arXiv subjects

Qing Li

Publications and source records attributed to Qing Li.

At least 469 records · Page 26Linked to original sources

arXiv2022

Cycle Encoding of a StyleGAN Encoder for Improved Reconstruction and Editability

GAN inversion aims to invert an input image into the latent space of a pre-trained GAN. Despite the recent advances in GAN inversion, there remain challenges to mitigate the tradeoff between distortion and editability, i.e. reconstructing the input image accurately and editing the inverted image with a small visual quality drop. The recently proposed pivotal tuning model makes significant progress towards reconstruction and editability, by using a two-step approach that first inverts the input image into a latent code, called pivot code, and then alters the generator so that the input image can be accurately mapped into the pivot code. Here, we show that both reconstruction and editability can be improved by a proper design of the pivot code. We present a simple yet effective method, named cycle encoding, for a high-quality pivot code. The key idea of our method is to progressively train an encoder in varying spaces according to a cycle scheme: W->W+->W. This training methodology preserves the properties of both W and W+ spaces, i.e. high editability of W and low distortion of W+. To further decrease the distortion, we also propose to refine the pivot code with an optimization-based method, where a regularization term is introduced to reduce the degradation in editability. Qualitative and quantitative comparisons to several state-of-the-art methods demonstrate the superiority of our approach.

cs.CV↗

arXiv2022

Mass and Age determination of the LAMOST data with different Machine Learning methods

We present a catalog of 948,216 stars with mass label and a catalog of 163,105 red clump (RC) stars with mass and age labels simultaneously. The training dataset is cross matched from the LAMOST (The Large Sky Area Multi-Object Fiber Spectroscopic Telescope) DR5 and high resolution asteroseismology data, mass and age are predicted by random forest method or convex hull algorithm. The stellar parameters with high correlation with mass and age are extracted and the test dataset shows that the median relative error of the prediction model for the mass of large sample is 3\% and meanwhile, the mass and age of red clump stars are 4\% and 7\%. We also compare the predicted age of red clump stars with the recent works and find that the final uncertainty of the RC sample could reach 18\% for age and 9\% for mass, in the meantime, final precision of the mass for large sample with different type of stars could reach 13\% without considering systematics, all these are implying that this method could be widely used in the future. Moreover, we explore the performance of different machine learning methods for our sample, including bayesian linear regression (BYS), gradient boosting decision Tree (GBDT), multilayer perceptron (MLP), multiple linear regression (MLR), random forest (RF) and support vector regression (SVR). Finally we find that the performance of nonlinear model is generally better than that of linear model, and the GBDT and RF methods are relatively better.

astro-ph.GA↗

arXiv2022

Pareto Optimization for Active Learning under Out-of-Distribution Data Scenarios

Pool-based Active Learning (AL) has achieved great success in minimizing labeling cost by sequentially selecting informative unlabeled samples from a large unlabeled data pool and querying their labels from oracle/annotators. However, existing AL sampling strategies might not work well in out-of-distribution (OOD) data scenarios, where the unlabeled data pool contains some data samples that do not belong to the classes of the target task. Achieving good AL performance under OOD data scenarios is a challenging task due to the natural conflict between AL sampling strategies and OOD sample detection. AL selects data that are hard to be classified by the current basic classifier (e.g., samples whose predicted class probabilities have high entropy), while OOD samples tend to have more uniform predicted class probabilities (i.e., high entropy) than in-distribution (ID) data. In this paper, we propose a sampling scheme, Monte-Carlo Pareto Optimization for Active Learning (POAL), which selects optimal subsets of unlabeled samples with fixed batch size from the unlabeled data pool. We cast the AL sampling task as a multi-objective optimization problem, and thus we utilize Pareto optimization based on two conflicting objectives: (1) the normal AL data sampling scheme (e.g., maximum entropy), and (2) the confidence of not being an OOD sample. Experimental results show its effectiveness on both classical Machine Learning (ML) and Deep Learning (DL) tasks.

cs.LG↗

arXiv2022

Saliency Attack: Towards Imperceptible Black-box Adversarial Attack

Deep neural networks are vulnerable to adversarial examples, even in the black-box setting where the attacker is only accessible to the model output. Recent studies have devised effective black-box attacks with high query efficiency. However, such performance is often accompanied by compromises in attack imperceptibility, hindering the practical use of these approaches. In this paper, we propose to restrict the perturbations to a small salient region to generate adversarial examples that can hardly be perceived. This approach is readily compatible with many existing black-box attacks and can significantly improve their imperceptibility with little degradation in attack success rate. Further, we propose the Saliency Attack, a new black-box attack aiming to refine the perturbations in the salient region to achieve even better imperceptibility. Extensive experiments show that compared to the state-of-the-art black-box attacks, our approach achieves much better imperceptibility scores, including most apparent distortion (MAD), $L_0$ and $L_2$ distances, and also obtains significantly higher success rates judged by a human-like threshold on MAD. Importantly, the perturbations generated by our approach are interpretable to some extent. Finally, it is also demonstrated to be robust to different detection-based defenses.

cs.LG↗

arXiv2022

SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing

Motivated by the success of T5 (Text-To-Text Transfer Transformer) in pre-trained natural language processing models, we propose a unified-modal SpeechT5 framework that explores the encoder-decoder pre-training for self-supervised speech/text representation learning. The SpeechT5 framework consists of a shared encoder-decoder network and six modal-specific (speech/text) pre/post-nets. After preprocessing the input speech/text through the pre-nets, the shared encoder-decoder network models the sequence-to-sequence transformation, and then the post-nets generate the output in the speech/text modality based on the output of the decoder. Leveraging large-scale unlabeled speech and text data, we pre-train SpeechT5 to learn a unified-modal representation, hoping to improve the modeling capability for both speech and text. To align the textual and speech information into this unified semantic space, we propose a cross-modal vector quantization approach that randomly mixes up speech/text states with latent units as the interface between encoder and decoder. Extensive evaluations show the superiority of the proposed SpeechT5 framework on a wide variety of spoken language processing tasks, including automatic speech recognition, speech synthesis, speech translation, voice conversion, speech enhancement, and speaker identification. We release our code and model at https://github.com/microsoft/SpeechT5.

eess.AS↗

arXiv2022

Simultaneous Double Q-learning with Conservative Advantage Learning for Actor-Critic Methods

Actor-critic Reinforcement Learning (RL) algorithms have achieved impressive performance in continuous control tasks. However, they still suffer two nontrivial obstacles, i.e., low sample efficiency and overestimation bias. To this end, we propose Simultaneous Double Q-learning with Conservative Advantage Learning (SDQ-CAL). Our SDQ-CAL boosts the Double Q-learning for off-policy actor-critic RL based on a modification of the Bellman optimality operator with Advantage Learning. Specifically, SDQ-CAL improves sample efficiency by modifying the reward to facilitate the distinction from experience between the optimal actions and the others. Besides, it mitigates the overestimation issue by updating a pair of critics simultaneously upon double estimators. Extensive experiments reveal that our algorithm realizes less biased value estimation and achieves state-of-the-art performance in a range of continuous control benchmark tasks. We release the source code of our method at: \url{https://github.com/LQNew/SDQ-CAL}.

cs.LG↗

arXiv2022

Attacking Black-box Recommendations via Copying Cross-domain User Profiles

Recently, recommender systems that aim to suggest personalized lists of items for users to interact with online have drawn a lot of attention. In fact, many of these state-of-the-art techniques have been deep learning based. Recent studies have shown that these deep learning models (in particular for recommendation systems) are vulnerable to attacks, such as data poisoning, which generates users to promote a selected set of items. However, more recently, defense strategies have been developed to detect these generated users with fake profiles. Thus, advanced injection attacks of creating more `realistic' user profiles to promote a set of items is still a key challenge in the domain of deep learning based recommender systems. In this work, we present our framework CopyAttack, which is a reinforcement learning based black-box attack method that harnesses real users from a source domain by copying their profiles into the target domain with the goal of promoting a subset of items. CopyAttack is constructed to both efficiently and effectively learn policy gradient networks that first select, and then further refine/craft, user profiles from the source domain to ultimately copy into the target domain. CopyAttack's goal is to maximize the hit ratio of the targeted items in the Top-$k$ recommendation list of the users in the target domain. We have conducted experiments on two real-world datasets and have empirically verified the effectiveness of our proposed framework and furthermore performed a thorough model analysis.

cs.IR↗

arXiv2022

Graph Trend Filtering Networks for Recommendations

Recommender systems aim to provide personalized services to users and are playing an increasingly important role in our daily lives. The key of recommender systems is to predict how likely users will interact with items based on their historical online behaviors, e.g., clicks, add-to-cart, purchases, etc. To exploit these user-item interactions, there are increasing efforts on considering the user-item interactions as a user-item bipartite graph and then performing information propagation in the graph via Graph Neural Networks (GNNs). Given the power of GNNs in graph representation learning, these GNNs-based recommendation methods have remarkably boosted the recommendation performance. Despite their success, most existing GNNs-based recommender systems overlook the existence of interactions caused by unreliable behaviors (e.g., random/bait clicks) and uniformly treat all the interactions, which can lead to sub-optimal and unstable performance. In this paper, we investigate the drawbacks (e.g., non-adaptive propagation and non-robustness) of existing GNN-based recommendation methods. To address these drawbacks, we introduce a principled graph trend collaborative filtering method and propose the Graph Trend Filtering Networks for recommendations (GTN) that can capture the adaptive reliability of the interactions. Comprehensive experiments and ablation studies are presented to verify and understand the effectiveness of the proposed framework. Our implementation based on PyTorch is available at https://github.com/wenqifan03/GTN-SIGIR2022.

cs.IR↗

arXiv2022

Comparative studies on unconventional superconductivity in Cr$_3$Ru compounds with bcc and A15 structures

Chromium (Cr) is a transition metal element with 3$d$ orbital electrons. In most compounds containing Cr, due to the correlation effect, twofold features of localization and itinerancy are expected. The localization gives rise to a magnetic moment, while the latter exhibits as the effective coherent weight for conductivity. Here we report the physical properties of Cr$_3$Ru compounds with bcc or A15 phases by using multiple experimental tools. The resistivity measurements show sharp superconducting transitions at $T_{\rm c}$ = 2.77 K and $T_{\rm c}$ = 3.37 K for the bcc and A15 structures. A high residual resistivity exists for both samples with the mean-free-path about 2 nm. Magnetization measurements also show rather narrow transitions, with a clear hump structure at high temperatures around 150 K, which may be ascribed to the remaining antiferromagnetic spin fluctuations. The specific heat coefficient reveals a clear jump at $T_{\rm c}$. We find that s-wave gaps can be adopted to fit the low temperature specific heat data yielding ratios of about $2Δ/k_{\scriptscriptstyle B}T_{\rm c}$ $\approx$ 3.6, indicating a moderate pairing strength. Interestingly, the Wilson ratios are 3.81 and 3.62 for the bcc and A15 phases, suggesting a moderate correlation effect of conducting electron in the normal state. Besides, for samples with A15 structure, another specific heat anomaly occurs at about 0.85 K and is sensitive to magnetic fields. By applying high pressures, both system exhibit an enhancement of $T_{\rm c}$ with a rate of about 0.019 K/GPa and 0.013 K/GPa for the bcc and A15 phases. We also conduct tunneling spectrum measurements and found that the coherence peaks are strongly smeared out. The strong suppression to the coherence peaks may be ascribed to the strong scattering. Our combinatory results point to an unconventional superconductivity in these Cr based compounds.

cond-mat.supr-con↗

arXiv2022

Indicative Image Retrieval: Turning Blackbox Learning into Grey

Deep learning became the game changer for image retrieval soon after it was introduced. It promotes the feature extraction (by representation learning) as the core of image retrieval, with the relevance/matching evaluation being degenerated into simple similarity metrics. In many applications, we need the matching evidence to be indicated rather than just have the ranked list (e.g., the locations of the target proteins/cells/lesions in medical images). It is like the matched words need to be highlighted in search engines. However, this is not easy to implement without explicit relevance/matching modeling. The deep representation learning models are not feasible because of their blackbox nature. In this paper, we revisit the importance of relevance/matching modeling in deep learning era with an indicative retrieval setting. The study shows that it is possible to skip the representation learning and model the matching evidence directly. By removing the dependency on the pre-trained models, it has avoided a lot of related issues (e.g., the domain gap between classification and retrieval, the detail-diffusion caused by convolution, and so on). More importantly, the study demonstrates that the matching can be explicitly modeled and backtracked later for generating the matching evidence indications. It can improve the explainability of deep inference. Our method obtains a best performance in literature on both Oxford-5k and Paris-6k, and sets a new record of 97.77% on Oxford-5k (97.81% on Paris-6k) without extracting any deep features.

cs.CV↗

arXiv2022

Observation of anomalous amplitude modes in the kagome metal CsV$_3$Sb$_5$

The charge-density wave (CDW) phase is often accompanied by the condensation of a soft acoustic phonon mode, giving rise to lattice distortion and charge density modulation. This picture was challenged for the recently discovered kagome metal CsV$_3$Sb$_5$, based on the evidence of absence of soft phonons. Here we report the observation of Raman-active CDW amplitude modes in this material, which are collective excitations typically thought to emerge out of frozen soft phonons. The amplitude modes strongly hybridize with other superlattice modes, imparting them with clear temperature-dependent frequency shift and broadening, rarely seen in other known CDW materials. Both the mode mixing and the large amplitude mode frequencies suggest that the CDW exhibits the character of strong electron-phonon coupling, a regime in which acoustic phonon softening can cease to exist. The observation of amplitude modes in the absence of soft phonons highlights the unconventional nature of the CDW in CsV$_3$Sb$_5$.

cond-mat.mtrl-sci↗

arXiv2022

Towards Sustainable Satellite Edge Computing

Recently, Low Earth Orbit (LEO) satellites experience rapid development and satellite edge computing emerges to address the limitation of bent-pipe architecture in existing satellite systems. Introducing energy-consuming computing components in satellite edge computing increases the depth of battery discharge. This will shorten batteries' life and influences the satellites' operation in orbit. In this paper, we aim to extend batteries' life by minimizing the depth of discharge for Earth observation missions. Facing the challenges of wireless uncertainty and energy harvesting dynamics, our work develops an online energy scheduling algorithm within an online convex optimization framework. Our algorithm achieves sub-linear regret and the constraint violation asymptotically approaches zero. Simulation results show that our algorithm can reduce the depth of discharge significantly.

eess.SY↗

arXiv2022

Tuning the competition between superconductivity and charge order in kagome superconductor Cs(V1-xNbx)3Sb5

The recently discovered coexistence of superconductivity and charge density wave order in the kagome systems AV3Sb5 (A = K, Rb, Cs) has stimulated enormous interest. According to theory, a vanadium-based kagome system may host a flat band, nontrivial linear dispersive Dirac surface states and electronic correlation. Despite intensive investigations, it remains controversial about the origin of the charge density wave (CDW) order, how does the superconductivity relate to the CDW, and whether the anomalous Hall effect (AHE) arises primarily from the kagome lattice or the CDW order. We report an extensive investigation on Cs(V1-xNbx)3Sb5 samples with systematic Nb doping. Our results show that the Nb doping induces apparent suppression of CDW order and promotes superconductivity; meanwhile, the AHE and magnetoresistance (MR) will be significantly weakened together with the CDW order. Combining with our density functional calculations, we interpret these effects by an antiphase shift of the Fermi energy with respect to the saddle points near M and the Fermi surface centered around Γ. It is found that the former depletes the filled states for the CDW instability and worsens the nesting condition for CDW order; while the latter lifts the Fermi level upward and enlarges the Fermi surface surrounding the Γ point, and thus promotes superconductivity. Our results uncover a delicate but unusual competition between the CDW order and superconductivity.

cond-mat.supr-con↗

arXiv2022

Superconductivity arising from pressure induced Lifshitz transition in Rb$_2$Pd$_3$Se$_4$ with kagome lattice

According to the Bardeen-Cooper-Schrieffer (BCS) theory, superconductivity usually needs well defined Fermi surface(s) with strong electron-phonon coupling and moderate quasiparticle density of states (DOS). A kagome lattice can host flat bands and topological Dirac bands; meanwhile, due to the parallel Fermi surfaces and the saddle points, many interesting orders are expected. Here, we report the observation of superconductivity by pressurizing a kagome compound Rb$_2$Pd$_3$Se$_4$ using a DAC anvil. The parent compound shows an insulating behavior; however, it gradually becomes metallic and turns to a superconducting state when a high pressure is applied. High pressure synchrotron measurements show that there is no structural transition occurring during this transition. The density-functional-theory (DFT) calculations illustrate that the insulating behavior of the parent phase is due to the crystalline field splitting of the partial Pd-4d $t_{2g}$ bands and the Se-derivative 4$p$-band. However, the threshold of metallicity and superconductivity are reached when the Lifshitz transition occurs, leading to the emergence of tiny Fermi surface at $Γ$ point. Our results point to an unconventional superconductivity and shed new light on understanding the electronic evolution of a kagome material.

cond-mat.supr-con↗

arXiv2022

An End-to-End Cascaded Image Deraining and Object Detection Neural Network

While the deep learning-based image deraining methods have made great progress in recent years, there are two major shortcomings in their application in real-world situations. Firstly, the gap between the low-level vision task represented by rain removal and the high-level vision task represented by object detection is significant, and the low-level vision task can hardly contribute to the high-level vision task. Secondly, the quality of the deraining dataset needs to be improved. In fact, the rain lines in many baselines have a large gap with the real rain lines, and the resolution of the deraining dataset images is generally not ideally. Meanwhile, there are few common datasets for both the low-level vision task and the high-level vision task. In this paper, we explore the combination of the low-level vision task with the high-level vision task. Specifically, we propose an end-to-end object detection network for reducing the impact of rainfall, which consists of two cascaded networks, an improved image deraining network and an object detection network, respectively. We also design the components of the loss function to accommodate the characteristics of the different sub-networks. We then propose a dataset based on the KITTI dataset for rainfall removal and object detection, on which our network surpasses the state-of-the-art with a significant improvement in metrics. Besides, our proposed network is measured on driving videos collected by self-driving vehicles and shows positive results for rain removal and object detection.

cs.CV↗

arXiv2022

Learning Bi-typed Multi-relational Heterogeneous Graph via Dual Hierarchical Attention Networks

Bi-type multi-relational heterogeneous graph (BMHG) is one of the most common graphs in practice, for example, academic networks, e-commerce user behavior graph and enterprise knowledge graph. It is a critical and challenge problem on how to learn the numerical representation for each node to characterize subtle structures. However, most previous studies treat all node relations in BMHG as the same class of relation without distinguishing the different characteristics between the intra-class relations and inter-class relations of the bi-typed nodes, causing the loss of significant structure information. To address this issue, we propose a novel Dual Hierarchical Attention Networks (DHAN) based on the bi-typed multi-relational heterogeneous graphs to learn comprehensive node representations with the intra-class and inter-class attention-based encoder under a hierarchical mechanism. Specifically, the former encoder aggregates information from the same type of nodes, while the latter aggregates node representations from its different types of neighbors. Moreover, to sufficiently model node multi-relational information in BMHG, we adopt a newly proposed hierarchical mechanism. By doing so, the proposed dual hierarchical attention operations enable our model to fully capture the complex structures of the bi-typed multi-relational heterogeneous graphs. Experimental results on various tasks against the state-of-the-arts sufficiently confirm the capability of DHAN in learning node representations on the BMHGs.

cs.LG↗

arXiv2022

Multi-View Self-Attention Based Transformer for Speaker Recognition

Initially developed for natural language processing (NLP), Transformer model is now widely used for speech processing tasks such as speaker recognition, due to its powerful sequence modeling capabilities. However, conventional self-attention mechanisms are originally designed for modeling textual sequence without considering the characteristics of speech and speaker modeling. Besides, different Transformer variants for speaker recognition have not been well studied. In this work, we propose a novel multi-view self-attention mechanism and present an empirical study of different Transformer variants with or without the proposed attention mechanism for speaker recognition. Specifically, to balance the capabilities of capturing global dependencies and modeling the locality, we propose a multi-view self-attention mechanism for speaker Transformer, in which different attention heads can attend to different ranges of the receptive field. Furthermore, we introduce and compare five Transformer variants with different network architectures, embedding locations, and pooling methods to learn speaker embeddings. Experimental results on the VoxCeleb1 and VoxCeleb2 datasets show that the proposed multi-view self-attention mechanism achieves improvement in the performance of speaker recognition, and the proposed speaker Transformer network attains excellent results compared with state-of-the-art models.

eess.AS↗

arXiv2022

Towards General Deep Leakage in Federated Learning

Unlike traditional central training, federated learning (FL) improves the performance of the global model by sharing and aggregating local models rather than local data to protect the users' privacy. Although this training approach appears secure, some research has demonstrated that an attacker can still recover private data based on the shared gradient information. This on-the-fly reconstruction attack deserves to be studied in depth because it can occur at any stage of training, whether at the beginning or at the end of model training; no relevant dataset is required and no additional models need to be trained. We break through some unrealistic assumptions and limitations to apply this reconstruction attack in a broader range of scenarios. We propose methods that can reconstruct the training data from shared gradients or weights, corresponding to the FedSGD and FedAvg usage scenarios, respectively. We propose a zero-shot approach to restore labels even if there are duplicate labels in the batch. We study the relationship between the label and image restoration. We find that image restoration fails even if there is only one incorrectly inferred label in the batch; we also find that when batch images have the same label, the corresponding image is restored as a fusion of that class of images. Our approaches are evaluated on classic image benchmarks, including CIFAR-10 and ImageNet. The batch size, image quality, and the adaptability of the label distribution of our approach exceed those of GradInversion, the state-of-the-art.

cs.LG↗