arXiv subjects

Xun Gong

Publications and source records attributed to Xun Gong.

At least 55 records · Page 3Linked to original sources

arXiv2023

Adversarial Driving Behavior Generation Incorporating Human Risk Cognition for Autonomous Vehicle Evaluation

Autonomous vehicle (AV) evaluation has been the subject of increased interest in recent years both in industry and in academia. This paper focuses on the development of a novel framework for generating adversarial driving behavior of background vehicle interfering against the AV to expose effective and rational risky events. Specifically, the adversarial behavior is learned by a reinforcement learning (RL) approach incorporated with the cumulative prospect theory (CPT) which allows representation of human risk cognition. Then, the extended version of deep deterministic policy gradient (DDPG) technique is proposed for training the adversarial policy while ensuring training stability as the CPT action-value function is leveraged. A comparative case study regarding the cut-in scenario is conducted on a high fidelity Hardware-in-the-Loop (HiL) platform and the results demonstrate the adversarial effectiveness to infer the weakness of the tested AV.

cs.AI↗

arXiv2023

On Data-Driven Modeling and Control in Modern Power Grids Stability: Survey and Perspective

Modern power grids are fast evolving with the increasing volatile renewable generation, distributed energy resources (DERs) and time-varying operating conditions. The DERs include rooftop photovoltaic (PV), small wind turbines, energy storages, flexible loads, electric vehicles (EVs), etc. The grid control is confronted with low inertia, uncertainty and nonlinearity that challenge the operation security, efficacy and efficiency. The ongoing digitization of power grids provides opportunities to address the challenges with data-driven and control. This paper provides a comprehensive review of emerging data-driven dynamical modeling and control methods and their various applications in power grid. Future trends are also discussed based on advances in data-driven control.

eess.SY↗

arXiv2023

An arithmetic Kontsevich--Zorich monodromy of a symmetric origami in genus 4

We demonstrate the existence of a certain genus four origami whose Kontsevich--Zorich monodromy is arithmetic in the sense of Sarnak. The surface is interesting because its Veech group is as large as possible and given by $\mathrm{SL}(2,\mathbb Z)$. When compared to other surfaces with Veech group $\mathrm{SL}(2,\mathbb Z)$ such as the Eierlegendre Wollmichsau and the Ornithorynque, an arithmetic Kontsevich--Zorich monodromy is surprising and indicates that there is little relationship between the Veech group and monodromy group of origamis. Additionally, we record the index and congruence level in the ambient symplectic group which gives data on what can appear in genus 4.

math.DS↗

arXiv2023

SpeechLM: Enhanced Speech Pre-Training with Unpaired Textual Data

How to boost speech pre-training with textual data is an unsolved problem due to the fact that speech and text are very different modalities with distinct characteristics. In this paper, we propose a cross-modal Speech and Language Model (SpeechLM) to explicitly align speech and text pre-training with a pre-defined unified discrete representation. Specifically, we introduce two alternative discrete tokenizers to bridge the speech and text modalities, including phoneme-unit and hidden-unit tokenizers, which can be trained using a small amount of paired speech-text data. Based on the trained tokenizers, we convert the unlabeled speech and text data into tokens of phoneme units or hidden units. The pre-training objective is designed to unify the speech and the text into the same discrete semantic space with a unified Transformer network. We evaluate SpeechLM on various spoken language processing tasks including speech recognition, speech translation, and universal representation evaluation framework SUPERB, demonstrating significant improvements on content-related tasks. Code and models are available at https://aka.ms/SpeechLM.

cs.CL↗

arXiv2023

RSIR Transformer: Hierarchical Vision Transformer using Random Sampling Windows and Important Region Windows

Recently, Transformers have shown promising performance in various vision tasks. However, the high costs of global self-attention remain challenging for Transformers, especially for high-resolution vision tasks. Local self-attention runs attention computation within a limited region for the sake of efficiency, resulting in insufficient context modeling as their receptive fields are small. In this work, we introduce two new attention modules to enhance the global modeling capability of the hierarchical vision transformer, namely, random sampling windows (RS-Win) and important region windows (IR-Win). Specifically, RS-Win sample random image patches to compose the window, following a uniform distribution, i.e., the patches in RS-Win can come from any position in the image. IR-Win composes the window according to the weights of the image patches in the attention map. Notably, RS-Win is able to capture global information throughout the entire model, even in earlier, high-resolution stages. IR-Win enables the self-attention module to focus on important regions of the image and capture more informative features. Incorporated with these designs, RSIR-Win Transformer demonstrates competitive performance on common vision tasks.

cs.CV↗

arXiv2023

Positional Label for Self-Supervised Vision Transformer

Positional encoding is important for vision transformer (ViT) to capture the spatial structure of the input image. General effectiveness has been proven in ViT. In our work we propose to train ViT to recognize the positional label of patches of the input image, this apparently simple task actually yields a meaningful self-supervisory task. Based on previous work on ViT positional encoding, we propose two positional labels dedicated to 2D images including absolute position and relative position. Our positional labels can be easily plugged into various current ViT variants. It can work in two ways: (a) As an auxiliary training target for vanilla ViT (e.g., ViT-B and Swin-B) for better performance. (b) Combine the self-supervised ViT (e.g., MAE) to provide a more powerful self-supervised signal for semantic feature learning. Experiments demonstrate that with the proposed self-supervised methods, ViT-B and Swin-B gain improvements of 1.20% (top-1 Acc) and 0.74% (top-1 Acc) on ImageNet, respectively, and 6.15% and 1.14% improvement on Mini-ImageNet.

cs.CV↗

arXiv2023

A Novel Koopman-Inspired Method for the Secondary Control of Microgrids with Grid-Forming and Grid-Following Sources

This paper proposes an online data-driven Koopman-inspired identification and control method for microgrid secondary voltage and frequency control. Unlike typical data-driven methods, the proposed method requires no warm-up training yet with guaranteed bounded-input-bounded-output (BIBO) stability and even asymptotic stability under some mild conditions. The proposed method estimates the Koopman state space model adaptively so as to perform effective secondary voltage and frequency control that can handle microgrid nonlinearity and uncertainty. Case studies in the 4-bus and 13-bus microgrid test systems (with grid-forming and grid-following sources) demonstrate the effectiveness and robustness of the proposed identification and control method subject to the change of operating conditions and large disturbances (e.g., microgrid mode transitions, generation/load variations) even with measurement noises and time delays.

eess.SY↗

arXiv2022

LongFNT: Long-form Speech Recognition with Factorized Neural Transducer

Traditional automatic speech recognition~(ASR) systems usually focus on individual utterances, without considering long-form speech with useful historical information, which is more practical in real scenarios. Simply attending longer transcription history for a vanilla neural transducer model shows no much gain in our preliminary experiments, since the prediction network is not a pure language model. This motivates us to leverage the factorized neural transducer structure, containing a real language model, the vocabulary predictor. We propose the {LongFNT-Text} architecture, which fuses the sentence-level long-form features directly with the output of the vocabulary predictor and then embeds token-level long-form features inside the vocabulary predictor, with a pre-trained contextual encoder RoBERTa to further boost the performance. Moreover, we propose the {LongFNT} architecture by extending the long-form speech to the original speech input and achieve the best performance. The effectiveness of our LongFNT approach is validated on LibriSpeech and GigaSpeech corpora with 19% and 12% relative word error rate~(WER) reduction, respectively.

cs.SD↗

arXiv2022

Axially Expanded Windows for Local-Global Interaction in Vision Transformers

Recently, Transformers have shown promising performance in various vision tasks. A challenging issue in Transformer design is that global self-attention is very expensive to compute, especially for the high-resolution vision tasks. Local self-attention performs attention computation within a local region to improve its efficiency, which leads to their receptive fields in a single attention layer are not large enough, resulting in insufficient context modeling. When observing a scene, humans usually focus on a local region while attending to non-attentional regions at coarse granularity. Based on this observation, we develop the axially expanded window self-attention mechanism that performs fine-grained self-attention within the local window and coarse-grained self-attention in the horizontal and vertical axes, and thus can effectively capturing both short- and long-range visual dependencies.

cs.CV↗

arXiv2022

BoundaryFace: A mining framework with noise label self-correction for Face Recognition

Face recognition has made tremendous progress in recent years due to the advances in loss functions and the explosive growth in training sets size. A properly designed loss is seen as key to extract discriminative features for classification. Several margin-based losses have been proposed as alternatives of softmax loss in face recognition. However, two issues remain to consider: 1) They overlook the importance of hard sample mining for discriminative learning. 2) Label noise ubiquitously exists in large-scale datasets, which can seriously damage the model's performance. In this paper, starting from the perspective of decision boundary, we propose a novel mining framework that focuses on the relationship between a sample's ground truth class center and its nearest negative class center. Specifically, a closed-set noise label self-correction module is put forward, making this framework work well on datasets containing a lot of label noise. The proposed method consistently outperforms SOTA methods in various face recognition benchmarks. Training code has been released at https://github.com/SWJTU-3DVision/BoundaryFace.

cs.CV↗

arXiv2022

Self-Supervised Implicit Attention: Guided Attention by The Model Itself

We propose Self-Supervised Implicit Attention (SSIA), a new approach that adaptively guides deep neural network models to gain attention by exploiting the properties of the models themselves. SSIA is a novel attention mechanism that does not require any extra parameters, computation, or memory access costs during inference, which is in contrast to existing attention mechanism. In short, by considering attention weights as higher-level semantic information, we reconsidered the implementation of existing attention mechanisms and further propose generating supervisory signals from higher network layers to guide lower network layers for parameter updates. We achieved this by building a self-supervised learning task using the hierarchical features of the network itself, which only works at the training stage. To verify the effectiveness of SSIA, we performed a particular implementation (called an SSIA block) in convolutional neural network models and validated it on several image classification datasets. The experimental results show that an SSIA block can significantly improve the model performance, even outperforms many popular attention methods that require additional parameters and computation costs, such as Squeeze-and-Excitation and Convolutional Block Attention Module. Our implementation will be available on GitHub.

cs.CV↗

arXiv2022

Knowledge Transfer and Distillation from Autoregressive to Non-Autoregressive Speech Recognition

Modern non-autoregressive~(NAR) speech recognition systems aim to accelerate the inference speed; however, they suffer from performance degradation compared with autoregressive~(AR) models as well as the huge model size issue. We propose a novel knowledge transfer and distillation architecture that leverages knowledge from AR models to improve the NAR performance while reducing the model's size. Frame- and sequence-level objectives are well-designed for transfer learning. To further boost the performance of NAR, a beam search method on Mask-CTC is developed to enlarge the search space during the inference stage. Experiments show that the proposed NAR beam search relatively reduces CER by over 5% on AISHELL-1 benchmark with a tolerable real-time-factor~(RTF) increment. By knowledge transfer, the NAR student who has the same size as the AR teacher obtains relative CER reductions of 8/16% on AISHELL-1 dev/test sets, and over 25% relative WER reductions on LibriSpeech test-clean/other sets. Moreover, the ~9x smaller NAR models achieve ~25% relative CER/WER reductions on both AISHELL-1 and LibriSpeech benchmarks with the proposed knowledge transfer and distillation.

cs.SD↗

arXiv2022

An Online Data-Driven Method for Microgrid Secondary Voltage and Frequency Control with Ensemble Koopman Modeling

Low inertia, nonlinearity and a high level of uncertainty (varying topologies and operating conditions) pose challenges to microgrid (MG) systemwide operation. This paper proposes an online adaptive Koopman operator optimal control (AKOOC) method for MG secondary voltage and frequency control. Unlike typical data-driven methods that are data-hungry and lack guaranteed stability, the proposed AKOOC requires no warm-up training yet with guaranteed bounded-input-bounded-output (BIBO) stability and even asymptotical stability under some mild conditions. The proposed AKOOC is developed based on an ensemble Koopman state space modeling with full basis functions that combines both linear and nonlinear bases without the need of event detection or switching. An iterative learning method is also developed to exploit model parameters, ensuring the effectiveness and the adaptiveness of the designed control. Simulation studies in the 4-bus (with detailed inner-loop control) MG system and the 34-bus MG system showed improved modeling accuracy and control, verifying the effectiveness of the proposed method subject to various changes of operating conditions even with time delay, measurement noise, and missing measurements.

eess.SY↗

arXiv2022

The Fixed Sub-Center: A Better Way to Capture Data Complexity

Treating class with a single center may hardly capture data distribution complexities. Using multiple sub-centers is an alternative way to address this problem. However, highly correlated sub-classes, the classifier's parameters grow linearly with the number of classes, and lack of intra-class compactness are three typical issues that need to be addressed in existing multi-subclass methods. To this end, we propose to use Fixed Sub-Center (F-SC), which allows the model to create more discrepant sub-centers while saving memory and cutting computational costs considerably. The F-SC specifically, first samples a class center Ui for each class from a uniform distribution, and then generates a normal distribution for each class, where the mean is equal to Ui. Finally, the sub-centers are sampled based on the normal distribution corresponding to each class, and the sub-centers are fixed during the training process avoiding the overhead of gradient calculation. Moreover, F-SC penalizes the Euclidean distance between the samples and their corresponding sub-centers, it helps remain intra-compactness. The experimental results show that F-SC significantly improves the accuracy of both image classification and fine-grained recognition tasks.

cs.CV↗

arXiv2022

Layer-wise Fast Adaptation for End-to-End Multi-Accent Speech Recognition

Accent variability has posed a huge challenge to automatic speech recognition~(ASR) modeling. Although one-hot accent vector based adaptation systems are commonly used, they require prior knowledge about the target accent and cannot handle unseen accents. Furthermore, simply concatenating accent embeddings does not make good use of accent knowledge, which has limited improvements. In this work, we aim to tackle these problems with a novel layer-wise adaptation structure injected into the E2E ASR model encoder. The adapter layer encodes an arbitrary accent in the accent space and assists the ASR model in recognizing accented speech. Given an utterance, the adaptation structure extracts the corresponding accent information and transforms the input acoustic feature into an accent-related feature through the linear combination of all accent bases. We further explore the injection position of the adaptation layer, the number of accent bases, and different types of accent bases to achieve better accent adaptation. Experimental results show that the proposed adaptation structure brings 12\% and 10\% relative word error rate~(WER) reduction on the AESRC2020 accent dataset and the Librispeech dataset, respectively, compared to the baseline.

cs.SD↗

arXiv2022

ReplaceBlock: An improved regularization method based on background information

Attention mechanism, being frequently used to train networks for better feature representations, can effectively disentangle the target object from irrelevant objects in the background. Given an arbitrary image, we find that the background's irrelevant objects are most likely to occlude/block the target object. We propose, based on this finding, a ReplaceBlock to simulate the situations when the target object is partially occluded by the objects that are deemed as background. Specifically, ReplaceBlock erases the target object in the image, and then generates a feature map with only irrelevant objects and background by the model. Finally, some regions in the background feature map are used to replace some regions of the target object in the original image feature map. In this way, ReplaceBlock can effectively simulate the feature map of the occluded image. The experimental results show that ReplaceBlock works better than DropBlock in regularizing convolutional networks.

cs.CV↗

arXiv2021

Cyclic three-level-pulse-area theorem for enantioselective state transfer of chiral molecules

We derive a pulse-area theorem for a cyclic three-level system, an archetypal model for exploring enantioselective state transfer (ESST) in chiral molecules driven by three linearly polarized microwave pulses. By dividing the closed-loop excitation into two separate stages, we obtain both amplitude and phase conditions of three control fields to generate high fidelity of ESST. As a proof of principle, we apply this pulse-area theorem to the cyclohexylmethanol molecules ($\text{C}_{7}\text{H}_{14}\text{O}$), for which three rotational states are connected by the $a$-type, $b$-type, and $c$-type components of the transition dipole moments in both center-frequency resonant and detuned conditions. Our results show that two enantiomers with opposite handedness can be transferred to different target states by designing three microwave pulses that satisfy the amplitude and phase conditions at the transition frequencies. The corresponding control schemes are robust against the time delays between the two stages. We suggest that the two control fields used in the second stage should be applied simultaneously for practical applications. This work contributes an alternative pulse-area theorem to the field of quantum control, which has the potential to determine the chirality of enantiomers in a mixture.

quant-ph↗

arXiv2021

An Irreversible Synthetic Route to an Ultra-Strong Two-Dimensional Polymer

Polymers that extend covalently in two dimensions have attracted recent attention as a means of combining the mechanical strength and in-plane energy conduction of conventional two-dimensional (2D) materials with the low densities, synthetic processability, and organic composition of their one-dimensional counterparts. Efforts to date have proven successful in forms that do not allow full realization of these properties, such as polymerization at flat interfaces or fixation of monomers in immobilized lattices. A frequently employed synthetic approach is to introduce microscopic reversibility, at the cost of bond stability, to achieve 2D crystals after extensive error correction. Herein we demonstrate a synthetic route to 2D irreversible polycondensation directly in the solution phase, resulting in covalently bonded 2D polymer platelets that are chemically stable and highly processable. Further fabrication offers highly oriented, free-standing films which exhibit exceptional 2D elastic modulus and yield strength at 50.9 +- 15.0 GPa and 0.976 +- 0.113 GPa, respectively. Platelet alignment is evidenced by polarized photoluminescence centered at 580 and 680 nm from different dipole transitions. This new synthetic route provides opportunities for 2D polymers in applications ranging from composite structures to molecular sieving membranes.

cond-mat.soft↗