PyTorchKR Papers

“Image Classification”

총 5,746편
태그·기간 등 상세 필터

카테고리 탐색

arXiv:2103.10626공식 코드 · Papers with Code

Cluster-to-Conquer: A Framework for End-to-End Multi-Instance Learning for Whole Slide Image Classification

In recent years, the availability of digitized Whole Slide Images (WSIs) has enabled the use of deep learning-based computer vision techniques for automated disease diagnosis. However, WSIs present unique computational and algorithmic challenges. WSIs are gigapixel-sized ($\sim$100K pixels), making them infeasible to be used directly for training deep neural networks. Also, often only slide-level labels are available for training as detailed annotations are tedious and can be time-consuming for experts. Approaches using multiple-instance learning (MIL) frameworks have been shown to overcome these challenges. Current state-of-the-art approaches divide the learning framework into two decoupled parts: a convolutional neural network (CNN) for encoding the patches followed by an independent aggregation approach for slide-level prediction. In this approach, the aggregation step has no bearing on the representations learned by the CNN encoder. We have proposed an end-to-end framework that clusters the patches from a WSI into ${k}$-groups, samples ${k}'$ patches from each group for training, and uses an adaptive attention mechanism for slide level prediction; Cluster-to-Conquer (C2C). We have demonstrated that dividing a WSI into clusters can improve the model training by exposing it to diverse discriminative features extracted from the patches. We regularized the clustering mechanism by introducing a KL-divergence loss between the attention weights of patches in a cluster and the uniform distribution. The framework is optimized end-to-end on slide-level cross-entropy, patch-level cross-entropy, and KL-divergence loss (Implementation: https://github.com/YashSharma/C2C).

arXiv:2303.05517공식 코드 · Papers with Code

On the Soundness of XAI in Prognostics and Health Management (PHM)

The aim of Predictive Maintenance, within the field of Prognostics and Health Management (PHM), is to identify and anticipate potential issues in the equipment before these become critical. The main challenge to be addressed is to assess the amount of time a piece of equipment will function effectively before it fails, which is known as Remaining Useful Life (RUL). Deep Learning (DL) models, such as Deep Convolutional Neural Networks (DCNN) and Long Short-Term Memory (LSTM) networks, have been widely adopted to address the task, with great success. However, it is well known that this kind of black box models are opaque decision systems, and it may be hard to explain its outputs to stakeholders (experts in the industrial equipment). Due to the large number of parameters that determine the behavior of these complex models, understanding the reasoning behind the predictions is challenging. This work presents a critical and comparative revision on a number of XAI methods applied on time series regression model for PM. The aim is to explore XAI methods within time series regression, which have been less studied than those for time series classification. The model used during the experimentation is a DCNN trained to predict the RUL of an aircraft engine. The methods are reviewed and compared using a set of metrics that quantifies a number of desirable properties that any XAI method should fulfill. The results show that GRAD-CAM is the most robust method, and that the best layer is not the bottom one, as is commonly seen within the context of Image Processing.

arXiv:2309.14203공식 코드 · Papers with Code

Detecting and Grounding Multi-Modal Media Manipulation and Beyond

Misinformation has become a pressing issue. Fake media, in both visual and textual forms, is widespread on the web. While various deepfake detection and text fake news detection methods have been proposed, they are only designed for single-modality forgery based on binary classification, let alone analyzing and reasoning subtle forgery traces across different modalities. In this paper, we highlight a new research problem for multi-modal fake media, namely Detecting and Grounding Multi-Modal Media Manipulation (DGM^4). DGM^4 aims to not only detect the authenticity of multi-modal media, but also ground the manipulated content, which requires deeper reasoning of multi-modal media manipulation. To support a large-scale investigation, we construct the first DGM^4 dataset, where image-text pairs are manipulated by various approaches, with rich annotation of diverse manipulations. Moreover, we propose a novel HierArchical Multi-modal Manipulation rEasoning tRansformer (HAMMER) to fully capture the fine-grained interaction between different modalities. HAMMER performs 1) manipulation-aware contrastive learning between two uni-modal encoders as shallow manipulation reasoning, and 2) modality-aware cross-attention by multi-modal aggregator as deep manipulation reasoning. Dedicated manipulation detection and grounding heads are integrated from shallow to deep levels based on the interacted multi-modal information. To exploit more fine-grained contrastive learning for cross-modal semantic alignment, we further integrate Manipulation-Aware Contrastive Loss with Local View and construct a more advanced model HAMMER++. Finally, we build an extensive benchmark and set up rigorous evaluation metrics for this new research problem. Comprehensive experiments demonstrate the superiority of HAMMER and HAMMER++.

arXiv:2401.02020공식 코드 · Papers with Code

Spikformer V2: Join the High Accuracy Club on ImageNet with an SNN Ticket

Spiking Neural Networks (SNNs), known for their biologically plausible architecture, face the challenge of limited performance. The self-attention mechanism, which is the cornerstone of the high-performance Transformer and also a biologically inspired structure, is absent in existing SNNs. To this end, we explore the potential of leveraging both self-attention capability and biological properties of SNNs, and propose a novel Spiking Self-Attention (SSA) and Spiking Transformer (Spikformer). The SSA mechanism eliminates the need for softmax and captures the sparse visual feature employing spike-based Query, Key, and Value. This sparse computation without multiplication makes SSA efficient and energy-saving. Further, we develop a Spiking Convolutional Stem (SCS) with supplementary convolutional layers to enhance the architecture of Spikformer. The Spikformer enhanced with the SCS is referred to as Spikformer V2. To train larger and deeper Spikformer V2, we introduce a pioneering exploration of Self-Supervised Learning (SSL) within the SNN. Specifically, we pre-train Spikformer V2 with masking and reconstruction style inspired by the mainstream self-supervised Transformer, and then finetune the Spikformer V2 on the image classification on ImageNet. Extensive experiments show that Spikformer V2 outperforms other previous surrogate training and ANN2SNN methods. An 8-layer Spikformer V2 achieves an accuracy of 80.38% using 4 time steps, and after SSL, a 172M 16-layer Spikformer V2 reaches an accuracy of 81.10% with just 1 time step. To the best of our knowledge, this is the first time that the SNN achieves 80+% accuracy on ImageNet. The code will be available at Spikformer V2.

arXiv:2408.03703공식 코드 · Papers with Code

CAS-ViT: Convolutional Additive Self-attention Vision Transformers for Efficient Mobile Applications

Vision Transformers (ViTs) mark a revolutionary advance in neural networks with their token mixer's powerful global context capability. However, the pairwise token affinity and complex matrix operations limit its deployment on resource-constrained scenarios and real-time applications, such as mobile devices, although considerable efforts have been made in previous works. In this paper, we introduce CAS-ViT: Convolutional Additive Self-attention Vision Transformers, to achieve a balance between efficiency and performance in mobile applications. Firstly, we argue that the capability of token mixers to obtain global contextual information hinges on multiple information interactions, such as spatial and channel domains. Subsequently, we propose Convolutional Additive Token Mixer (CATM) employing underlying spatial and channel attention as novel interaction forms. This module eliminates troublesome complex operations such as matrix multiplication and Softmax. We introduce Convolutional Additive Self-attention(CAS) block hybrid architecture and utilize CATM for each block. And further, we build a family of lightweight networks, which can be easily extended to various downstream tasks. Finally, we evaluate CAS-ViT across a variety of vision tasks, including image classification, object detection, instance segmentation, and semantic segmentation. Our M and T model achieves 83.0\%/84.1\% top-1 with only 12M/21M parameters on ImageNet-1K. Meanwhile, throughput evaluations on GPUs, ONNX, and iPhones also demonstrate superior results compared to other state-of-the-art backbones. Extensive experiments demonstrate that our approach achieves a better balance of performance, efficient inference and easy-to-deploy. Our code and model are available at: \url{https://github.com/Tianfang-Zhang/CAS-ViT}

arXiv:2501.09705공식 코드 · Papers with Code

Practical Continual Forgetting for Pre-trained Vision Models

For privacy and security concerns, the need to erase unwanted information from pre-trained vision models is becoming evident nowadays. In real-world scenarios, erasure requests originate at any time from both users and model owners, and these requests usually form a sequence. Therefore, under such a setting, selective information is expected to be continuously removed from a pre-trained model while maintaining the rest. We define this problem as continual forgetting and identify three key challenges. (i) For unwanted knowledge, efficient and effective deleting is crucial. (ii) For remaining knowledge, the impact brought by the forgetting procedure should be minimal. (iii) In real-world scenarios, the training samples may be scarce or partially missing during the process of forgetting. To address them, we first propose Group Sparse LoRA (GS-LoRA). Specifically, towards (i), we introduce LoRA modules to fine-tune the FFN layers in Transformer blocks for each forgetting task independently, and towards (ii), a simple group sparse regularization is adopted, enabling automatic selection of specific LoRA groups and zeroing out the others. To further extend GS-LoRA to more practical scenarios, we incorporate prototype information as additional supervision and introduce a more practical approach, GS-LoRA++. For each forgotten class, we move the logits away from its original prototype. For the remaining classes, we pull the logits closer to their respective prototypes. We conduct extensive experiments on face recognition, object detection and image classification and demonstrate that our method manages to forget specific classes with minimal impact on other classes. Codes have been released on https://github.com/bjzhb666/GS-LoRA.

arXiv:2312.05632공식 코드 · Papers with Code

Subject-Based Domain Adaptation for Facial Expression Recognition

Adapting a deep learning model to a specific target individual is a challenging facial expression recognition (FER) task that may be achieved using unsupervised domain adaptation (UDA) methods. Although several UDA methods have been proposed to adapt deep FER models across source and target data sets, multiple subject-specific source domains are needed to accurately represent the intra- and inter-person variability in subject-based adaption. This paper considers the setting where domains correspond to individuals, not entire datasets. Unlike UDA, multi-source domain adaptation (MSDA) methods can leverage multiple source datasets to improve the accuracy and robustness of the target model. However, previous methods for MSDA adapt image classification models across datasets and do not scale well to a more significant number of source domains. This paper introduces a new MSDA method for subject-based domain adaptation in FER. It efficiently leverages information from multiple source subjects (labeled source domain data) to adapt a deep FER model to a single target individual (unlabeled target domain data). During adaptation, our subject-based MSDA first computes a between-source discrepancy loss to mitigate the domain shift among data from several source subjects. Then, a new strategy is employed to generate augmented confident pseudo-labels for the target subject, allowing a reduction in the domain shift between source and target subjects. Experiments performed on the challenging BioVid heat and pain dataset with 87 subjects and the UNBC-McMaster shoulder pain dataset with 25 subjects show that our subject-based MSDA can outperform state-of-the-art methods yet scale well to multiple subject-based source domains.

arXiv:2410.06645공식 코드 · Papers with Code

Continual Learning in the Frequency Domain

Continual learning (CL) is designed to learn new tasks while preserving existing knowledge. Replaying samples from earlier tasks has proven to be an effective method to mitigate the forgetting of previously acquired knowledge. However, the current research on the training efficiency of rehearsal-based methods is insufficient, which limits the practical application of CL systems in resource-limited scenarios. The human visual system (HVS) exhibits varying sensitivities to different frequency components, enabling the efficient elimination of visually redundant information. Inspired by HVS, we propose a novel framework called Continual Learning in the Frequency Domain (CLFD). To our knowledge, this is the first study to utilize frequency domain features to enhance the performance and efficiency of CL training on edge devices. For the input features of the feature extractor, CLFD employs wavelet transform to map the original input image into the frequency domain, thereby effectively reducing the size of input feature maps. Regarding the output features of the feature extractor, CLFD selectively utilizes output features for distinct classes for classification, thereby balancing the reusability and interference of output features based on the frequency domain similarity of the classes across various tasks. Optimizing only the input and output features of the feature extractor allows for seamless integration of CLFD with various rehearsal-based methods. Extensive experiments conducted in both cloud and edge environments demonstrate that CLFD consistently improves the performance of state-of-the-art (SOTA) methods in both precision and training efficiency. Specifically, CLFD can increase the accuracy of the SOTA CL method by up to 6.83% and reduce the training time by 2.6$\times$.

arXiv:2502.07542공식 코드 · Papers with Code

Exoplanet Transit Candidate Identification in TESS Full-Frame Images via a Transformer-Based Algorithm

The Transiting Exoplanet Survey Satellite (TESS) is surveying a large fraction of the sky, generating a vast database of photometric time series data that requires thorough analysis to identify exoplanetary transit signals. Automated learning approaches have been successfully applied to identify transit signals. However, most existing methods focus on the classification and validation of candidates, while few efforts have explored new techniques for the search of candidates. To search for new exoplanet transit candidates, we propose an approach to identify exoplanet transit signals without the need for phase folding or assuming periodicity in the transit signals, such as those observed in multi-transit light curves. To achieve this, we implement a new neural network inspired by Transformers to directly process Full Frame Image (FFI) light curves to detect exoplanet transits. Transformers, originally developed for natural language processing, have recently demonstrated significant success in capturing long-range dependencies compared to previous approaches focused on sequential data. This ability allows us to employ multi-head self-attention to identify exoplanet transit signals directly from the complete light curves, combined with background and centroid time series, without requiring prior transit parameters. The network is trained to learn characteristics of the transit signal, like the dip shape, which helps distinguish planetary transits from other variability sources. Our model successfully identified 214 new planetary system candidates, including 122 multi-transit light curves, 88 single-transit and 4 multi-planet systems from TESS sectors 1-26 with a radius > 0.27 $R_{\mathrm{Jupiter}}$, demonstrating its ability to detect transits regardless of their periodicity.

arXiv:2110.10389공식 코드 · Papers with Code

Does Data Repair Lead to Fair Models? Curating Contextually Fair Data To Reduce Model Bias

Contextual information is a valuable cue for Deep Neural Networks (DNNs) to learn better representations and improve accuracy. However, co-occurrence bias in the training dataset may hamper a DNN model's generalizability to unseen scenarios in the real world. For example, in COCO, many object categories have a much higher co-occurrence with men compared to women, which can bias a DNN's prediction in favor of men. Recent works have focused on task-specific training strategies to handle bias in such scenarios, but fixing the available data is often ignored. In this paper, we propose a novel and more generic solution to address the contextual bias in the datasets by selecting a subset of the samples, which is fair in terms of the co-occurrence with various classes for a protected attribute. We introduce a data repair algorithm using the coefficient of variation, which can curate fair and contextually balanced data for a protected class(es). This helps in training a fair model irrespective of the task, architecture or training methodology. Our proposed solution is simple, effective, and can even be used in an active learning setting where the data labels are not present or being generated incrementally. We demonstrate the effectiveness of our algorithm for the task of object detection and multi-label image classification across different datasets. Through a series of experiments, we validate that curating contextually fair data helps make model predictions fair by balancing the true positive rate for the protected class across groups without compromising on the model's overall performance.

arXiv:2212.10105공식 코드 · Papers with Code

On the Applicability of Synthetic Data for Re-Identification

This contribution demonstrates the feasibility of applying Generative Adversarial Networks (GANs) on images of EPAL pallet blocks for dataset enhancement in the context of re-identification. For many industrial applications of re-identification methods, datasets of sufficient volume would otherwise be unattainable in non-laboratory settings. Using a state-of-the-art GAN architecture, namely CycleGAN, images of pallet blocks rotated to their left-hand side were generated from images of visually centered pallet blocks, based on images of rotated pallet blocks that were recorded as part of a previously recorded and published dataset. In this process, the unique chipwood pattern of the pallet block surface structure was retained, only changing the orientation of the pallet block itself. By doing so, synthetic data for re-identification testing and training purposes was generated, in a manner that is distinct from ordinary data augmentation. In total, 1,004 new images of pallet blocks were generated. The quality of the generated images was gauged using a perspective classifier that was trained on the original images and then applied to the synthetic ones, comparing the accuracy between the two sets of images. The classification accuracy was 98% for the original images and 92% for the synthetic images. In addition, the generated images were also used in a re-identification task, in order to re-identify original images based on synthetic ones. The accuracy in this scenario was up to 88% for synthetic images, compared to 96% for original images. Through this evaluation, it is established, whether or not a generated pallet block image closely resembles its original counterpart.

arXiv:2308.04588공식 코드 · Papers with Code

ScatterUQ: Interactive Uncertainty Visualizations for Multiclass Deep Learning Problems

Recently, uncertainty-aware deep learning methods for multiclass labeling problems have been developed that provide calibrated class prediction probabilities and out-of-distribution (OOD) indicators, letting machine learning (ML) consumers and engineers gauge a model's confidence in its predictions. However, this extra neural network prediction information is challenging to scalably convey visually for arbitrary data sources under multiple uncertainty contexts. To address these challenges, we present ScatterUQ, an interactive system that provides targeted visualizations to allow users to better understand model performance in context-driven uncertainty settings. ScatterUQ leverages recent advances in distance-aware neural networks, together with dimensionality reduction techniques, to construct robust, 2-D scatter plots explaining why a model predicts a test example to be (1) in-distribution and of a particular class, (2) in-distribution but unsure of the class, and (3) out-of-distribution. ML consumers and engineers can visually compare the salient features of test samples with training examples through the use of a ``hover callback'' to understand model uncertainty performance and decide follow up courses of action. We demonstrate the effectiveness of ScatterUQ to explain model uncertainty for a multiclass image classification on a distance-aware neural network trained on Fashion-MNIST and tested on Fashion-MNIST (in distribution) and MNIST digits (out of distribution), as well as a deep learning model for a cyber dataset. We quantitatively evaluate dimensionality reduction techniques to optimize our contextually driven UQ visualizations. Our results indicate that the ScatterUQ system should scale to arbitrary, multiclass datasets. Our code is available at https://github.com/mit-ll-responsible-ai/equine-webapp

arXiv:2404.14109공식 코드 · Papers with Code

CKD: Contrastive Knowledge Distillation from A Sample-wise Perspective

In this paper, we present a simple yet effective contrastive knowledge distillation approach, which can be formulated as a sample-wise alignment problem with intra- and inter-sample constraints. Unlike traditional knowledge distillation methods that concentrate on maximizing feature similarities or preserving class-wise semantic correlations between teacher and student features, our method attempts to recover the "dark knowledge" by aligning sample-wise teacher and student logits. Specifically, our method first minimizes logit differences within the same sample by considering their numerical values, thus preserving intra-sample similarities. Next, we bridge semantic disparities by leveraging dissimilarities across different samples. Note that constraints on intra-sample similarities and inter-sample dissimilarities can be efficiently and effectively reformulated into a contrastive learning framework with newly designed positive and negative pairs. The positive pair consists of the teacher's and student's logits derived from an identical sample, while the negative pairs are formed by using logits from different samples. With this formulation, our method benefits from the simplicity and efficiency of contrastive learning through the optimization of InfoNCE, yielding a run-time complexity that is far less than $O(n^2)$, where $n$ represents the total number of training samples. Furthermore, our method can eliminate the need for hyperparameter tuning, particularly related to temperature parameters and large batch sizes. We conduct comprehensive experiments on three datasets including CIFAR-100, ImageNet-1K, and MS COCO. Experimental results clearly confirm the effectiveness of the proposed method on both image classification and object detection tasks. Our source codes will be publicly available at https://github.com/wencheng-zhu/CKD.

arXiv:2501.11253공식 코드 · Papers with Code

How Well Do Supervised 3D Models Transfer to Medical Imaging Tasks?

The pre-training and fine-tuning paradigm has become prominent in transfer learning. For example, if the model is pre-trained on ImageNet and then fine-tuned to PASCAL, it can significantly outperform that trained on PASCAL from scratch. While ImageNet pre-training has shown enormous success, it is formed in 2D, and the learned features are for classification tasks; when transferring to more diverse tasks, like 3D image segmentation, its performance is inevitably compromised due to the deviation from the original ImageNet context. A significant challenge lies in the lack of large, annotated 3D datasets rivaling the scale of ImageNet for model pre-training. To overcome this challenge, we make two contributions. Firstly, we construct AbdomenAtlas 1.1 that comprises 9,262 three-dimensional computed tomography (CT) volumes with high-quality, per-voxel annotations of 25 anatomical structures and pseudo annotations of seven tumor types. Secondly, we develop a suite of models that are pre-trained on our AbdomenAtlas 1.1 for transfer learning. Our preliminary analyses indicate that the model trained only with 21 CT volumes, 672 masks, and 40 GPU hours has a transfer learning ability similar to the model trained with 5,050 (unlabeled) CT volumes and 1,152 GPU hours. More importantly, the transfer learning ability of supervised models can further scale up with larger annotated datasets, achieving significantly better performance than preexisting pre-trained models, irrespective of their pre-training methodologies or data sources. We hope this study can facilitate collective efforts in constructing larger 3D medical datasets and more releases of supervised pre-trained models.

arXiv:1901.00366공식 코드 · Papers with Code

Learning Efficient Detector with Semi-supervised Adaptive Distillation

Knowledge Distillation (KD) has been used in image classification for model compression. However, rare studies apply this technology on single-stage object detectors. Focal loss shows that the accumulated errors of easily-classified samples dominate the overall loss in the training process. This problem is also encountered when applying KD in the detection task. For KD, the teacher-defined hard samples are far more important than any others. We propose ADL to address this issue by adaptively mimicking the teacher's logits, with more attention paid on two types of hard samples: hard-to-learn samples predicted by teacher with low certainty and hard-to-mimic samples with a large gap between the teacher's and the student's prediction. ADL enlarges the distillation loss for hard-to-learn and hard-to-mimic samples and reduces distillation loss for the dominant easy samples, enabling distillation to work on the single-stage detector first time, even if the student and the teacher are identical. Besides, ADL is effective in both the supervised setting and the semi-supervised setting, even when the labeled data and unlabeled data are from different distributions. For distillation on unlabeled data, ADL achieves better performance than existing data distillation which simply utilizes hard targets, making the student detector surpass its teacher. On the COCO database, semi-supervised adaptive distillation (SAD) makes a student detector with a backbone of ResNet-50 surpasses its teacher with a backbone of ResNet-101, while the student has half of the teacher's computation complexity. The code is avaiable at https://github.com/Tangshitao/Semi-supervised-Adaptive-Distillation

arXiv:2106.12472공식 코드 · Papers with Code

Finding simplicity: unsupervised discovery of features, patterns, and order parameters via shift-invariant variational autoencoders

Recent advances in scanning tunneling and transmission electron microscopies (STM and STEM) have allowed routine generation of large volumes of imaging data containing information on the structure and functionality of materials. The experimental data sets contain signatures of long-range phenomena such as physical order parameter fields, polarization and strain gradients in STEM, or standing electronic waves and carrier-mediated exchange interactions in STM, all superimposed onto scanning system distortions and gradual changes of contrast due to drift and/or mis-tilt effects. Correspondingly, while the human eye can readily identify certain patterns in the images such as lattice periodicities, repeating structural elements, or microstructures, their automatic extraction and classification are highly non-trivial and universal pathways to accomplish such analyses are absent. We pose that the most distinctive elements of the patterns observed in STM and (S)TEM images are similarity and (almost-) periodicity, behaviors stemming directly from the parsimony of elementary atomic structures, superimposed on the gradual changes reflective of order parameter distributions. However, the discovery of these elements via global Fourier methods is non-trivial due to variability and lack of ideal discrete translation symmetry. To address this problem, we develop shift-invariant variational autoencoders (shift-VAE) that allow disentangling characteristic repeating features in the images, their variations, and shifts inevitable for random sampling of image space. Shift-VAEs balance the uncertainty in the position of the object of interest with the uncertainty in shape reconstruction. This approach is illustrated for model 1D data, and further extended to synthetic and experimental STM and STEM 2D data.

arXiv:1810.06583공식 코드 · Papers with Code

Concise Explanations of Neural Networks using Adversarial Training

We show new connections between adversarial learning and explainability for deep neural networks (DNNs). One form of explanation of the output of a neural network model in terms of its input features, is a vector of feature-attributions. Two desirable characteristics of an attribution-based explanation are: (1) $\textit{sparseness}$: the attributions of irrelevant or weakly relevant features should be negligible, thus resulting in $\textit{concise}$ explanations in terms of the significant features, and (2) $\textit{stability}$: it should not vary significantly within a small local neighborhood of the input. Our first contribution is a theoretical exploration of how these two properties (when using attributions based on Integrated Gradients, or IG) are related to adversarial training, for a class of 1-layer networks (which includes logistic regression models for binary and multi-class classification); for these networks we show that (a) adversarial training using an $\ell_\infty$-bounded adversary produces models with sparse attribution vectors, and (b) natural model-training while encouraging stable explanations (via an extra term in the loss function), is equivalent to adversarial training. Our second contribution is an empirical verification of phenomenon (a), which we show, somewhat surprisingly, occurs $\textit{not only}$ $\textit{in 1-layer networks}$, $\textit{but also DNNs}$ $\textit{trained on }$ $\textit{standard image datasets}$, and extends beyond IG-based attributions, to those based on DeepSHAP: adversarial training with $\ell_\infty$-bounded perturbations yields significantly sparser attribution vectors, with little degradation in performance on natural test data, compared to natural training. Moreover, the sparseness of the attribution vectors is significantly better than that achievable via $\ell_1$-regularized natural training.

arXiv:2004.07882공식 코드 · Papers with Code

Models Genesis

Transfer learning from natural images to medical images has been established as one of the most practical paradigms in deep learning for medical image analysis. To fit this paradigm, however, 3D imaging tasks in the most prominent imaging modalities (e.g., CT and MRI) have to be reformulated and solved in 2D, losing rich 3D anatomical information, thereby inevitably compromising its performance. To overcome this limitation, we have built a set of models, called Generic Autodidactic Models, nicknamed Models Genesis, because they are created ex nihilo (with no manual labeling), self-taught (learnt by self-supervision), and generic (served as source models for generating application-specific target models). Our extensive experiments demonstrate that our Models Genesis significantly outperform learning from scratch and existing pre-trained 3D models in all five target 3D applications covering both segmentation and classification. More importantly, learning a model from scratch simply in 3D may not necessarily yield performance better than transfer learning from ImageNet in 2D, but our Models Genesis consistently top any 2D/2.5D approaches including fine-tuning the models pre-trained from ImageNet as well as fine-tuning the 2D versions of our Models Genesis, confirming the importance of 3D anatomical information and significance of Models Genesis for 3D medical imaging. This performance is attributed to our unified self-supervised learning framework, built on a simple yet powerful observation: the sophisticated and recurrent anatomy in medical images can serve as strong yet free supervision signals for deep models to learn common anatomical representation automatically via self-supervision. As open science, all codes and pre-trained Models Genesis are available at https://github.com/MrGiovanni/ModelsGenesis.

arXiv:2106.09017공식 코드 · Papers with Code

Bridging Multi-Task Learning and Meta-Learning: Towards Efficient Training and Effective Adaptation

Multi-task learning (MTL) aims to improve the generalization of several related tasks by learning them jointly. As a comparison, in addition to the joint training scheme, modern meta-learning allows unseen tasks with limited labels during the test phase, in the hope of fast adaptation over them. Despite the subtle difference between MTL and meta-learning in the problem formulation, both learning paradigms share the same insight that the shared structure between existing training tasks could lead to better generalization and adaptation. In this paper, we take one important step further to understand the close connection between these two learning paradigms, through both theoretical analysis and empirical investigation. Theoretically, we first demonstrate that MTL shares the same optimization formulation with a class of gradient-based meta-learning (GBML) algorithms. We then prove that for over-parameterized neural networks with sufficient depth, the learned predictive functions of MTL and GBML are close. In particular, this result implies that the predictions given by these two models are similar over the same unseen task. Empirically, we corroborate our theoretical findings by showing that, with proper implementation, MTL is competitive against state-of-the-art GBML algorithms on a set of few-shot image classification benchmarks. Since existing GBML algorithms often involve costly second-order bi-level optimization, our first-order MTL method is an order of magnitude faster on large-scale datasets such as mini-ImageNet. We believe this work could help bridge the gap between these two learning paradigms, and provide a computationally efficient alternative to GBML that also supports fast task adaptation.

arXiv:2205.12956공식 코드 · Papers with Code

Inception Transformer

Recent studies show that Transformer has strong capability of building long-range dependencies, yet is incompetent in capturing high frequencies that predominantly convey local information. To tackle this issue, we present a novel and general-purpose Inception Transformer, or iFormer for short, that effectively learns comprehensive features with both high- and low-frequency information in visual data. Specifically, we design an Inception mixer to explicitly graft the advantages of convolution and max-pooling for capturing the high-frequency information to Transformers. Different from recent hybrid frameworks, the Inception mixer brings greater efficiency through a channel splitting mechanism to adopt parallel convolution/max-pooling path and self-attention path as high- and low-frequency mixers, while having the flexibility to model discriminative information scattered within a wide frequency range. Considering that bottom layers play more roles in capturing high-frequency details while top layers more in modeling low-frequency global information, we further introduce a frequency ramp structure, i.e. gradually decreasing the dimensions fed to the high-frequency mixer and increasing those to the low-frequency mixer, which can effectively trade-off high- and low-frequency components across different layers. We benchmark the iFormer on a series of vision tasks, and showcase that it achieves impressive performance on image classification, COCO detection and ADE20K segmentation. For example, our iFormer-S hits the top-1 accuracy of 83.4% on ImageNet-1K, much higher than DeiT-S by 3.6%, and even slightly better than much bigger model Swin-B (83.3%) with only 1/4 parameters and 1/3 FLOPs. Code and models will be released at https://github.com/sail-sg/iFormer.