Search

Associativity of a two-place function $T: [0,1]^2\rightarrow [0,1]$ defined by $T(x,y)=f^{(-1)}(F(f(x),f(y)))$ where $F:[0,\infty]^2\rightarrow[0,\infty]$ is an associative function, $f: [0,1]\rightarrow [0,\infty]$ is a monotone function which satisfies either $f(x)=f(x^{+})$ when $f(x^{+})\in \mbox{Ran}(f)$ or $f(x)\neq f(y)$ for any $y\neq x$ when $f(x^{+})\notin \mbox{Ran}(f)$ for all $x\in[0,1]$ and $f^{(-1)}:[0,\infty]\rightarrow[0,1]$ is a pseudo-inverse of $f$ depends only on properties of the range of $f$. The necessary and sufficient conditions for the $T$ to be associative are presented by applying the properties of the monotone function $f$., Comment: 16

Vision transformers (ViTs) excel in computer vision for modeling long-term dependencies, yet face two key challenges for image quality assessment (IQA): discarding fine details during patch embedding, and requiring extensive training data due to lack of inductive biases. In this study, we propose a Global-Local progressive INTegration network for IQA, called GlintIQA, to address these issues through three key components: 1) Hybrid feature extraction combines ViT-based global feature extractor (VGFE) and convolutional neural networks (CNNs)-based local feature extractor (CLFE) to capture global coarse-grained features and local fine-grained features, respectively. The incorporation of CNNs mitigates the patch-level information loss and inductive bias constraints inherent to ViT architectures. 2) Progressive feature integration leverages diverse kernel sizes in embedding to spatially align coarse- and fine-grained features, and progressively aggregate these features by interactively stacking channel-wise attention and spatial enhancement modules to build effective quality-aware representations. 3) Content similarity-based labeling approach is proposed that automatically assigns quality labels to images with diverse content based on subjective quality scores. This addresses the scarcity of labeled training data in synthetic datasets and bolsters model generalization. The experimental results demonstrate the efficacy of our approach, yielding 5.04% average SROCC gains on cross-authentic dataset evaluations. Moreover, our model and its counterpart pre-trained on the proposed dataset respectively exhibited 5.40% and 13.23% improvements on across-synthetic datasets evaluation. The codes and proposed dataset will be released at https://github.com/XiaoqiWang/GlintIQA.

Knowledge distillation plays a key role in compressing the Large Language Models (LLMs), which boosts a small-size student model under large teacher models' guidance. However, existing LLM distillation methods overly rely on student-generated outputs, which may introduce generation errors and misguide the distillation process. Moreover, the distillation loss functions introduced in previous art struggle to align the most informative part due to the complex distribution of LLMs' outputs. To address these problems, we propose a multi-granularity semantic revision method for LLM distillation. At the sequence level, we propose a sequence correction and re-generation (SCRG) strategy. SCRG first calculates the semantic cognitive difference between the teacher and student to detect the error token, then corrects it with the teacher-generated one, and re-generates the sequence to reduce generation errors and enhance generation diversity. At the token level, we design a distribution adaptive clipping Kullback-Leibler (DAC-KL) loss as the distillation objective function. DAC-KL loss exploits a learnable sub-network to adaptively extract semantically dense areas from the teacher's output, avoiding the interference of redundant information in the distillation process. Finally, at the span level, we leverage the span priors of a sequence to compute the probability correlations within spans, and constrain the teacher and student's probability correlations to be consistent, further enhancing the transfer of semantic information. Extensive experiments across different model families with parameters ranging from 0.1B to 13B demonstrate the superiority of our method compared to existing methods.

Point cloud has been the mainstream representation for advanced 3D applications, such as virtual reality and augmented reality. However, the massive data amounts of point clouds is one of the most challenging issues for transmission and storage. In this paper, we propose an end-to-end voxel Transformer and Sparse Convolution based Point Cloud Attribute Compression (TSC-PCAC) for 3D broadcasting. Firstly, we present a framework of the TSC-PCAC, which include Transformer and Sparse Convolutional Module (TSCM) based variational autoencoder and channel context module. Secondly, we propose a two-stage TSCM, where the first stage focuses on modeling local dependencies and feature representations of the point clouds, and the second stage captures global features through spatial and channel pooling encompassing larger receptive fields. This module effectively extracts global and local interpoint relevance to reduce informational redundancy. Thirdly, we design a TSCM based channel context module to exploit interchannel correlations, which improves the predicted probability distribution of quantized latent representations and thus reduces the bitrate. Experimental results indicate that the proposed TSC-PCAC method achieves an average of 38.53%, 21.30%, and 11.19% Bjontegaard Delta bitrate reductions compared to the Sparse-PCAC, NF-PCAC, and G-PCC v23 methods, respectively. The encoding/decoding time costs are reduced up to 97.68%/98.78% on average compared to the Sparse-PCAC. The source code and the trained models of the TSC-PCAC are available at https://github.com/igizuxo/TSC-PCAC.

The Near-Earth binary asteroid Didymos was the target of a planetary defense demonstration mission DART in September 2022. The smaller binary component, Dimorphos, was impacted by the spacecraft in order to measure momentum transfer in kinetic impacts into rubble piles. DART and associated Earth-based observation campaigns have provided a wealth of scientific data on the Didymos-Dimorphos binary. DART revealed a largely oblate and ellipsoidal shape of Dimorphos before the impact, while the post-impact observations suggest that Dimorphos now has a prolate shape. Here we add those data points to the known properties of small binary asteroids and propose new paradigms of the radiative binary YORP (BYORP) effect as well as tidal dissipation in small binaries. We find that relatively spheroidal bodies like Dimorphos made of small debris may experience a weaker and more size-dependent BYORP effect than previously thought. This could explain the observed values of period drift in several well-characterized binaries. We also propose that energy dissipation in small binaries is dominated by relatively brief episodes of large-scale movement of (likely surface) materials, rather than long-term steady-state tidal dissipation. We propose that one such episode was triggered on Dimorphos by the DART impact. Depending on the longevity of this high-dissipation regime, it is possible that Dimorphos will be more dynamically relaxed in time for the Hera mission than it was in the weeks following the impact., Comment: Accepted for PSJ

Ultralight bosons can condense to form the so-called bosonic clouds around spinning black holes by superradiance instability. When quantum effects are taken into account, the classical black holes were replaced by exotic compact objects including area quantized black holes. In this work, we consider the superradiant instabilities of massive scalar fields around area quantized Kerr black hole. We introduce the reflectivity of area quantized black hole possesses the distinct discrete feature, and the scalar fields have the superradiant modes solution only within the specific mass range. In addition, the area quantization may terminate the superradiance when the black hole spins down, or even suppress the formation of the bosionic cloud., Comment: 10 pages, 5 figures

This paper provides a comprehensive review of the NTIRE 2024 challenge, focusing on efficient single-image super-resolution (ESR) solutions and their outcomes. The task of this challenge is to super-resolve an input image with a magnification factor of x4 based on pairs of low and corresponding high-resolution images. The primary objective is to develop networks that optimize various aspects such as runtime, parameters, and FLOPs, while still maintaining a peak signal-to-noise ratio (PSNR) of approximately 26.90 dB on the DIV2K_LSDIR_valid dataset and 26.99 dB on the DIV2K_LSDIR_test dataset. In addition, this challenge has 4 tracks including the main track (overall performance), sub-track 1 (runtime), sub-track 2 (FLOPs), and sub-track 3 (parameters). In the main track, all three metrics (ie runtime, FLOPs, and parameter count) were considered. The ranking of the main track is calculated based on a weighted sum-up of the scores of all other sub-tracks. In sub-track 1, the practical runtime performance of the submissions was evaluated, and the corresponding score was used to determine the ranking. In sub-track 2, the number of FLOPs was considered. The score calculated based on the corresponding FLOPs was used to determine the ranking. In sub-track 3, the number of parameters was considered. The score calculated based on the corresponding parameters was used to determine the ranking. RLFN is set as the baseline for efficiency measurement. The challenge had 262 registered participants, and 34 teams made valid submissions. They gauge the state-of-the-art in efficient single-image super-resolution. To facilitate the reproducibility of the challenge and enable other researchers to build upon these findings, the code and the pre-trained model of validated solutions are made publicly available at https://github.com/Amazingren/NTIRE2024_ESR/., Comment: The report paper of NTIRE2024 Efficient Super-resolution, accepted by CVPRW2024

The Weyl double copy builds the relation between gauge theory and gravity theory, especially the correspondence between gauge solutions and gravity solutions. In this paper, we obtain the slowly rotating charge solutions from Weyl double copy for the Kerr black hole with small Chern-Simons correction. Based on the Weyl double copy relation, for the Petrov type D solution, we find the additional correction to the electromagnetic field strength tensor of rotating charge. For the Petrov type I solution, we find that the additional electromagnetic field strength tensors have the exogenous properties, while the total sources vanish at the leading order., Comment: 9 pages

Knowledge distillation (KD) is a promising yet challenging model compression technique that transfers rich learning representations from a well-performing but cumbersome teacher model to a compact student model. Previous methods for image super-resolution (SR) mostly compare the feature maps directly or after standardizing the dimensions with basic algebraic operations (e.g. average, dot-product). However, the intrinsic semantic differences among feature maps are overlooked, which are caused by the disparate expressive capacity between the networks. This work presents MiPKD, a multi-granularity mixture of prior KD framework, to facilitate efficient SR model through the feature mixture in a unified latent space and stochastic network block mixture. Extensive experiments demonstrate the effectiveness of the proposed MiPKD method.

Adversarial training (AT) is currently one of the most effective ways to obtain the robustness of deep neural networks against adversarial attacks. However, most AT methods suffer from robust overfitting, i.e., a significant generalization gap in adversarial robustness between the training and testing curves. In this paper, we first identify a connection between robust overfitting and the excessive memorization of noisy labels in AT from a view of gradient norm. As such label noise is mainly caused by a distribution mismatch and improper label assignments, we are motivated to propose a label refinement approach for AT. Specifically, our Self-Guided Label Refinement first self-refines a more accurate and informative label distribution from over-confident hard labels, and then it calibrates the training by dynamically incorporating knowledge from self-distilled models into the current model and thus requiring no external teachers. Empirical results demonstrate that our method can simultaneously boost the standard accuracy and robust performance across multiple benchmark datasets, attack types, and architectures. In addition, we also provide a set of analyses from the perspectives of information theory to dive into our method and suggest the importance of soft labels for robust generalization., Comment: Accepted to CVPR 2024

In this paper, we retarget video stitching to an emerging issue, named warping shake, when extending image stitching to video stitching. It unveils the temporal instability of warped content in non-overlapping regions, despite image stitching having endeavored to preserve the natural structures. Therefore, in most cases, even if the input videos to be stitched are stable, the stitched video will inevitably cause undesired warping shakes and affect the visual experience. To eliminate the shakes, we propose StabStitch to simultaneously realize video stitching and video stabilization in a unified unsupervised learning framework. Starting from the camera paths in video stabilization, we first derive the expression of stitching trajectories in video stitching by elaborately integrating spatial and temporal warps. Then a warp smoothing model is presented to optimize them with a comprehensive consideration regarding content alignment, trajectory smoothness, spatial consistency, and online collaboration. To establish an evaluation benchmark and train the learning framework, we build a video stitching dataset with a rich diversity in camera motions and scenes. Compared with existing stitching solutions, StabStitch exhibits significant superiority in scene robustness and inference speed in addition to stitching and stabilization performance, contributing to a robust and real-time online video stitching system. The code and dataset are available at https://github.com/nie-lang/StabStitch., Comment: Aceepted to ECCV2024

The pulsar timings are sensitive to both the nanohertz gravitational-wave background and the oscillation of ultralight dark matter. The Hellings-Downs angular correlation curve provides a criterion to search for stochastic gravitational-wave backgrounds at nanohertz via pulsar timing arrays. We study the angular correlation of the timing residuals induced by the spin-2 ultralight dark matter, which is different from the usual Hellings-Downs correlation. At a typical frequency, we show that the spin-2 ultralight dark matter can give rise to the deformation of the Hellings-Downs correlation curve induced by the stochastic gravitational wave background., Comment: 7 pages, 4 figures

We explore binary asteroid formation by spin-up and rotational disruption considering the NASA DART mission's encounter with the Didymos-Dimorphos binary, which was the first small binary visited by a spacecraft. Using a suite of $N$-body simulations, we follow the gravitational accumulation of a satellite from meter-sized particles following a mass-shedding event from a rapidly rotating primary. The satellite's formation is chaotic, as it undergoes a series of collisions, mergers, and close gravitational encounters with other moonlets, leading to a wide range of outcomes in terms of the satellite's mass, shape, orbit, and rotation state. We find that a Dimorphos-like satellite can form rapidly, in a matter of days, following a realistic mass-shedding event in which only ${\sim}2-3\%$ of the primary's mass is shed. Satellites can form in synchronous rotation due to their formation near the Roche limit. There is a strong preference for forming prolate (elongated) satellites, although some simulations result in oblate spheroids like Dimorphos. The distribution of simulated secondary shapes is broadly consistent with other binary systems, measured through radar or lightcurves. Unless Dimorphos's shape is an outlier, and considering the observational bias against lightcurve-based determination of secondary elongations for oblate bodies, we suggest there could be a significant population of oblate secondaries. If these satellites initially form with elongated shapes, a yet-unidentified pathway is needed to explain how they become oblate. Finally, we show that this chaotic formation pathway occasionally forms asteroid pairs and stable triples, including co-orbital satellites and satellites in mean motion resonances., Comment: 41 pages, 23 figures, accepted to PSJ, movies available at https://zenodo.org/doi/10.5281/zenodo.8387043

Graph Neural Networks (GNNs) are widely applied across various domains, yet they perform poorly in deep layers. Existing research typically attributes this problem to node over-smoothing, where node representations become indistinguishable after multiple rounds of propagation. In this paper, we delve into the neighborhood propagation mechanism of GNNs and discover that the real root cause of GNNs' performance degradation in deep layers lies in ineffective neighborhood feature propagation. This propagation leads to an exponential growth of a node's current representation at every propagation step, making it extremely challenging to capture valuable dependencies between long-distance nodes. To address this issue, we introduce Graph Elimination Networks (GENs), which employ a specific algorithm to eliminate redundancies during neighborhood propagation. We demonstrate that GENs can enhance nodes' perception of distant neighborhoods and extend the depth of network propagation. Extensive experiments show that GENs outperform the state-of-the-art methods on various graph-level and node-level datasets., Comment: Includes 8 pages of main text and 4 pages of appendices