Streaming Perception

The perceptive models of autonomous driving require fast inference within a low latency for safety. While existing works ignore the inevitable environmental changes after processing, streaming perception jointly evaluates the latency and accuracy into a single metric for video online perception, guiding the previous works to search trade-offs between accuracy and speed. In this paper, we explore the performance of real time models on this metric and endow the models with the capacity of predicting the future, significantly improving the results for streaming perception. Specifically, we build a simple framework with two effective modules. One is a Dual Flow Perception module (DFP). It consists of dynamic flow and static flow in parallel to capture moving tendency and basic detection feature, respectively. Trend Aware Loss (TAL) is the other module which adaptively generates loss weight for each object with its moving speed. Realistically, we consider multiple velocities driving scene and further propose Velocity-awared streaming AP (VsAP) to jointly evaluate the accuracy. In this realistic setting, we design a efficient mix-velocity training strategy to guide detector perceive any velocities. Our simple method achieves the state-of-the-art performance on Argoverse-HD dataset and improves the sAP and VsAP by 4.7% and 8.2% respectively compared to the strong baseline, validating its effectiveness.

cs.CV↗

arXiv2022

PETR: Position Embedding Transformation for Multi-View 3D Object Detection

In this paper, we develop position embedding transformation (PETR) for multi-view 3D object detection. PETR encodes the position information of 3D coordinates into image features, producing the 3D position-aware features. Object query can perceive the 3D position-aware features and perform end-to-end object detection. PETR achieves state-of-the-art performance (50.4% NDS and 44.1% mAP) on standard nuScenes dataset and ranks 1st place on the benchmark. It can serve as a simple yet strong baseline for future research. Code is available at \url{https://github.com/megvii-research/PETR}.

cs.CV↗

arXiv2022

Dense Teacher: Dense Pseudo-Labels for Semi-supervised Object Detection

To date, the most powerful semi-supervised object detectors (SS-OD) are based on pseudo-boxes, which need a sequence of post-processing with fine-tuned hyper-parameters. In this work, we propose replacing the sparse pseudo-boxes with the dense prediction as a united and straightforward form of pseudo-label. Compared to the pseudo-boxes, our Dense Pseudo-Label (DPL) does not involve any post-processing method, thus retaining richer information. We also introduce a region selection technique to highlight the key information while suppressing the noise carried by dense labels. We name our proposed SS-OD algorithm that leverages the DPL as Dense Teacher. On COCO and VOC, Dense Teacher shows superior performance under various settings compared with the pseudo-box-based methods.

cs.CV↗

arXiv2022

Data-driven Self-triggered Control via Trajectory Prediction

Self-triggered control, a well-documented technique for reducing the communication overhead while ensuring desired system performance, is gaining increasing popularity. However, existing methods for self-triggered control require explicit system models that are assumed perfectly known a priori. An end-to-end control paradigm known as data-driven control learns control laws directly from data, and offers a competing alternative to the routine system identification-then-control method. In this context, the present paper puts forth data-driven self-triggered control schemes for unknown linear systems using data collected offline. Specifically, for output feedback control systems, a data-driven model predictive control (MPC) scheme is proposed, which computes a sequence of control inputs while generating a predicted system trajectory. A data-driven self-triggering law is designed using the predicted trajectory, to determine the next triggering time once a new measurement becomes available. For state feedback control systems, instead of capitalizing on MPC to predict the trajectory, a data-fitting problem using the pre-collected input-state data is solved, whose solution is employed to construct the self-triggering mechanism. Both feasibility and stability are established for the proposed self-triggered controllers, which are validated using numerical examples.

eess.SY↗

arXiv2022

Improving Meta-learning for Low-resource Text Classification and Generation via Memory Imitation

Building models of natural language processing (NLP) is challenging in low-resource scenarios where only limited data are available. Optimization-based meta-learning algorithms achieve promising results in low-resource scenarios by adapting a well-generalized model initialization to handle new tasks. Nonetheless, these approaches suffer from the memorization overfitting issue, where the model tends to memorize the meta-training tasks while ignoring support sets when adapting to new tasks. To address this issue, we propose a memory imitation meta-learning (MemIML) method that enhances the model's reliance on support sets for task adaptation. Specifically, we introduce a task-specific memory module to store support set information and construct an imitation module to force query sets to imitate the behaviors of some representative support-set samples stored in the memory. A theoretical analysis is provided to prove the effectiveness of our method, and empirical results also demonstrate that our method outperforms competitive baselines on both text classification and generation tasks.

cs.CL↗

arXiv2022

Layout-Aware Information Extraction for Document-Grounded Dialogue: Dataset, Method and Demonstration

Building document-grounded dialogue systems have received growing interest as documents convey a wealth of human knowledge and commonly exist in enterprises. Wherein, how to comprehend and retrieve information from documents is a challenging research problem. Previous work ignores the visual property of documents and treats them as plain text, resulting in incomplete modality. In this paper, we propose a Layout-aware document-level Information Extraction dataset, LIE, to facilitate the study of extracting both structural and semantic knowledge from visually rich documents (VRDs), so as to generate accurate responses in dialogue systems. LIE contains 62k annotations of three extraction tasks from 4,061 pages in product and official documents, becoming the largest VRD-based information extraction dataset to the best of our knowledge. We also develop benchmark methods that extend the token-based language model to consider layout features like humans. Empirical results show that layout is critical for VRD-based extraction, and system demonstration also verifies that the extracted knowledge can help locate the answers that users care about.

cs.CL↗

arXiv2022

GPUMD: A package for constructing accurate machine-learned potentials and performing highly efficient atomistic simulations

We present our latest advancements of machine-learned potentials (MLPs) based on the neuroevolution potential (NEP) framework introduced in [Fan et al., Phys. Rev. B 104, 104309 (2021)] and their implementation in the open-source package GPUMD. We increase the accuracy of NEP models both by improving the radial functions in the atomic-environment descriptor using a linear combination of Chebyshev basis functions and by extending the angular descriptor with some four-body and five-body contributions as in the atomic cluster expansion approach. We also detail our efficient implementation of the NEP approach in graphics processing units as well as our workflow for the construction of NEP models, and we demonstrate their application in large-scale atomistic simulations. By comparing to state-of-the-art MLPs, we show that the NEP approach not only achieves above-average accuracy but also is far more computationally efficient. These results demonstrate that the GPUMD package is a promising tool for solving challenging problems requiring highly accurate, large-scale atomistic simulations. To enable the construction of MLPs using a minimal training set, we propose an active-learning scheme based on the latent space of a pre-trained NEP model. Finally, we introduce three separate Python packages, GPYUMD, CALORINE, and PYNEP, which enable the integration of GPUMD into Python workflows.

physics.comp-ph↗

arXiv2022

A Survey on Neural Open Information Extraction: Current Status and Future Directions

Open Information Extraction (OpenIE) facilitates domain-independent discovery of relational facts from large corpora. The technique well suits many open-world natural language understanding scenarios, such as automatic knowledge base construction, open-domain question answering, and explicit reasoning. Thanks to the rapid development in deep learning technologies, numerous neural OpenIE architectures have been proposed and achieve considerable performance improvement. In this survey, we provide an extensive overview of the-state-of-the-art neural OpenIE models, their key design decisions, strengths and weakness. Then, we discuss limitations of current solutions and the open issues in OpenIE problem itself. Finally we list recent trends that could help expand its scope and applicability, setting up promising directions for future research in OpenIE. To our best knowledge, this paper is the first review on this specific topic.

cs.CL↗

arXiv2022

Rebalanced Siamese Contrastive Mining for Long-Tailed Recognition

Deep neural networks perform poorly on heavily class-imbalanced datasets. Given the promising performance of contrastive learning, we propose Rebalanced Siamese Contrastive Mining (ResCom) to tackle imbalanced recognition. Based on the mathematical analysis and simulation results, we claim that supervised contrastive learning suffers a dual class-imbalance problem at both the original batch and Siamese batch levels, which is more serious than long-tailed classification learning. In this paper, at the original batch level, we introduce a class-balanced supervised contrastive loss to assign adaptive weights for different classes. At the Siamese batch level, we present a class-balanced queue, which maintains the same number of keys for all classes. Furthermore, we note that the imbalanced contrastive loss gradient with respect to the contrastive logits can be decoupled into the positives and negatives, and easy positives and easy negatives will make the contrastive gradient vanish. We propose supervised hard positive and negative pairs mining to pick up informative pairs for contrastive computation and improve representation learning. Finally, to approximately maximize the mutual information between the two views, we propose Siamese Balanced Softmax and joint it with the contrastive loss for one-stage training. Extensive experiments demonstrate that ResCom outperforms the previous methods by large margins on multiple long-tailed recognition benchmarks. Our code and models are made publicly available at: https://github.com/dvlab-research/ResCom.

cs.CV↗