We describe a simple and effective method (Spectral Attribute removaL; SAL) to remove private or guarded information from neural representations. Our method uses matrix decomposition to project the input representations into directions with reduced covariance with the guarded information rather than maximal covariance as factorization methods normally use. We begin with linear information removal and proceed to generalize our algorithm to the case of nonlinear information removal using kernels. Our experiments demonstrate that our algorithm retains better main task performance after removing the guarded information compared to previous work. In addition, our experiments demonstrate that we need a relatively small amount of guarded attribute data to remove information about these attributes, which lowers the exposure to sensitive data and is more suitable for low-resource scenarios.

Gold Doesn't Always Glitter: Spectral Removal of Linear and Nonlinear Guarded Attribute Information

Self-supervised learning (SSL) techniques have recently been integrated into the few-shot learning (FSL) framework and have shown promising results in improving the few-shot image classification performance. However, existing SSL approaches used in FSL typically seek the supervision signals from the global embedding of every single image. Therefore, during the episodic training of FSL, these methods cannot capture and fully utilize the local visual information in image samples and the data structure information of the whole episode, which are beneficial to FSL. To this end, we propose to augment the few-shot learning objective with a novel self-supervised Episodic Spatial Pretext Task (ESPT). Specifically, for each few-shot episode, we generate its corresponding transformed episode by applying a random geometric transformation to all the images in it. Based on these, our ESPT objective is defined as maximizing the local spatial relationship consistency between the original episode and the transformed one. With this definition, the ESPT-augmented FSL objective promotes learning more transferable feature representations that capture the local spatial features of different images and their inter-relational structural information in each input episode, thus enabling the model to generalize better to new categories with only a few samples. Extensive experiments indicate that our ESPT method achieves new state-of-the-art performance for few-shot image classification on three mainstay benchmark datasets. The source code will be available at: https://github.com/Whut-YiRong/ESPT (Pytorch) and https://gitee.com/Yi-Rong/ESPT (MindSpore).

ESPT: A Self-Supervised Episodic Spatial Pretext Task for Improving Few-Shot Learning

Recent studies show that task distribution plays a vital role in the meta-learner's performance. Conventional wisdom is that task diversity should improve the performance of meta-learning. In this work, we find evidence to the contrary; (i) our experiments draw into question the efficacy of our learned models: similar manifolds can be learned with a subset of the data (lower task diversity). This finding questions the advantage of providing more data to the model, and (ii) adding diversity to the task distribution (higher task diversity) sometimes hinders the model and does not lead to a significant improvement in performance as previously believed. To strengthen our findings, we provide both empirical and theoretical evidence.

The Effect of Diversity in Meta-Learning

This paper focuses on the prevalent stage interference and stage performance imbalance of incremental learning. To avoid obvious stage learning bottlenecks, we propose a new incremental learning framework, which leverages a series of stage-isolated classifiers to perform the learning task at each stage, without interference from others. To be concrete, to aggregate multiple stage classifiers as a uniform one impartially, we first introduce a temperature-controlled energy metric for indicating the confidence score levels of the stage classifiers. We then propose an anchor-based energy self-normalization strategy to ensure the stage classifiers work at the same energy level. Finally, we design a voting-based inference augmentation strategy for robust inference. The proposed method is rehearsal-free and can work for almost all incremental learning scenarios. We evaluate the proposed method on four large datasets. Extensive results demonstrate the superiority of the proposed method in setting up new state-of-the-art overall performance. Code is available at https://github.com/iamwangyabin/ESN.

Isolation and Impartial Aggregation: A Paradigm of Incremental Learning without Interference

Unsupervised foreground-background segmentation aims at extracting salient objects from cluttered backgrounds, where Generative Adversarial Network (GAN) approaches, especially layered GANs, show great promise. However, without human annotations, they are typically prone to produce foreground and background layers with non-negligible semantic and visual confusion, dubbed "information leakage", resulting in notable degeneration of the generated segmentation mask. To alleviate this issue, we propose a simple-yet-effective explicit layer independence modeling approach, termed Independent Layer Synthesis GAN (ILSGAN), pursuing independent foreground-background layer generation by encouraging their discrepancy. Specifically, it targets minimizing the mutual information between visible and invisible regions of the foreground and background to spur interlayer independence. Through in-depth theoretical and experimental analyses, we justify that explicit layer independence modeling is critical to suppressing information leakage and contributes to impressive segmentation performance gains. Also, our ILSGAN achieves strong state-of-the-art generation quality and segmentation performance on complex real-world data. 

ILSGAN: Independent Layer Synthesis for Unsupervised Foreground-Background Segmentation

Deep networks trained on the source domain show degraded performance when tested on unseen target domain data. To enhance the model's generalization ability, most existing domain generalization methods learn domain invariant features by suppressing domain sensitive features. Different from them, we propose a Domain Projection and Contrastive Learning (DPCL) approach for generalized semantic segmentation, which includes two modules: Self-supervised Source Domain Projection (SSDP) and Multi-level Contrastive Learning (MLCL). SSDP aims to reduce domain gap by projecting data to the source domain, while MLCL is a learning scheme to learn discriminative and generalizable features on the projected data. During test time, we first project the target data by SSDP to mitigate domain shift, then generate the segmentation results by the learned segmentation network based on MLCL. At test time, we can update the projected data for minimizing our proposed pixel-to-pixel contrastive loss to obtain better results. Extensive experiments for semantic segmentation demonstrate the favorable generalization capability of our method on benchmark datasets.

Generalized Semantic Segmentation by Self-Supervised Source Domain Projection and Multi-Level Contrastive Learning

Handling out-of-distribution samples is a long-lasting challenge for deep visual models. In particular, domain generalization (DG) is one of the most relevant tasks that aims to train a model with a generalization capability on novel domains. Most existing DG approaches share the same philosophy to minimize the discrepancy between domains by finding the domain-invariant representations. On the contrary, our proposed method called POEM acquires a strong DG capability by learning domain-invariant and domain-specific representations and polarizing them. Specifically, POEM co-trains category-classifying and domain-classifying embeddings while regularizing them to be orthogonal via minimizing the cosine-similarity between their features, i.e., the polarization of embeddings. The clear separation of embeddings suppresses domain-specific features in the domain-invariant embeddings. The concept of POEM shows a unique direction to enhance the domain robustness of representations that brings considerable and consistent performance gains when combined with existing DG methods. Extensive simulation results in popular DG benchmarks with the PACS, VLCS, OfficeHome, TerraInc, and DomainNet datasets show that POEM indeed facilitates the category-classifying embedding to be more domain-invariant.

POEM: Polarization of Embeddings for Domain-Invariant Representations

Generative Adversarial Networks (GANs) have shown compelling results in various tasks and applications in recent years. However, mode collapse remains a critical problem in GANs. In this paper, we propose a novel training pipeline to address the mode collapse issue of GANs. Different from existing methods, we propose to generalize the discriminator as feature embedding and maximize the entropy of distributions in the embedding space learned by the discriminator. Specifically, two regularization terms, i.e., Deep Local Linear Embedding (DLLE) and Deep Isometric feature Mapping (DIsoMap), are designed to encourage the discriminator to learn the structural information embedded in the data, such
that the embedding space learned by the discriminator can be well-formed. Based on the well-learned embedding space supported by the discriminator, a non-parametric entropy estimator is designed to efficiently maximize the entropy of embedding vectors, playing as an approximation of maximizing the entropy of the generated distribution. By improving the discriminator and maximizing the distance of the most similar samples in the embedding space, our pipeline effectively reduces the mode collapse without sacrificing the quality of generated samples. Extensive experimental results show the effectiveness of our method which outperforms the GAN baseline, MaF-GAN on CelebA (9.13 vs. 12.43 in FID) and surpasses the recent state-of-the-art energy-based model on the ANIMEFACE dataset (2.80 vs. 2.26 in Inception score).

Combating Mode Collapse via Offline Manifold Entropy Estimation

In this work, we address the problem of scene-aware 3D human avatar generation based on human-scene interactions. In particular, we pay attention to the fact that physical contact between a 3D human and a scene (i.e., physical human-scene interactions) requires a geometrical alignment to generate natural 3D human avatar. Motivated by this fact, we present a new 3D human generation framework that considers geometric alignment on potential contact areas between 3D human avatars and their surroundings. In addition, we introduce a compact yet effective human pose classifier that classifies the human pose and provides potential contact areas of the 3D human avatar. It allows us to adaptively use geometric alignment loss according to the classified human pose. Compared to state-of-the-art method, our method can generate physically and semantically plausible 3D humans that interact naturally with 3D scenes without additional post-processing. In our evaluations, we achieve the improvements with more plausible interactions and more variety of poses than prior research in qualitative and quantitative analysis. Project page: https://bupyeonghealer.github.io/phin/.

Pose-guided 3D Human Generation in Indoor Scene

Hiding information in text documents has been a hot topic recently, with the most typical schemes of utilizing fonts. By constructing several fonts with similar appearances, information can be effectively represented and embedded in documents. However, due to the unstructured characteristic, font vectors are more difficult to synthesize than font images. Existing methods mainly use handcrafted features to design the fonts manually, which is time-consuming and labor-intensive. Moreover, due to the diversity of fonts, handcrafted features are not generalizable to different fonts. Besides, in practice, since documents might be distorted through transmission, ensuring extractability under distortions is also an important requirement. Therefore, three requirements are imposed on vector font generation in this domain: automaticity, generalizability, and robustness. However, none of the existing methods can satisfy these requirements well and simultaneously.
To satisfy the above requirements, we propose AutoStegaFont, an automatic vector font synthesis scheme for hiding information in documents. Specifically, we design a two-stage and dual-modality learning framework. In the first stage, we jointly train an encoder and a decoder to invisibly encode the font images with different information. To ensure robustness, we target designing a noise layer to work with the encoder and decoder during training. In the second stage, we employ a differentiable rasterizer to establish a connection between the image and the vector modality. Then, we design an optimization algorithm to convey the information from the encoded image to the corresponding vector. Thus the encoded font vectors can be automatically generated. Extensive experiments demonstrate the superior performance of our scheme in automatically synthesizing vector fonts for hiding information in documents, with robustness to distortions caused by low-resolution screenshots, printing, and photography. Besides, the proposed framework has better generalizability to fonts with diverse styles and languages.

AutoStegaFont: Synthesizing Vector Fonts for Hiding Information in Documents

Recent years we have witnessed rapid development in NeRF-based image rendering due to its high quality. However, point clouds rendering is somehow less explored. Compared to NeRF-based rendering which suffers from dense spatial sampling, point clouds rendering is naturally less computation intensive, which enables its deployment in mobile computing device. In this work, we focus on boosting the image quality of point clouds rendering with a compact model design. We first analyze the adaption of the volume rendering formulation on point clouds. Based on the analysis, we simplify the NeRF representation to a spatial mapping function which only requires single evaluation per pixel. Further, motivated by ray marching, we rectify the the noisy raw point clouds to the estimated intersection between rays and surfaces as queried coordinates, which could avoid \textit{spatial frequency collapse} and neighbor point disturbance. Composed of rasterization, spatial mapping and the refinement stages, our method achieves the state-of-the-art performance on point clouds rendering, outperforming prior works by notable margins, with a smaller model size. We obtain a PSNR of 31.74 on NeRF-Synthetic, 25.88 on ScanNet and 30.81 on DTU. Code and data are publicly available\footnote{\url{https://github.com/seanywang0408/RadianceMapping}}.

Boosting Point Clouds Rendering via Radiance Mapping

This paper explores a better prediction target for  BERT pre-training of vision transformers. We observe that current prediction targets disagree with human perception judgment. This contradiction motivates us to learn a perceptual prediction target. We argue that perceptually similar images should stay close to each other in the prediction target space. We surprisingly find one simple yet effective idea: enforcing perceptual similarity during the dVAE training. 
Moreover, we adopt a self-supervised transformer model for deep feature extraction and show that it works well for calculating perceptual similarity. We demonstrate that such learned visual tokens indeed exhibit better semantic meanings, and help pre-training achieve superior transfer performance in various downstream tasks. For example, we achieve $\textbf{84.5\%}$ Top-1 accuracy on ImageNet-1K with ViT-B backbone, outperforming the competitive method BEiT by +1.3% under the same pre-training epochs. Our approach also gets significant improvement on object detection and segmentation on COCO and semantic segmentation on ADE20K. Equipped with a larger backbone ViT-H, we achieve the state-of-the-art ImageNet accuracy (88.3%) among methods using only ImageNet-1K data.

Underline digital video library

Gold Doesn't Always Glitter: Spectral Removal of Linear and Nonlinear Guarded Attribute Information

ESPT: A Self-Supervised Episodic Spatial Pretext Task for Improving Few-Shot Learning

The Effect of Diversity in Meta-Learning

Isolation and Impartial Aggregation: A Paradigm of Incremental Learning without Interference

ILSGAN: Independent Layer Synthesis for Unsupervised Foreground-Background Segmentation

Generalized Semantic Segmentation by Self-Supervised Source Domain Projection and Multi-Level Contrastive Learning

POEM: Polarization of Embeddings for Domain-Invariant Representations

Combating Mode Collapse via Offline Manifold Entropy Estimation

Pose-guided 3D Human Generation in Indoor Scene

AutoStegaFont: Synthesizing Vector Fonts for Hiding Information in Documents

Boosting Point Clouds Rendering via Radiance Mapping

PeCo: Perceptual Codebook for BERT Pre-training of Vision Transformers

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES