PyTorchKR Papers

최신 논문

총 5,360편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2211.10908

ESTAS: Effective and Stable Trojan Attacks in Self-supervised Encoders with One Target Unlabelled Sample

Emerging self-supervised learning (SSL) has become a popular image representation encoding method to obviate the reliance on labeled data and learn rich representations from large-scale, ubiquitous unlabelled data. Then one can train a downstream classifier on top of the pre-trained SSL image encoder with few or no labeled downstream data. Although extensive works show that SSL has achieved remarkable and competitive performance on different downstream tasks, its security concerns, e.g, Trojan attacks in SSL encoders, are still not well-studied. In this work, we present a novel Trojan Attack method, denoted by ESTAS, that can enable an effective and stable attack in SSL encoders with only one target unlabeled sample. In particular, we propose consistent trigger poisoning and cascade optimization in ESTAS to improve attack efficacy and model accuracy, and eliminate the expensive target-class data sample extraction from large-scale disordered unlabelled data. Our substantial experiments on multiple datasets show that ESTAS stably achieves > 99% attacks success rate (ASR) with one target-class sample. Compared to prior works, ESTAS attains > 30% ASR increase and > 8.3% accuracy improvement on average.

arXiv:2211.10641공식 코드 · Papers with Code

Domain-Adaptive Self-Supervised Pre-Training for Face & Body Detection in Drawings

Drawings are powerful means of pictorial abstraction and communication. Understanding diverse forms of drawings, including digital arts, cartoons, and comics, has been a major problem of interest for the computer vision and computer graphics communities. Although there are large amounts of digitized drawings from comic books and cartoons, they contain vast stylistic variations, which necessitate expensive manual labeling for training domain-specific recognizers. In this work, we show how self-supervised learning, based on a teacher-student network with a modified student network update design, can be used to build face and body detectors. Our setup allows exploiting large amounts of unlabeled data from the target domain when labels are provided for only a small subset of it. We further demonstrate that style transfer can be incorporated into our learning pipeline to bootstrap detectors using a vast amount of out-of-domain labeled images from natural images (i.e., images from the real world). Our combined architecture yields detectors with state-of-the-art (SOTA) and near-SOTA performance using minimal annotation effort. Our code can be accessed from https://github.com/barisbatuhan/DASS_Detector.

arXiv:2211.10636공식 코드 · Papers with Code

EVEREST: Efficient Masked Video Autoencoder by Removing Redundant Spatiotemporal Tokens

Masked Video Autoencoder (MVA) approaches have demonstrated their potential by significantly outperforming previous video representation learning methods. However, they waste an excessive amount of computations and memory in predicting uninformative tokens/frames due to random masking strategies. (e.g., over 16 nodes with 128 NVIDIA A100 GPUs). To resolve this issue, we exploit the unequal information density among the patches in videos and propose EVEREST, a surprisingly efficient MVA approach for video representation learning that finds tokens containing rich motion features and discards uninformative ones during both pre-training and fine-tuning. We further present an information-intensive frame selection strategy that allows the model to focus on informative and causal frames with minimal redundancy. Our method significantly reduces the computation and memory requirements of MVA, enabling the pre-training and fine-tuning on a single machine with 8 GPUs while achieving comparable performance to computation- and memory-heavy baselines on multiple benchmarks and the uncurated Ego4D dataset. We hope that our work contributes to reducing the barrier to further research on video understanding.

arXiv:2211.11182

Deep Projective Rotation Estimation through Relative Supervision

Orientation estimation is the core to a variety of vision and robotics tasks such as camera and object pose estimation. Deep learning has offered a way to develop image-based orientation estimators; however, such estimators often require training on a large labeled dataset, which can be time-intensive to collect. In this work, we explore whether self-supervised learning from unlabeled data can be used to alleviate this issue. Specifically, we assume access to estimates of the relative orientation between neighboring poses, such that can be obtained via a local alignment method. While self-supervised learning has been used successfully for translational object keypoints, in this work, we show that naively applying relative supervision to the rotational group $SO(3)$ will often fail to converge due to the non-convexity of the rotational space. To tackle this challenge, we propose a new algorithm for self-supervised orientation estimation which utilizes Modified Rodrigues Parameters to stereographically project the closed manifold of $SO(3)$ to the open manifold of $\mathbb{R}^{3}$, allowing the optimization to be done in an open Euclidean space. We empirically validate the benefits of the proposed algorithm for rotational averaging problem in two settings: (1) direct optimization on rotation parameters, and (2) optimization of parameters of a convolutional neural network that predicts object orientations from images. In both settings, we demonstrate that our proposed algorithm is able to converge to a consistent relative orientation frame much faster than algorithms that purely operate in the $SO(3)$ space. Additional information can be found at https://sites.google.com/view/deep-projective-rotation/home .

arXiv:2211.10549

Local Contrastive Feature learning for Tabular Data

Contrastive self-supervised learning has been successfully used in many domains, such as images, texts, graphs, etc., to learn features without requiring label information. In this paper, we propose a new local contrastive feature learning (LoCL) framework, and our theme is to learn local patterns/features from tabular data. In order to create a niche for local learning, we use feature correlations to create a maximum-spanning tree, and break the tree into feature subsets, with strongly correlated features being assigned next to each other. Convolutional learning of the features is used to learn latent feature space, regulated by contrastive and reconstruction losses. Experiments on public tabular datasets show the effectiveness of the proposed method versus state-of-the-art baseline methods.

arXiv:2211.10946공식 코드 · Papers with Code

Normalizing Flows for Human Pose Anomaly Detection

Video anomaly detection is an ill-posed problem because it relies on many parameters such as appearance, pose, camera angle, background, and more. We distill the problem to anomaly detection of human pose, thus decreasing the risk of nuisance parameters such as appearance affecting the result. Focusing on pose alone also has the side benefit of reducing bias against distinct minority groups. Our model works directly on human pose graph sequences and is exceptionally lightweight (~1K parameters), capable of running on any machine able to run the pose estimation with negligible additional resources. We leverage the highly compact pose representation in a normalizing flows framework, which we extend to tackle the unique characteristics of spatio-temporal pose data and show its advantages in this use case. The algorithm is quite general and can handle training data of only normal examples as well as a supervised setting that consists of labeled normal and abnormal examples. We report state-of-the-art results on two anomaly detection benchmarks - the unsupervised ShanghaiTech dataset and the recent supervised UBnormal dataset.

arXiv:2211.09988

Exploring WavLM on Speech Enhancement

There is a surge in interest in self-supervised learning approaches for end-to-end speech encoding in recent years as they have achieved great success. Especially, WavLM showed state-of-the-art performance on various speech processing tasks. To better understand the efficacy of self-supervised learning models for speech enhancement, in this work, we design and conduct a series of experiments with three resource conditions by combining WavLM and two high-quality speech enhancement systems. Also, we propose a regression-based WavLM training objective and a noise-mixing data configuration to further boost the downstream enhancement performance. The experiments on the DNS challenge dataset and a simulation dataset show that the WavLM benefits the speech enhancement task in terms of both speech quality and speech recognition accuracy, especially for low fine-tuning resources. For the high fine-tuning resource condition, only the word error rate is substantially improved.

arXiv:2211.10408공식 코드 · Papers with Code

CroCo v2: Improved Cross-view Completion Pre-training for Stereo Matching and Optical Flow

Despite impressive performance for high-level downstream tasks, self-supervised pre-training methods have not yet fully delivered on dense geometric vision tasks such as stereo matching or optical flow. The application of self-supervised concepts, such as instance discrimination or masked image modeling, to geometric tasks is an active area of research. In this work, we build on the recent cross-view completion framework, a variation of masked image modeling that leverages a second view from the same scene which makes it well suited for binocular downstream tasks. The applicability of this concept has so far been limited in at least two ways: (a) by the difficulty of collecting real-world image pairs -- in practice only synthetic data have been used -- and (b) by the lack of generalization of vanilla transformers to dense downstream tasks for which relative position is more meaningful than absolute position. We explore three avenues of improvement. First, we introduce a method to collect suitable real-world image pairs at large scale. Second, we experiment with relative positional embeddings and show that they enable vision transformers to perform substantially better. Third, we scale up vision transformer based cross-completion architectures, which is made possible by the use of large amounts of data. With these improvements, we show for the first time that state-of-the-art results on stereo matching and optical flow can be reached without using any classical task-specific techniques like correlation volume, iterative estimation, image warping or multi-scale reasoning, thus paving the way towards universal vision models.

arXiv:2211.09949공식 코드 · Papers with Code

Is Smaller Always Faster? Tradeoffs in Compressing Self-Supervised Speech Transformers

Transformer-based self-supervised models have achieved remarkable success in speech processing, but their large size and high inference cost present significant challenges for real-world deployment. While numerous compression techniques have been proposed, inconsistent evaluation metrics make it difficult to compare their practical effectiveness. In this work, we conduct a comprehensive study of four common compression methods, including weight pruning, head pruning, low-rank approximation, and knowledge distillation on self-supervised speech Transformers. We evaluate each method under three key metrics: parameter count, multiply-accumulate operations, and real-time factor. Results show that each method offers distinct advantages. In addition, we contextualize recent compression techniques, comparing DistilHuBERT, FitHuBERT, LightHuBERT, ARMHuBERT, and STaRHuBERT under the same framework, offering practical guidance on compression for deployment.

arXiv:2211.09944공식 코드 · Papers with Code

MelHuBERT: A simplified HuBERT on Mel spectrograms

Self-supervised models have had great success in learning speech representations that can generalize to various downstream tasks. However, most self-supervised models require a large amount of compute and multiple GPUs to train, significantly hampering the development of self-supervised learning. In an attempt to reduce the computation of training, we revisit the training of HuBERT, a highly successful self-supervised model. We improve and simplify several key components, including the loss function, input representation, and training in multiple stages. Our model, MelHuBERT, is able to achieve favorable performance on phone recognition, speaker identification, and automatic speech recognition against HuBERT, while saving 31.2% of the pre-training time, or equivalently 33.5% MACs per one second speech. The code and pre-trained models are available in https://github.com/nervjack2/MelHuBERT.

arXiv:2211.09861

Self-Supervised Visual Representation Learning via Residual Momentum

Self-supervised learning (SSL) approaches have shown promising capabilities in learning the representation from unlabeled data. Amongst them, momentum-based frameworks have attracted significant attention. Despite being a great success, these momentum-based SSL frameworks suffer from a large gap in representation between the online encoder (student) and the momentum encoder (teacher), which hinders performance on downstream tasks. This paper is the first to investigate and identify this invisible gap as a bottleneck that has been overlooked in the existing SSL frameworks, potentially preventing the models from learning good representation. To solve this problem, we propose "residual momentum" to directly reduce this gap to encourage the student to learn the representation as close to that of the teacher as possible, narrow the performance gap with the teacher, and significantly improve the existing SSL. Our method is straightforward, easy to implement, and can be easily plugged into other SSL frameworks. Extensive experimental results on numerous benchmark datasets and diverse network architectures have demonstrated the effectiveness of our method over the state-of-the-art contrastive learning baselines.

arXiv:2211.09981

Weighted Ensemble Self-Supervised Learning

Ensembling has proven to be a powerful technique for boosting model performance, uncertainty estimation, and robustness in supervised learning. Advances in self-supervised learning (SSL) enable leveraging large unlabeled corpora for state-of-the-art few-shot and supervised learning performance. In this paper, we explore how ensemble methods can improve recent SSL techniques by developing a framework that permits data-dependent weighted cross-entropy losses. We refrain from ensembling the representation backbone; this choice yields an efficient ensemble method that incurs a small training cost and requires no architectural changes or computational overhead to downstream evaluation. The effectiveness of our method is demonstrated with two state-of-the-art SSL methods, DINO (Caron et al., 2021) and MSN (Assran et al., 2022). Our method outperforms both in multiple evaluation metrics on ImageNet-1K, particularly in the few-shot setting. We explore several weighting schemes and find that those which increase the diversity of ensemble heads lead to better downstream evaluation results. Thorough experiments yield improved prior art baselines which our method still surpasses; e.g., our overall improvement with MSN ViT-B/16 is 3.9 p.p. for 1-shot learning.

arXiv:2211.09376

Balanced Deep CCA for Bird Vocalization Detection

Event detection improves when events are captured by two different modalities rather than just one. But to train detection systems on multiple modalities is challenging, in particular when there is abundance of unlabelled data but limited amounts of labeled data. We develop a novel self-supervised learning technique for multi-modal data that learns (hidden) correlations between simultaneously recorded microphone (sound) signals and accelerometer (body vibration) signals. The key objective of this work is to learn useful embeddings associated with high performance in downstream event detection tasks when labeled data is scarce and the audio events of interest (songbird vocalizations) are sparse. We base our approach on deep canonical correlation analysis (DCCA) that suffers from event sparseness. We overcome the sparseness of positive labels by first learning a data sampling model from the labelled data and by applying DCCA on the output it produces. This method that we term balanced DCCA (b-DCCA) improves the performance of the unsupervised embeddings on the downstream supervised audio detection task compared to classsical DCCA. Because data labels are frequently imbalanced, our method might be of broad utility in low-resource scenarios.

arXiv:2211.09703공식 코드 · Papers with Code

EfficientTrain: Exploring Generalized Curriculum Learning for Training Visual Backbones

The superior performance of modern deep networks usually comes with a costly training procedure. This paper presents a new curriculum learning approach for the efficient training of visual backbones (e.g., vision Transformers). Our work is inspired by the inherent learning dynamics of deep networks: we experimentally show that at an earlier training stage, the model mainly learns to recognize some 'easier-to-learn' discriminative patterns within each example, e.g., the lower-frequency components of images and the original information before data augmentation. Driven by this phenomenon, we propose a curriculum where the model always leverages all the training data at each epoch, while the curriculum starts with only exposing the 'easier-to-learn' patterns of each example, and introduces gradually more difficult patterns. To implement this idea, we 1) introduce a cropping operation in the Fourier spectrum of the inputs, which enables the model to learn from only the lower-frequency components efficiently, 2) demonstrate that exposing the features of original images amounts to adopting weaker data augmentation, and 3) integrate 1) and 2) and design a curriculum learning schedule with a greedy-search algorithm. The resulting approach, EfficientTrain, is simple, general, yet surprisingly effective. As an off-the-shelf method, it reduces the wall-time training cost of a wide variety of popular models (e.g., ResNet, ConvNeXt, DeiT, PVT, Swin, and CSWin) by >1.5x on ImageNet-1K/22K without sacrificing accuracy. It is also effective for self-supervised learning (e.g., MAE). Code is available at https://github.com/LeapLabTHU/EfficientTrain.

arXiv:2211.09643

CPT-V: A Contrastive Approach to Post-Training Quantization of Vision Transformers

When considering post-training quantization, prior work has typically focused on developing a mixed precision scheme or learning the best way to partition a network for quantization. In our work, CPT-V, we look at a general way to improve the accuracy of networks that have already been quantized, simply by perturbing the quantization scales. Borrowing the idea of contrastive loss from self-supervised learning, we find a robust way to jointly minimize a loss function using just 1,000 calibration images. In order to determine the best performing quantization scale, CPT-V contrasts the features of quantized and full precision models in a self-supervised fashion. Unlike traditional reconstruction-based loss functions, the use of a contrastive loss function not only rewards similarity between the quantized and full precision outputs but also helps in distinguishing the quantized output from other outputs within a given batch. In addition, in contrast to prior works, CPT-V proposes a block-wise evolutionary search to minimize a global contrastive loss objective, allowing for accuracy improvement of existing vision transformer (ViT) quantization schemes. For example, CPT-V improves the top-1 accuracy of a fully quantized ViT-Base by 10.30%, 0.78%, and 0.15% for 3-bit, 4-bit, and 8-bit weight quantization levels. Extensive experiments on a variety of other ViT architectures further demonstrate its robustness in extreme quantization scenarios. Our code is available at <link>.

arXiv:2211.08989

Avoid Overthinking in Self-Supervised Models for Speech Recognition

Self-supervised learning (SSL) models reshaped our approach to speech, language and vision. However their huge size and the opaque relations between their layers and tasks result in slow inference and network overthinking, where predictions made from the last layer of large models is worse than those made from intermediate layers. Early exit (EE) strategies can solve both issues by dynamically reducing computations at inference time for certain samples. Although popular for classification tasks in vision and language, EE has seen less use for sequence-to-sequence speech recognition (ASR) tasks where outputs from early layers are often degenerate. This challenge is further compounded when speech SSL models are applied on out-of-distribution (OOD) data. This paper first shows that SSL models do overthinking in ASR. We then motivate further research in EE by computing an optimal bound for performance versus speed trade-offs. To approach this bound we propose two new strategies for ASR: (1) we adapt the recently proposed patience strategy to ASR; and (2) we design a new EE strategy specific to ASR that performs better than all strategies previously introduced.

arXiv:2211.09027

LLEDA -- Lifelong Self-Supervised Domain Adaptation

Humans and animals have the ability to continuously learn new information over their lifetime without losing previously acquired knowledge. However, artificial neural networks struggle with this due to new information conflicting with old knowledge, resulting in catastrophic forgetting. The complementary learning systems (CLS) theory suggests that the interplay between hippocampus and neocortex systems enables long-term and efficient learning in the mammalian brain, with memory replay facilitating the interaction between these two systems to reduce forgetting. The proposed Lifelong Self-Supervised Domain Adaptation (LLEDA) framework draws inspiration from the CLS theory and mimics the interaction between two networks: a DA network inspired by the hippocampus that quickly adjusts to changes in data distribution and an SSL network inspired by the neocortex that gradually learns domain-agnostic general representations. LLEDA's latent replay technique facilitates communication between these two networks by reactivating and replaying the past memory latent representations to stabilise long-term generalisation and retention without interfering with the previously learned information. Extensive experiments demonstrate that the proposed method outperforms several other methods resulting in a long-term adaptation while being less prone to catastrophic forgetting when transferred to new domains.

arXiv:2211.09013

Masked Reconstruction Contrastive Learning with Information Bottleneck Principle

Contrastive learning (CL) has shown great power in self-supervised learning due to its ability to capture insight correlations among large-scale data. Current CL models are biased to learn only the ability to discriminate positive and negative pairs due to the discriminative task setting. However, this bias would lead to ignoring its sufficiency for other downstream tasks, which we call the discriminative information overfitting problem. In this paper, we propose to tackle the above problems from the aspect of the Information Bottleneck (IB) principle, further pushing forward the frontier of CL. Specifically, we present a new perspective that CL is an instantiation of the IB principle, including information compression and expression. We theoretically analyze the optimal information situation and demonstrate that minimum sufficient augmentation and information-generalized representation are the optimal requirements for achieving maximum compression and generalizability to downstream tasks. Therefore, we propose the Masked Reconstruction Contrastive Learning~(MRCL) model to improve CL models. For implementation in practice, MRCL utilizes the masking operation for stronger augmentation, further eliminating redundant and noisy information. In order to alleviate the discriminative information overfitting problem effectively, we employ the reconstruction task to regularize the discriminative task. We conduct comprehensive experiments and show the superiority of the proposed model on multiple tasks, including image classification, semantic segmentation and objective detection.

arXiv:2211.08559

Robust Alzheimer's Progression Modeling using Cross-Domain Self-Supervised Deep Learning

Developing successful artificial intelligence systems in practice depends on both robust deep learning models and large, high-quality data. However, acquiring and labeling data can be prohibitively expensive and time-consuming in many real-world applications, such as clinical disease models. Self-supervised learning has demonstrated great potential in increasing model accuracy and robustness in small data regimes. In addition, many clinical imaging and disease modeling applications rely heavily on regression of continuous quantities. However, the applicability of self-supervised learning for these medical-imaging regression tasks has not been extensively studied. In this study, we develop a cross-domain self-supervised learning approach for disease prognostic modeling as a regression problem using medical images as input. We demonstrate that self-supervised pretraining can improve the prediction of Alzheimer's Disease progression from brain MRI. We also show that pretraining on extended (but not labeled) brain MRI data outperforms pretraining on natural images. We further observe that the highest performance is achieved when both natural images and extended brain-MRI data are used for pretraining.

arXiv:2211.08904

SelfOdom: Self-supervised Egomotion and Depth Learning via Bi-directional Coarse-to-Fine Scale Recovery

Accurately perceiving location and scene is crucial for autonomous driving and mobile robots. Recent advances in deep learning have made it possible to learn egomotion and depth from monocular images in a self-supervised manner, without requiring highly precise labels to train the networks. However, monocular vision methods suffer from a limitation known as scale-ambiguity, which restricts their application when absolute-scale is necessary. To address this, we propose SelfOdom, a self-supervised dual-network framework that can robustly and consistently learn and generate pose and depth estimates in global scale from monocular images. In particular, we introduce a novel coarse-to-fine training strategy that enables the metric scale to be recovered in a two-stage process. Furthermore, SelfOdom is flexible and can incorporate inertial data with images, which improves its robustness in challenging scenarios, using an attention-based fusion module. Our model excels in both normal and challenging lighting conditions, including difficult night scenes. Extensive experiments on public datasets have demonstrated that SelfOdom outperforms representative traditional and learning-based VO and VIO models.