Yu Zhang

Hao Chen, Chenyuan Qu, Yu Zhang et al.

With its significant performance improvements, the deep learning paradigm has become a standard tool for modern image denoisers. While promising performance has been shown on seen noise distributions, existing approaches often suffer from generalisation to unseen noise types or general and real noise. It is understandable as the model is designed to learn paired mapping (e.g. from a noisy image to its clean version). In this paper, we instead propose to learn to disentangle the noisy image, under the intuitive assumption that different corrupted versions of the same clean image share a common latent space. A self-supervised learning framework is proposed to achieve the goal, without looking at the latent clean image. By taking two different corrupted versions of the same image as input, the proposed Multi-view Self-supervised Disentanglement (MeD) approach learns to disentangle the latent clean features from the corruptions and recover the clean image consequently. Extensive experimental analysis on both synthetic and real noise shows the superiority of the proposed method over prior self-supervised approaches, especially on unseen novel noise types. On real noise, the proposed method even outperforms its supervised counterparts by over 3 dB.

Hong Zhang, Yu Zhang

Spiking neural networks (SNNs) are potential competitors to artificial neural networks (ANNs) due to their high energy-efficiency on neuromorphic hardware. However, SNNs are unfolded over simulation time steps during the training process. Thus, SNNs require much more memory than ANNs, which impedes the training of deeper SNN models. In this paper, we propose the reversible spiking neural network to reduce the memory cost of intermediate activations and membrane potentials during training. Firstly, we extend the reversible architecture along temporal dimension and propose the reversible spiking block, which can reconstruct the computational graph and recompute all intermediate variables in forward pass with a reverse process. On this basis, we adopt the state-of-the-art SNN models to the reversible variants, namely reversible spiking ResNet (RevSResNet) and reversible spiking transformer (RevSFormer). Through experiments on static and neuromorphic datasets, we demonstrate that the memory cost per image of our reversible SNNs does not increase with the network depth. On CIFAR10 and CIFAR100 datasets, our RevSResNet37 and RevSFormer-4-384 achieve comparable accuracies and consume 3.79x and 3.00x lower GPU memory per image than their counterparts with roughly identical model complexity and parameters. We believe that this work can unleash the memory constraints in SNN training and pave the way for training extremely large and deep SNNs. The code is available at https://github.com/mi804/RevSNN.git.

Lei Lan, Tianjia Shao, Zixuan Lu et al.

3D Gaussian Splatting (3DGS) has emerged as a mainstream solution for novel view synthesis and 3D reconstruction. By explicitly encoding a 3D scene using a collection of Gaussian kernels, 3DGS achieves high-quality rendering with superior efficiency. As a learning-based approach, 3DGS training has been dealt with the standard stochastic gradient descent (SGD) method, which offers at most linear convergence. Consequently, training often requires tens of minutes, even with GPU acceleration. This paper introduces a (near) second-order convergent training algorithm for 3DGS, leveraging its unique properties. Our approach is inspired by two key observations. First, the attributes of a Gaussian kernel contribute independently to the image-space loss, which endorses isolated and local optimization algorithms. We exploit this by splitting the optimization at the level of individual kernel attributes, analytically constructing small-size Newton systems for each parameter group, and efficiently solving these systems on GPU threads. This achieves Newton-like convergence per training image without relying on the global Hessian. Second, kernels exhibit sparse and structured coupling across input images. This property allows us to effectively utilize spatial information to mitigate overshoot during stochastic training. Our method converges an order faster than standard GPU-based 3DGS training, requiring over $10\times$ fewer iterations while maintaining or surpassing the quality of the compared with the SGD-based 3DGS reconstructions.

Yizhen Lao, Yu Zhang, Ziting Wang et al.

Deep learning models have been shown to be susceptible to adversarial attacks with visually imperceptible perturbations. Even this poses a serious security challenge for the localization of self-driving cars, there has been very little exploration of attack on it, as most of adversarial attacks have been applied to 3D perception. In this work, we propose a novel adversarial attack framework called DisorientLiDAR targeting LiDAR-based localization. By reverse-engineering localization models (e.g., feature extraction networks), adversaries can identify critical keypoints and strategically remove them, thereby disrupting LiDAR-based localization. Our proposal is first evaluated on three state-of-the-art point-cloud registration models (HRegNet, D3Feat, and GeoTransformer) using the KITTI dataset. Experimental results demonstrate that removing regions containing Top-K keypoints significantly degrades their registration accuracy. We further validate the attack's impact on the Autoware autonomous driving platform, where hiding merely a few critical regions induces noticeable localization drift. Finally, we extended our attacks to the physical world by hiding critical regions with near-infrared absorptive materials, thereby successfully replicate the attack effects observed in KITTI data. This step has been closer toward the realistic physical-world attack that demonstrate the veracity and generality of our proposal.

Jing Jiang, Yiran Ling, Binzhu Li et al.

Text-to-image generation models often struggle with key element loss or semantic confusion in tasks involving Chinese classical poetry.Addressing this issue through fine-tuning models needs considerable training costs. Additionally, manual prompts for re-diffusion adjustments need professional knowledge. To solve this problem, we propose Poetry2Image, an iterative correction framework for images generated from Chinese classical poetry. Utilizing an external poetry dataset, Poetry2Image establishes an automated feedback and correction loop, which enhances the alignment between poetry and image through image generation models and subsequent re-diffusion modifications suggested by large language models (LLM). Using a test set of 200 sentences of Chinese classical poetry, the proposed method--when integrated with five popular image generation models--achieves an average element completeness of 70.63%, representing an improvement of 25.56% over direct image generation. In tests of semantic correctness, our method attains an average semantic consistency of 80.09%. The study not only promotes the dissemination of ancient poetry culture but also offers a reference for similar non-fine-tuning methods to enhance LLM generation.

Guangneng Hu, Yu Zhang, Qiang Yang

Collaborative filtering (CF) is the key technique for recommender systems. Pure CF approaches exploit the user-item interaction data (e.g., clicks, likes, and views) only and suffer from the sparsity issue. Items are usually associated with content information such as unstructured text (e.g., abstracts of articles and reviews of products). CF can be extended to leverage text. In this paper, we develop a unified neural framework to exploit interaction data and content information seamlessly. The proposed framework, called LCMR, is based on memory networks and consists of local and centralized memories for exploiting content information and interaction data, respectively. By modeling content information as local memories, LCMR attentively learns what to exploit with the guidance of user-item interaction. On real-world datasets, LCMR shows better performance by comparing with various baselines in terms of the hit ratio and NDCG metrics. We further conduct analyses to understand how local and centralized memories work for the proposed framework.