arXiv subjects

Sheng Li

Publications and source records attributed to Sheng Li.

At least 199 records · Page 11Linked to original sources

arXiv2023

Image Protection for Robust Cropping Localization and Recovery

Existing image cropping detection schemes ignore that recovering the cropped-out contents can unveil the purpose of the behaved cropping attack. This paper presents \textbf{CLR}-Net, a novel image protection scheme addressing the combined challenge of image \textbf{C}ropping \textbf{L}ocalization and \textbf{R}ecovery. We first protect the original image by introducing imperceptible perturbations. Then, typical image post-processing attacks are simulated to erode the protected image. On the recipient's side, we predict the cropping mask and recover the original image. Besides, we propose a novel \textbf{F}ine-\textbf{G}rained generative \textbf{JPEG} simulator (FG-JPEG) as well as a feature alignment network to improve the real-world robustness. Comprehensive experiments prove that the quality of the recovered image and the accuracy of crop localization are both satisfactory.

eess.IV↗

arXiv2023

Mask-guided BERT for Few Shot Text Classification

Transformer-based language models have achieved significant success in various domains. However, the data-intensive nature of the transformer architecture requires much labeled data, which is challenging in low-resource scenarios (i.e., few-shot learning (FSL)). The main challenge of FSL is the difficulty of training robust models on small amounts of samples, which frequently leads to overfitting. Here we present Mask-BERT, a simple and modular framework to help BERT-based architectures tackle FSL. The proposed approach fundamentally differs from existing FSL strategies such as prompt tuning and meta-learning. The core idea is to selectively apply masks on text inputs and filter out irrelevant information, which guides the model to focus on discriminative tokens that influence prediction results. In addition, to make the text representations from different categories more separable and the text representations from the same category more compact, we introduce a contrastive learning loss function. Experimental results on public-domain benchmark datasets demonstrate the effectiveness of Mask-BERT.

cs.CL↗

arXiv2023

Coarse-to-fine Knowledge Graph Domain Adaptation based on Distantly-supervised Iterative Training

Modern supervised learning neural network models require a large amount of manually labeled data, which makes the construction of domain-specific knowledge graphs time-consuming and labor-intensive. In parallel, although there has been much research on named entity recognition and relation extraction based on distantly supervised learning, constructing a domain-specific knowledge graph from large collections of textual data without manual annotations is still an urgent problem to be solved. In response, we propose an integrated framework for adapting and re-learning knowledge graphs from one coarse domain (biomedical) to a finer-define domain (oncology). In this framework, we apply distant-supervision on cross-domain knowledge graph adaptation. Consequently, no manual data annotation is required to train the model. We introduce a novel iterative training strategy to facilitate the discovery of domain-specific named entities and triples. Experimental results indicate that the proposed framework can perform domain adaptation and construction of knowledge graph efficiently.

cs.AI↗

arXiv2023

Continual Causal Inference with Incremental Observational Data

The era of big data has witnessed an increasing availability of observational data from mobile and social networking, online advertising, web mining, healthcare, education, public policy, marketing campaigns, and so on, which facilitates the development of causal effect estimation. Although significant advances have been made to overcome the challenges in the academic area, such as missing counterfactual outcomes and selection bias, they only focus on source-specific and stationary observational data, which is unrealistic in most industrial applications. In this paper, we investigate a new industrial problem of causal effect estimation from incrementally available observational data and present three new evaluation criteria accordingly, including extensibility, adaptability, and accessibility. We propose a Continual Causal Effect Representation Learning method for estimating causal effects with observational data, which are incrementally available from non-stationary data distributions. Instead of having access to all seen observational data, our method only stores a limited subset of feature representations learned from previous data. Combining selective and balanced representation learning, feature representation distillation, and feature transformation, our method achieves the continual causal effect estimation for new data without compromising the estimation capability for original data. Extensive experiments demonstrate the significance of continual causal effect estimation and the effectiveness of our method.

cs.LG↗

arXiv2023

Steganography of Steganographic Networks

Steganography is a technique for covert communication between two parties. With the rapid development of deep neural networks (DNN), more and more steganographic networks are proposed recently, which are shown to be promising to achieve good performance. Unlike the traditional handcrafted steganographic tools, a steganographic network is relatively large in size. It raises concerns on how to covertly transmit the steganographic network in public channels, which is a crucial stage in the pipeline of steganography in real world applications. To address such an issue, we propose a novel scheme for steganography of steganographic networks in this paper. Unlike the existing steganographic schemes which focus on the subtle modification of the cover data to accommodate the secrets. We propose to disguise a steganographic network (termed as the secret DNN model) into a stego DNN model which performs an ordinary machine learning task (termed as the stego task). During the model disguising, we select and tune a subset of filters in the secret DNN model to preserve its function on the secret task, where the remaining filters are reactivated according to a partial optimization strategy to disguise the whole secret DNN model into a stego DNN model. The secret DNN model can be recovered from the stego DNN model when needed. Various experiments have been conducted to demonstrate the advantage of our proposed method for covert communication of steganographic networks as well as general DNN models.

cs.CR↗

arXiv2023

Multiple intercalation stages and universal Tc enhancement through polar organic species in electron-doped 1T-SnSe2

In this work, we report multiple intercalation stages and universal Tc enhancement of superconductivity in 1T-SnSe2 through Li and organic molecules cointercalation. We observe significantly increased lattice parameters up to 40 Å and dramatically enlarged interlayer distance up to ~11Å in Li and N,N-dimethylformamide (DMF) cointercalated SnSe2. Well-separated cointercalation stages with different stacking patterns have been discovered by carefully controlled reaction time and concentration of solutions. These cointercalation stages are superconductors showing different superconducting signals. In addition, Li and various organic species such as Acetone, Dimethyl sulfoxide (DMSO) and Tetrahydrofuran (THF) have been cointercalated into SnSe2 crystals, all of which show enhanced superconducting Tc compared to solely Li intercalated SnSe2. Our findings may provide more insight to effectively tune electronic structure of the lamellar structure through organic molecules co-regulation, and open a new strategy to engineer the physical properties of these layered materials by controlling their different intercalation stages.

cond-mat.supr-con↗

arXiv2023

Causal Effect Estimation: Recent Advances, Challenges, and Opportunities

Causal inference has numerous real-world applications in many domains, such as health care, marketing, political science, and online advertising. Treatment effect estimation, a fundamental problem in causal inference, has been extensively studied in statistics for decades. However, traditional treatment effect estimation methods may not well handle large-scale and high-dimensional heterogeneous data. In recent years, an emerging research direction has attracted increasing attention in the broad artificial intelligence field, which combines the advantages of traditional treatment effect estimation approaches (e.g., propensity score, matching, and reweighing) and advanced machine learning approaches (e.g., representation learning, adversarial learning, and graph neural networks). Although the advanced machine learning approaches have shown extraordinary performance in treatment effect estimation, it also comes with a lot of new topics and new research questions. In view of the latest research efforts in the causal inference field, we provide a comprehensive discussion of challenges and opportunities for the three core components of the treatment effect estimation task, i.e., treatment, covariates, and outcome. In addition, we showcase the promising research directions of this topic from multiple perspectives.

cs.LG↗

arXiv2023

Ideal Weak Topological Insulator and Protected Helical Saddle Points

The paradigm of classifying three-dimensional (3D) topological insulators into strong and weak ones (STI and WTI) opens the door for the discovery of various topological phases of matter protected by different symmetries and defined in different dimensions. However, in contrast to the vast realization of STIs, very few materials have been experimentally identified as being close to WTI. Even amongst those identified, none exists with topological surface states (TSS) exposed in a global bulk band gap that is stable at all temperatures. Here we report the design and observation of an ideal WTI in a quasi-one-dimensional (quasi-1D) bismuth halide, Bi$_{4}$I$_{1.2}$Br$_{2.8}$ (BIB). Via angle-resolved photoemission spectroscopy (ARPES), we identify that BIB hosts TSS on the (100)$\prime$ side surface in the form of two anisotropic $π$-offset Dirac cones (DCs) separated in momentum while topologically dark on the (001) top surface. The ARPES data fully determine a unique side-surface Hamiltonian and thereby identify two pairs of non-degenerate helical saddle points and a series of four Lifshitz transitions. The fact that both the surface Dirac and saddle points are in the global bulk band gap of 195 meV, combined with the small Dirac velocities, nontrivial spin texture, and the near-gap chemical potential, qualifies BIB to be not only an ideal WTI but also a fertile ground for topological many-body physics.

cond-mat.mtrl-sci↗

arXiv2023

Trojaning semi-supervised learning model via poisoning wild images on the web

Wild images on the web are vulnerable to backdoor (also called trojan) poisoning, causing machine learning models learned on these images to be injected with backdoors. Most previous attacks assumed that the wild images are labeled. In reality, however, most images on the web are unlabeled. Specifically, we study the effects of unlabeled backdoor images under semi-supervised learning (SSL) on widely studied deep neural networks. To be realistic, we assume that the adversary is zero-knowledge and that the semi-supervised learning model is trained from scratch. Firstly, we find the fact that backdoor poisoning always fails when poisoned unlabeled images come from different classes, which is different from poisoning the labeled images. The reason is that the SSL algorithms always strive to correct them during training. Therefore, for unlabeled images, we implement backdoor poisoning on images from the target class. Then, we propose a gradient matching strategy to craft poisoned images such that their gradients match the gradients of target images on the SSL model, which can fit poisoned images to the target class and realize backdoor injection. To the best of our knowledge, this may be the first approach to backdoor poisoning on unlabeled images of trained-from-scratch SSL models. Experiments show that our poisoning achieves state-of-the-art attack success rates on most SSL algorithms while bypassing modern backdoor defenses.

cs.CY↗

arXiv2022

Exploring Depth Information for Face Manipulation Detection

Face manipulation detection has been receiving a lot of attention for the reliability and security of the face images. Recent studies focus on using auxiliary information or prior knowledge to capture robust manipulation traces, which are shown to be promising. As one of the important face features, the face depth map, which has shown to be effective in other areas such as the face recognition or face detection, is unfortunately paid little attention to in literature for detecting the manipulated face images. In this paper, we explore the possibility of incorporating the face depth map as auxiliary information to tackle the problem of face manipulation detection in real world applications. To this end, we first propose a Face Depth Map Transformer (FDMT) to estimate the face depth map patch by patch from a RGB face image, which is able to capture the local depth anomaly created due to manipulation. The estimated face depth map is then considered as auxiliary information to be integrated with the backbone features using a Multi-head Depth Attention (MDA) mechanism that is newly designed. Various experiments demonstrate the advantage of our proposed method for face manipulation detection.

cs.CV↗

arXiv2022

Speech-text based multi-modal training with bidirectional attention for improved speech recognition

To let the state-of-the-art end-to-end ASR model enjoy data efficiency, as well as much more unpaired text data by multi-modal training, one needs to address two problems: 1) the synchronicity of feature sampling rates between speech and language (aka text data); 2) the homogeneity of the learned representations from two encoders. In this paper we propose to employ a novel bidirectional attention mechanism (BiAM) to jointly learn both ASR encoder (bottom layers) and text encoder with a multi-modal learning method. The BiAM is to facilitate feature sampling rate exchange, realizing the quality of the transformed features for the one kind to be measured in another space, with diversified objective functions. As a result, the speech representations are enriched with more linguistic information, while the representations generated by the text encoder are more similar to corresponding speech ones, and therefore the shared ASR models are more amenable for unpaired text data pretraining. To validate the efficacy of the proposed method, we perform two categories of experiments with or without extra unpaired text data. Experimental results on Librispeech corpus show it can achieve up to 6.15% word error rate reduction (WERR) with only paired data learning, while 9.23% WERR when more unpaired text data is employed.

eess.AS↗

arXiv2022

Learning to Immunize Images for Tamper Localization and Self-Recovery

Digital images are vulnerable to nefarious tampering attacks such as content addition or removal that severely alter the original meaning. It is somehow like a person without protection that is open to various kinds of viruses. Image immunization (Imuge) is a technology of protecting the images by introducing trivial perturbation, so that the protected images are immune to the viruses in that the tampered contents can be auto-recovered. This paper presents Imuge+, an enhanced scheme for image immunization. By observing the invertible relationship between image immunization and the corresponding self-recovery, we employ an invertible neural network to jointly learn image immunization and recovery respectively in the forward and backward pass. We also introduce an efficient attack layer that involves both malicious tamper and benign image post-processing, where a novel distillation-based JPEG simulator is proposed for improved JPEG robustness. Our method achieves promising results in real-world tests where experiments show accurate tamper localization as well as high-fidelity content recovery. Additionally, we show superior performance on tamper localization compared to state-of-the-art schemes based on passive forensics.

eess.IV↗

arXiv2022

Layer Freezing & Data Sieving: Missing Pieces of a Generic Framework for Sparse Training

Recently, sparse training has emerged as a promising paradigm for efficient deep learning on edge devices. The current research mainly devotes efforts to reducing training costs by further increasing model sparsity. However, increasing sparsity is not always ideal since it will inevitably introduce severe accuracy degradation at an extremely high sparsity level. This paper intends to explore other possible directions to effectively and efficiently reduce sparse training costs while preserving accuracy. To this end, we investigate two techniques, namely, layer freezing and data sieving. First, the layer freezing approach has shown its success in dense model training and fine-tuning, yet it has never been adopted in the sparse training domain. Nevertheless, the unique characteristics of sparse training may hinder the incorporation of layer freezing techniques. Therefore, we analyze the feasibility and potentiality of using the layer freezing technique in sparse training and find it has the potential to save considerable training costs. Second, we propose a data sieving method for dataset-efficient training, which further reduces training costs by ensuring only a partial dataset is used throughout the entire training process. We show that both techniques can be well incorporated into the sparse training algorithm to form a generic framework, which we dub SpFDE. Our extensive experiments demonstrate that SpFDE can significantly reduce training costs while preserving accuracy from three dimensions: weight sparsity, layer freezing, and dataset sieving.

cs.LG↗

arXiv2022

Room temperature spin-orbit torque efficiency and magnetization switching in SrRuO3-based heterostructures

Spin-orbit torques (SOTs) from transition metal oxides (TMOs) in conjunction with magnetic materials have recently attracted tremendous attention for realizing high-efficient spintronic devices. SrRuO3 is a promising candidate among TMOs due to its large and tunable SOT-efficiency as well as high conductivity and chemical stability. However, a further study for benchmarking the SOT-efficiency and realizing SOT-driven magnetization switching in SrRuO3 is still highly desired so far. Here, we systematically study the SOT properties of high-quality SrRuO3 thin film heterostructuring with different magnetic alloys of both IMA and PMA configuration by the harmonic Hall voltage technique. Our results indicate that SrRuO3 possesses pronounced SOT-efficiency of about 0.2 at room temperature regardless of the magnetic alloys, which is comparable to typical heavy metals (HMs). Furthermore, we achieve SOT-driven magnetization switching with a low threshold current density of 3.8x10^10 A/m^2, demonstrating the promising potential of SrRuO3 for practical devices. By making a comprehensive comparison with HMs, our work unambiguously benchmarks the SOT properties and concludes the advantages of SrRuO3, which may bring more diverse choices for SOT applications by utilizing hybrid-oxide/metal and all-oxide systems.

cond-mat.mtrl-sci↗

arXiv2022

Anisotropic linear and nonlinear charge-spin conversion in topological semimetal SrIrO3

Over the past decade, utilizing spin currents in the linear response of electric field to manipulate magnetization states via spin-orbit torques (SOTs) is one of the core concepts for realizing a multitude of spintronic devices. Besides the linear regime, recently, nonlinear charge-spin conversion under the square of electric field has been recognized in a wide variety of materials with nontrivial spin textures, opening an emerging field of nonlinear spintronics. Here, we report the investigation of both linear and nonlinear charge-spin conversion in one single topological semimetal SrIrO3(110) thin film that hosts strong spin-orbit coupling and nontrivial spin textures in the momentum space. In the nonlinear regime, the observation of crystalline direction dependent response indicates the presence of anisotropic surface states induced spin-momentum locking near the Fermi level. Such anisotropic spin textures also give rise to spin currents in the linear response regime, which mainly contributes to the fieldlike SOT component. Our work demonstrates the power of combination of linear and nonlinear approaches in understanding and utilizing charge-spin conversion in topological materials.

cond-mat.mtrl-sci↗

arXiv2022

Hierarchical Capsule Prediction Network for Marketing Campaigns Effect

Marketing campaigns are a set of strategic activities that can promote a business's goal. The effect prediction for marketing campaigns in a real industrial scenario is very complex and challenging due to the fact that prior knowledge is often learned from observation data, without any intervention for the marketing campaign. Furthermore, each subject is always under the interference of several marketing campaigns simultaneously. Therefore, we cannot easily parse and evaluate the effect of a single marketing campaign. To the best of our knowledge, there are currently no effective methodologies to solve such a problem, i.e., modeling an individual-level prediction task based on a hierarchical structure with multiple intertwined events. In this paper, we provide an in-depth analysis of the underlying parse tree-like structure involved in the effect prediction task and we further establish a Hierarchical Capsule Prediction Network (HapNet) for predicting the effects of marketing campaigns. Extensive results based on both the synthetic data and real data demonstrate the superiority of our model over the state-of-the-art methods and show remarkable practicability in real industrial applications.

stat.ML↗

arXiv2022

Generative Steganography Network

Steganography usually modifies cover media to embed secret data. A new steganographic approach called generative steganography (GS) has emerged recently, in which stego images (images containing secret data) are generated from secret data directly without cover media. However, existing GS schemes are often criticized for their poor performances. In this paper, we propose an advanced generative steganography network (GSN) that can generate realistic stego images without using cover images. We firstly introduce the mutual information mechanism in GS, which helps to achieve high secret extraction accuracy. Our model contains four sub-networks, i.e., an image generator ($G$), a discriminator ($D$), a steganalyzer ($S$), and a data extractor ($E$). $D$ and $S$ act as two adversarial discriminators to ensure the visual quality and security of generated stego images. $E$ is to extract the hidden secret from generated stego images. The generator $G$ is flexibly constructed to synthesize either cover or stego images with different inputs. It facilitates covert communication by concealing the function of generating stego images in a normal generator. A module named secret block is designed to hide secret data in the feature maps during image generation, with which high hiding capacity and image fidelity are achieved. In addition, a novel hierarchical gradient decay (HGD) skill is developed to resist steganalysis detection. Experiments demonstrate the superiority of our work over existing methods.

cs.CV↗

arXiv2022

On Generating Identifiable Virtual Faces

Face anonymization with generative models have become increasingly prevalent since they sanitize private information by generating virtual face images, ensuring both privacy and image utility. Such virtual face images are usually not identifiable after the removal or protection of the original identity. In this paper, we formalize and tackle the problem of generating identifiable virtual face images. Our virtual face images are visually different from the original ones for privacy protection. In addition, they are bound with new virtual identities, which can be directly used for face recognition. We propose an Identifiable Virtual Face Generator (IVFG) to generate the virtual face images. The IVFG projects the latent vectors of the original face images into virtual ones according to a user specific key, based on which the virtual face images are generated. To make the virtual face images identifiable, we propose a multi-task learning objective as well as a triplet styled training strategy to learn the IVFG. We evaluate the performance of our virtual face images using different face recognizers on diffident face image datasets, all of which demonstrate the effectiveness of the IVFG for generate identifiable virtual face images.

cs.CV↗