arXiv subjects

Hong Chen

Publications and source records attributed to Hong Chen.

At least 199 records · Page 11Linked to original sources

arXiv2023

DIRE for Diffusion-Generated Image Detection

Diffusion models have shown remarkable success in visual synthesis, but have also raised concerns about potential abuse for malicious purposes. In this paper, we seek to build a detector for telling apart real images from diffusion-generated images. We find that existing detectors struggle to detect images generated by diffusion models, even if we include generated images from a specific diffusion model in their training data. To address this issue, we propose a novel image representation called DIffusion Reconstruction Error (DIRE), which measures the error between an input image and its reconstruction counterpart by a pre-trained diffusion model. We observe that diffusion-generated images can be approximately reconstructed by a diffusion model while real images cannot. It provides a hint that DIRE can serve as a bridge to distinguish generated and real images. DIRE provides an effective way to detect images generated by most diffusion models, and it is general for detecting generated images from unseen diffusion models and robust to various perturbations. Furthermore, we establish a comprehensive diffusion-generated benchmark including images generated by eight diffusion models to evaluate the performance of diffusion-generated image detectors. Extensive experiments on our collected benchmark demonstrate that DIRE exhibits superiority over previous generated-image detectors. The code and dataset are available at https://github.com/ZhendongWang6/DIRE.

cs.CV↗

arXiv2023

Stepdown SLOPE for Controlled Feature Selection

Sorted L-One Penalized Estimation (SLOPE) has shown the nice theoretical property as well as empirical behavior recently on the false discovery rate (FDR) control of high-dimensional feature selection by adaptively imposing the non-increasing sequence of tuning parameters on the sorted $\ell_1$ penalties. This paper goes beyond the previous concern limited to the FDR control by considering the stepdown-based SLOPE to control the probability of $k$ or more false rejections ($k$-FWER) and the false discovery proportion (FDP). Two new SLOPEs, called $k$-SLOPE and F-SLOPE, are proposed to realize $k$-FWER and FDP control respectively, where the stepdown procedure is injected into the SLOPE scheme. For the proposed stepdown SLOPEs, we establish their theoretical guarantees on controlling $k$-FWER and FDP under the orthogonal design setting, and also provide an intuitive guideline for the choice of regularization parameter sequence in much general setting. Empirical evaluations on simulated data validate the effectiveness of our approaches on controlled feature selection and support our theoretical findings.

math.ST↗

arXiv2023

On the Stability and Generalization of Triplet Learning

Triplet learning, i.e. learning from triplet data, has attracted much attention in computer vision tasks with an extremely large number of categories, e.g., face recognition and person re-identification. Albeit with rapid progress in designing and applying triplet learning algorithms, there is a lacking study on the theoretical understanding of their generalization performance. To fill this gap, this paper investigates the generalization guarantees of triplet learning by leveraging the stability analysis. Specifically, we establish the first general high-probability generalization bound for the triplet learning algorithm satisfying the uniform stability, and then obtain the excess risk bounds of the order $O(n^{-\frac{1}{2}} \mathrm{log}n)$ for both stochastic gradient descent (SGD) and regularized risk minimization (RRM), where $2n$ is approximately equal to the number of training samples. Moreover, an optimistic generalization bound in expectation as fast as $O(n^{-1})$ is derived for RRM in a low noise case via the on-average stability analysis. Finally, our results are applied to triplet metric learning to characterize its theoretical underpinning.

stat.ML↗

arXiv2023

Stability-based Generalization Analysis for Mixtures of Pointwise and Pairwise Learning

Recently, some mixture algorithms of pointwise and pairwise learning (PPL) have been formulated by employing the hybrid error metric of "pointwise loss + pairwise loss" and have shown empirical effectiveness on feature selection, ranking and recommendation tasks. However, to the best of our knowledge, the learning theory foundation of PPL has not been touched in the existing works. In this paper, we try to fill this theoretical gap by investigating the generalization properties of PPL. After extending the definitions of algorithmic stability to the PPL setting, we establish the high-probability generalization bounds for uniformly stable PPL algorithms. Moreover, explicit convergence rates of stochastic gradient descent (SGD) and regularized risk minimization (RRM) for PPL are stated by developing the stability analysis technique of pairwise learning. In addition, the refined generalization bounds of PPL are obtained by replacing uniform stability with on-average stability.

cs.LG↗

arXiv2023

Wireless and Service Allocation for Mobile Computation Offloading with Task Deadlines

In mobile computation offloading (MCO), mobile devices (MDs) can choose to either execute tasks locally or to have them executed on a remote edge server (ES). This paper addresses the problem of assigning both the wireless communication bandwidth needed, along with the ES capacity that is used for the task execution, so that task completion time constraints are satisfied. The objective is to obtain these allocations so that the average power consumption of the mobile devices is minimized, subject to a cost budget constraint. The paper includes contributions for both soft and hard task completion deadline constraints. The problems are first formulated as mixed integer nonlinear programs (MINLPs). Approximate solutions are then obtained by decomposing the problems into a collection of convex subproblems that can be efficiently solved. Results are presented that demonstrate the quality of the proposed solutions, which can achieve near optimum performance over a wide range of system parameters.

cs.NI↗

arXiv2023

Global Weighted Tensor Nuclear Norm for Tensor Robust Principal Component Analysis

Tensor Robust Principal Component Analysis (TRPCA), which aims to recover a low-rank tensor corrupted by sparse noise, has attracted much attention in many real applications. This paper develops a new Global Weighted TRPCA method (GWTRPCA), which is the first approach simultaneously considers the significance of intra-frontal slice and inter-frontal slice singular values in the Fourier domain. Exploiting this global information, GWTRPCA penalizes the larger singular values less and assigns smaller weights to them. Hence, our method can recover the low-tubal-rank components more exactly. Moreover, we propose an effective adaptive weight learning strategy by a Modified Cauchy Estimator (MCE) since the weight setting plays a crucial role in the success of GWTRPCA. To implement the GWTRPCA method, we devise an optimization algorithm using an Alternating Direction Method of Multipliers (ADMM) method. Experiments on real-world datasets validate the effectiveness of our proposed method.

cs.LG↗

arXiv2022

PrivateRec: Differentially Private Training and Serving for Federated News Recommendation

Collecting and training over sensitive personal data raise severe privacy concerns in personalized recommendation systems, and federated learning can potentially alleviate the problem by training models over decentralized user data.However, a theoretically private solution in both the training and serving stages of federated recommendation is essential but still lacking.Furthermore, naively applying differential privacy (DP) to the two stages in federated recommendation would fail to achieve a satisfactory trade-off between privacy and utility due to the high-dimensional characteristics of model gradients and hidden representations.In this work, we propose a federated news recommendation method for achieving a better utility in model training and online serving under a DP guarantee.We first clarify the DP definition over behavior data for each round in the life-circle of federated recommendation systems.Next, we propose a privacy-preserving online serving mechanism under this definition based on the idea of decomposing user embeddings with public basic vectors and perturbing the lower-dimensional combination coefficients. We apply a random behavior padding mechanism to reduce the required noise intensity for better utility. Besides, we design a federated recommendation model training method, which can generate effective and public basic vectors for serving while providing DP for training participants. We avoid the dimension-dependent noise for large models via label permutation and differentially private attention modules. Experiments on real-world news recommendation datasets validate that our method achieves superior utility under a DP guarantee in both training and serving of federated news recommendations.

cs.IR↗

arXiv2022

On the Global Solution of Soft k-Means

This paper presents an algorithm to solve the Soft k-Means problem globally. Unlike Fuzzy c-Means, Soft k-Means (SkM) has a matrix factorization-type objective and has been shown to have a close relation with the popular probability decomposition-type clustering methods, e.g., Left Stochastic Clustering (LSC). Though some work has been done for solving the Soft k-Means problem, they usually use an alternating minimization scheme or the projected gradient descent method, which cannot guarantee global optimality since the non-convexity of SkM. In this paper, we present a sufficient condition for a feasible solution of Soft k-Means problem to be globally optimal and show the output of the proposed algorithm satisfies it. Moreover, for the Soft k-Means problem, we provide interesting discussions on stability, solutions non-uniqueness, and connection with LSC. Then, a new model, named Minimal Volume Soft k-Means (MVSkM), is proposed to address the solutions non-uniqueness issue. Finally, experimental results support our theoretical results.

cs.LG↗

arXiv2022

Robust and Fast Measure of Information via Low-rank Representation

The matrix-based Rényi's entropy allows us to directly quantify information measures from given data, without explicit estimation of the underlying probability distribution. This intriguing property makes it widely applied in statistical inference and machine learning tasks. However, this information theoretical quantity is not robust against noise in the data, and is computationally prohibitive in large-scale applications. To address these issues, we propose a novel measure of information, termed low-rank matrix-based Rényi's entropy, based on low-rank representations of infinitely divisible kernel matrices. The proposed entropy functional inherits the specialty of of the original definition to directly quantify information from data, but enjoys additional advantages including robustness and effective calculation. Specifically, our low-rank variant is more sensitive to informative perturbations induced by changes in underlying distributions, while being insensitive to uninformative ones caused by noises. Moreover, low-rank Rényi's entropy can be efficiently approximated by random projection and Lanczos iteration techniques, reducing the overall complexity from $\mathcal{O}(n^3)$ to $\mathcal{O}(n^2 s)$ or even $\mathcal{O}(ns^2)$, where $n$ is the number of data samples and $s \ll n$. We conduct large-scale experiments to evaluate the effectiveness of this new information measure, demonstrating superior results compared to matrix-based Rényi's entropy in terms of both performance and computational efficiency.

cs.LG↗

arXiv2022

Universal Relations of Energy Flow, Acoustic Spin and Torque for Near-Field Acoustic Tweezers

Acoustic spin, radiation torque, energy flow, and reactive power are of significant importance from both fundamental and practical aspects, responsible for flexible tweezer manipulations and near-field sound directionality. Nevertheless, the intrinsic relations among these physical quantities are far from clear. Here, we prove the universal geometric relations among them in acoustics, independent on wave structure details. Particularly, we connect acoustic spin and torque to the cross product of time-averaged energy flow and reactive power, as well as to the local vorticity of energy flow. These relations are universally valid, verified in a variety of different acoustic systems. We also demonstrate the multipole mechanical torques and forces generated in three acoustic near-field sources: Janus, Huygens and Spin sources, applying on small lossy particles. These universal geometric relations uncover hidden locking relations beyond simple spin-momentum locking of near-field waves, and show the basic principles between the acoustic spin, radiation torque, and energy flow, reactive power.

cond-mat.other↗

arXiv2022

Elastic Valley Spin Controlled Chiral Coupling in Topological Valley Phononic Crystals

Distinct from the phononic valley pseudo-spin, the real physical spin of elastic waves adds a novel tool-kit capable of envisaging the valley-spin physics of topological valley phononic crystals from a local viewpoint. Here, we report the observation of local elastic valley spin as well as the hidden elastic spin-valley locking mechanism overlooked before. We demonstrate that the selective one-way routing of valley phonon states along the topological interface can be reversed by imposing the elastic spin meta-source at different interface locations with opposite valley-spin correspondence. We unveil the physical mechanism of selective directionality as the elastic spin controlled chiral coupling of valley phonon states, through both analytical theory and experimental measurement of the opposite local elastic spin density at different interface locations for different transport directions. The elastic spin of valley topological edge phonons can be extended to other topological states and offers new tool to explore topological metamaterials.

cond-mat.mes-hall↗

arXiv2022

Oracle-guided Contrastive Clustering

Deep clustering aims to learn a clustering representation through deep architectures. Most of the existing methods usually conduct clustering with the unique goal of maximizing clustering performance, that ignores the personalized demand of clustering tasks.% and results in unguided clustering solutions. However, in real scenarios, oracles may tend to cluster unlabeled data by exploiting distinct criteria, such as distinct semantics (background, color, object, etc.), and then put forward personalized clustering tasks. To achieve task-aware clustering results, in this study, Oracle-guided Contrastive Clustering(OCC) is then proposed to cluster by interactively making pairwise ``same-cluster" queries to oracles with distinctive demands. Specifically, inspired by active learning, some informative instance pairs are queried, and evaluated by oracles whether the pairs are in the same cluster according to their desired orientation. And then these queried same-cluster pairs extend the set of positive instance pairs for contrastive learning, guiding OCC to extract orientation-aware feature representation. Accordingly, the query results, guided by oracles with distinctive demands, may drive the OCC's clustering results in a desired orientation. Theoretically, the clustering risk in an active learning manner is given with a tighter upper bound, that guarantees active queries to oracles do mitigate the clustering risk. Experimentally, extensive results verify that OCC can cluster accurately along the specific orientation and it substantially outperforms the SOTA clustering methods as well. To the best of our knowledge, it is the first deep framework to perform personalized clustering.

cs.CV↗

arXiv2022

Character-Centric Story Visualization via Visual Planning and Token Alignment

Story visualization advances the traditional text-to-image generation by enabling multiple image generation based on a complete story. This task requires machines to 1) understand long text inputs and 2) produce a globally consistent image sequence that illustrates the contents of the story. A key challenge of consistent story visualization is to preserve characters that are essential in stories. To tackle the challenge, we propose to adapt a recent work that augments Vector-Quantized Variational Autoencoders (VQ-VAE) with a text-tovisual-token (transformer) architecture. Specifically, we modify the text-to-visual-token module with a two-stage framework: 1) character token planning model that predicts the visual tokens for characters only; 2) visual token completion model that generates the remaining visual token sequence, which is sent to VQ-VAE for finalizing image generations. To encourage characters to appear in the images, we further train the two-stage framework with a character-token alignment objective. Extensive experiments and evaluations demonstrate that the proposed method excels at preserving characters and can produce higher quality image sequences compared with the strong baselines. Codes can be found in https://github.com/sairin1202/VP-CSV

cs.CV↗

arXiv2022

StoryER: Automatic Story Evaluation via Ranking, Rating and Reasoning

Existing automatic story evaluation methods place a premium on story lexical level coherence, deviating from human preference. We go beyond this limitation by considering a novel \textbf{Story} \textbf{E}valuation method that mimics human preference when judging a story, namely \textbf{StoryER}, which consists of three sub-tasks: \textbf{R}anking, \textbf{R}ating and \textbf{R}easoning. Given either a machine-generated or a human-written story, StoryER requires the machine to output 1) a preference score that corresponds to human preference, 2) specific ratings and their corresponding confidences and 3) comments for various aspects (e.g., opening, character-shaping). To support these tasks, we introduce a well-annotated dataset comprising (i) 100k ranked story pairs; and (ii) a set of 46k ratings and comments on various aspects of the story. We finetune Longformer-Encoder-Decoder (LED) on the collected dataset, with the encoder responsible for preference score and aspect prediction and the decoder for comment generation. Our comprehensive experiments result in a competitive benchmark for each task, showing the high correlation to human preference. In addition, we have witnessed the joint learning of the preference scores, the aspect ratings, and the comments brings gain in each single task. Our dataset and benchmarks are publicly available to advance the research of story evaluation tasks.\footnote{Dataset and pre-trained model demo are available at anonymous website \url{http://storytelling-lab.com/eval} and \url{https://github.com/sairin1202/StoryER}}

cs.CL↗

arXiv2022

Microcantilever-integrated photonic circuits for broadband laser beam scanning

Laser beam scanning is central to many applications, including displays, microscopy, three-dimensional mapping, and quantum information. Reducing the scanners to microchip form factors has spurred the development of very-large-scale photonic integrated circuits of optical phased arrays and focal plane switched arrays. An outstanding challenge remains to simultaneously achieve a compact footprint, broad wavelength operation, and low power consumption. Here, we introduce a laser beam scanner that meets these requirements. Using microcantilevers embedded with silicon nitride nanophotonic circuitry, we demonstrate broadband, one- and two-dimensional steering of light with wavelengths from 410 nm to 700 nm. The microcantilevers have ultracompact ~0.1 mm$^2$ areas, consume ~31 to 46 mW of power, are simple to control, and emit a single light beam. The microcantilevers are monolithically integrated in an active photonic platform on 200-mm silicon wafers. The microcantilever-integrated photonic circuits miniaturize and simplify light projectors to enable versatile, power-efficient, and broadband laser scanner microchips.

physics.optics↗

arXiv2022

Monolithically integrated, broadband, high-efficiency silicon nitride-on-silicon waveguide photodetectors in a visible-light integrated photonics platform

Visible and near-infrared spectrum photonic integrated circuits are quickly becoming a key technology to address the scaling challenges in quantum information and biosensing. Thus far, integrated photonic platforms in this spectral range have lacked integrated photodetectors. Here, we report the first silicon nitride-on-silicon waveguide photodetectors that are monolithically integrated in a visible light photonic platform on silicon. Owing to a leaky-wave silicon nitride-on-silicon design, the devices achieved a high external quantum efficiency of > 60% across a record wavelength span from $λ$ ~400 nm to ~640 nm, an opto-electronic bandwidth up to 9 GHz, and an avalanche gain-bandwidth product up to 173 $\pm$ 30 GHz. As an example, a photodetector was integrated with a wavelength-tunable microring in a single chip for on-chip power monitoring.

physics.optics↗

arXiv2022

NeurIPS'22 Cross-Domain MetaDL competition: Design and baseline results

We present the design and baseline results for a new challenge in the ChaLearn meta-learning series, accepted at NeurIPS'22, focusing on "cross-domain" meta-learning. Meta-learning aims to leverage experience gained from previous tasks to solve new tasks efficiently (i.e., with better performance, little training data, and/or modest computational resources). While previous challenges in the series focused on within-domain few-shot learning problems, with the aim of learning efficiently N-way k-shot tasks (i.e., N class classification problems with k training examples), this competition challenges the participants to solve "any-way" and "any-shot" problems drawn from various domains (healthcare, ecology, biology, manufacturing, and others), chosen for their humanitarian and societal impact. To that end, we created Meta-Album, a meta-dataset of 40 image classification datasets from 10 domains, from which we carve out tasks with any number of "ways" (within the range 2-20) and any number of "shots" (within the range 1-20). The competition is with code submission, fully blind-tested on the CodaLab challenge platform. The code of the winners will be open-sourced, enabling the deployment of automated machine learning solutions for few-shot image classification across several domains.

cs.LG↗

arXiv2022

Bag of Tricks for Out-of-Distribution Generalization

Recently, out-of-distribution (OOD) generalization has attracted attention to the robustness and generalization ability of deep learning based models, and accordingly, many strategies have been made to address different aspects related to this issue. However, most existing algorithms for OOD generalization are complicated and specifically designed for certain dataset. To alleviate this problem, nicochallenge-2022 provides NICO++, a large-scale dataset with diverse context information. In this paper, based on systematic analysis of different schemes on NICO++ dataset, we propose a simple but effective learning framework via coupling bag of tricks, including multi-objective framework design, data augmentations, training and inference strategies. Our algorithm is memory-efficient and easily-equipped, without complicated modules and does not require for large pre-trained models. It achieves an excellent performance with Top-1 accuracy of 88.16% on public test set and 75.65% on private test set, and ranks 1st in domain generalization task of nicochallenge-2022.

cs.CV↗