PyTorchKR Papers

“Image Classification”

총 5,746편
태그·기간 등 상세 필터

카테고리 탐색

arXiv:2205.01316공식 코드 · Papers with Code

HL-Net: Heterophily Learning Network for Scene Graph Generation

Scene graph generation (SGG) aims to detect objects and predict their pairwise relationships within an image. Current SGG methods typically utilize graph neural networks (GNNs) to acquire context information between objects/relationships. Despite their effectiveness, however, current SGG methods only assume scene graph homophily while ignoring heterophily. Accordingly, in this paper, we propose a novel Heterophily Learning Network (HL-Net) to comprehensively explore the homophily and heterophily between objects/relationships in scene graphs. More specifically, HL-Net comprises the following 1) an adaptive reweighting transformer module, which adaptively integrates the information from different layers to exploit both the heterophily and homophily in objects; 2) a relationship feature propagation module that efficiently explores the connections between relationships by considering heterophily in order to refine the relationship representation; 3) a heterophily-aware message-passing scheme to further distinguish the heterophily and homophily between objects/relationships, thereby facilitating improved message passing in graphs. We conducted extensive experiments on two public datasets: Visual Genome (VG) and Open Images (OI). The experimental results demonstrate the superiority of our proposed HL-Net over existing state-of-the-art approaches. In more detail, HL-Net outperforms the second-best competitors by 2.1$\%$ on the VG dataset for scene graph classification and 1.2$\%$ on the IO dataset for the final score. Code is available at https://github.com/siml3/HL-Net.

arXiv:2307.12280공식 코드 · Papers with Code

Downstream-agnostic Adversarial Examples

Self-supervised learning usually uses a large amount of unlabeled data to pre-train an encoder which can be used as a general-purpose feature extractor, such that downstream users only need to perform fine-tuning operations to enjoy the benefit of "large model". Despite this promising prospect, the security of pre-trained encoder has not been thoroughly investigated yet, especially when the pre-trained encoder is publicly available for commercial use. In this paper, we propose AdvEncoder, the first framework for generating downstream-agnostic universal adversarial examples based on the pre-trained encoder. AdvEncoder aims to construct a universal adversarial perturbation or patch for a set of natural images that can fool all the downstream tasks inheriting the victim pre-trained encoder. Unlike traditional adversarial example works, the pre-trained encoder only outputs feature vectors rather than classification labels. Therefore, we first exploit the high frequency component information of the image to guide the generation of adversarial examples. Then we design a generative attack framework to construct adversarial perturbations/patches by learning the distribution of the attack surrogate dataset to improve their attack success rates and transferability. Our results show that an attacker can successfully attack downstream tasks without knowing either the pre-training dataset or the downstream dataset. We also tailor four defenses for pre-trained encoders, the results of which further prove the attack ability of AdvEncoder.

arXiv:2307.15250공식 코드 · Papers with Code

D2S: Representing sparse descriptors and 3D coordinates for camera relocalization

State-of-the-art visual localization methods mostly rely on complex procedures to match local descriptors and 3D point clouds. However, these procedures can incur significant costs in terms of inference, storage, and updates over time. In this study, we propose a direct learning-based approach that utilizes a simple network named D2S to represent complex local descriptors and their scene coordinates. Our method is characterized by its simplicity and cost-effectiveness. It solely leverages a single RGB image for localization during the testing phase and only requires a lightweight model to encode a complex sparse scene. The proposed D2S employs a combination of a simple loss function and graph attention to selectively focus on robust descriptors while disregarding areas such as clouds, trees, and several dynamic objects. This selective attention enables D2S to effectively perform a binary-semantic classification for sparse descriptors. Additionally, we propose a simple outdoor dataset to evaluate the capabilities of visual localization methods in scene-specific generalization and self-updating from unlabeled observations. Our approach outperforms the previous regression-based methods in both indoor and outdoor environments. It demonstrates the ability to generalize beyond training data, including scenarios involving transitions from day to night and adapting to domain shifts. The source code, trained models, dataset, and demo videos are available at the following link: https://thpjp.github.io/d2s.

arXiv:2403.13319공식 코드 · Papers with Code

HyperFusion: A Hypernetwork Approach to Multimodal Integration of Tabular and Medical Imaging Data for Predictive Modeling

The integration of diverse clinical modalities such as medical imaging and the tabular data extracted from patients' Electronic Health Records (EHRs) is a crucial aspect of modern healthcare. Integrative analysis of multiple sources can provide a comprehensive understanding of the clinical condition of a patient, improving diagnosis and treatment decision. Deep Neural Networks (DNNs) consistently demonstrate outstanding performance in a wide range of multimodal tasks in the medical domain. However, the complex endeavor of effectively merging medical imaging with clinical, demographic and genetic information represented as numerical tabular data remains a highly active and ongoing research pursuit. We present a novel framework based on hypernetworks to fuse clinical imaging and tabular data by conditioning the image processing on the EHR's values and measurements. This approach aims to leverage the complementary information present in these modalities to enhance the accuracy of various medical applications. We demonstrate the strength and generality of our method on two different brain Magnetic Resonance Imaging (MRI) analysis tasks, namely, brain age prediction conditioned by subject's sex and multi-class Alzheimer's Disease (AD) classification conditioned by tabular data. We show that our framework outperforms both single-modality models and state-of-the-art MRI tabular data fusion methods. A link to our code can be found at https://github.com/daniel4725/HyperFusion

arXiv:2403.19067공식 코드 · Papers with Code

Low-Rank Rescaled Vision Transformer Fine-Tuning: A Residual Design Approach

Parameter-efficient fine-tuning for pre-trained Vision Transformers aims to adeptly tailor a model to downstream tasks by learning a minimal set of new adaptation parameters while preserving the frozen majority of pre-trained parameters. Striking a balance between retaining the generalizable representation capacity of the pre-trained model and acquiring task-specific features poses a key challenge. Currently, there is a lack of focus on guiding this delicate trade-off. In this study, we approach the problem from the perspective of Singular Value Decomposition (SVD) of pre-trained parameter matrices, providing insights into the tuning dynamics of existing methods. Building upon this understanding, we propose a Residual-based Low-Rank Rescaling (RLRR) fine-tuning strategy. This strategy not only enhances flexibility in parameter tuning but also ensures that new parameters do not deviate excessively from the pre-trained model through a residual design. Extensive experiments demonstrate that our method achieves competitive performance across various downstream image classification tasks, all while maintaining comparable new parameters. We believe this work takes a step forward in offering a unified perspective for interpreting existing methods and serves as motivation for the development of new approaches that move closer to effectively considering the crucial trade-off mentioned above. Our code is available at \href{https://github.com/zstarN70/RLRR.git}{https://github.com/zstarN70/RLRR.git}.

arXiv:2410.08003공식 코드 · Papers with Code

More Experts Than Galaxies: Conditionally-overlapping Experts With Biologically-Inspired Fixed Routing

The evolution of biological neural systems has led to both modularity and sparse coding, which enables energy efficiency and robustness across the diversity of tasks in the lifespan. In contrast, standard neural networks rely on dense, non-specialized architectures, where all model parameters are simultaneously updated to learn multiple tasks, leading to interference. Current sparse neural network approaches aim to alleviate this issue but are hindered by limitations such as 1) trainable gating functions that cause representation collapse, 2) disjoint experts that result in redundant computation and slow learning, and 3) reliance on explicit input or task IDs that limit flexibility and scalability. In this paper we propose Conditionally Overlapping Mixture of ExperTs (COMET), a general deep learning method that addresses these challenges by inducing a modular, sparse architecture with an exponential number of overlapping experts. COMET replaces the trainable gating function used in Sparse Mixture of Experts with a fixed, biologically inspired random projection applied to individual input representations. This design causes the degree of expert overlap to depend on input similarity, so that similar inputs tend to share more parameters. This results in faster learning \rev{per update step} and improved \rev{out-of-sample} generalization. We demonstrate the effectiveness of COMET on a range of tasks, including image classification, language modeling, and regression, using several popular deep learning architectures.

arXiv:2111.05955공식 코드 · Papers with Code

Keys to Accurate Feature Extraction Using Residual Spiking Neural Networks

Spiking neural networks (SNNs) have become an interesting alternative to conventional artificial neural networks (ANN) thanks to their temporal processing capabilities and energy efficient implementations in neuromorphic hardware. However the challenges involved in training SNNs have limited their performance in terms of accuracy and thus their applications. Improving learning algorithms and neural architectures for a more accurate feature extraction is therefore one of the current priorities in SNN research. In this paper we present a study on the key components of modern spiking architectures. We design a spiking version of the successful residual network architecture and provide an in-depth study on the possible implementations of spiking residual connections. This study shows how, depending on the use case, the optimal residual connection implementation may vary. Additionally, we empirically compare different techniques in image classification datasets taken from the best performing networks. Our results provide a state of the art guide to SNN design, which allows to make informed choices when trying to build the optimal visual feature extractor. Finally, our network outperforms previous SNN architectures in CIFAR-10 (94.14%) and CIFAR-100 (74.65%) datasets and matches the state of the art in DVS-CIFAR10 (72.98%), with less parameters than the previous state of the art and without the need for ANN-SNN conversion. Code available at https://github.com/VicenteAlex/Spiking_ResNet

arXiv:2112.07106공식 코드 · Papers with Code

E-CRF: Embedded Conditional Random Field for Boundary-caused Class Weights Confusion in Semantic Segmentation

Modern semantic segmentation methods devote much effect to adjusting image feature representations to improve the segmentation performance in various ways, such as architecture design, attention mechnism, etc. However, almost all those methods neglect the particularity of class weights (in the classification layer) in segmentation models. In this paper, we notice that the class weights of categories that tend to share many adjacent boundary pixels lack discrimination, thereby limiting the performance. We call this issue Boundary-caused Class Weights Confusion (BCWC). We try to focus on this problem and propose a novel method named Embedded Conditional Random Field (E-CRF) to alleviate it. E-CRF innovatively fuses the CRF into the CNN network as an organic whole for more effective end-to-end optimization. The reasons are two folds. It utilizes CRF to guide the message passing between pixels in high-level features to purify the feature representation of boundary pixels, with the help of inner pixels belonging to the same object. More importantly, it enables optimizing class weights from both scale and direction during backpropagation. We make detailed theoretical analysis to prove it. Besides, superpixel is integrated into E-CRF and served as an auxiliary to exploit the local object prior for more reliable message passing. Finally, our proposed method yields impressive results on ADE20K, Cityscapes, and Pascal Context datasets.

arXiv:2202.01179공식 코드 · Papers with Code

AntidoteRT: Run-time Detection and Correction of Poison Attacks on Neural Networks

We study backdoor poisoning attacks against image classification networks, whereby an attacker inserts a trigger into a subset of the training data, in such a way that at test time, this trigger causes the classifier to predict some target class. %There are several techniques proposed in the literature that aim to detect the attack but only a few also propose to defend against it, and they typically involve retraining the network which is not always possible in practice. We propose lightweight automated detection and correction techniques against poisoning attacks, which are based on neuron patterns mined from the network using a small set of clean and poisoned test samples with known labels. The patterns built based on the mis-classified samples are used for run-time detection of new poisoned inputs. For correction, we propose an input correction technique that uses a differential analysis to identify the trigger in the detected poisoned images, which is then reset to a neutral color. Our detection and correction are performed at run-time and input level, which is in contrast to most existing work that is focused on offline model-level defenses. We demonstrate that our technique outperforms existing defenses such as NeuralCleanse and STRIP on popular benchmarks such as MNIST, CIFAR-10, and GTSRB against the popular BadNets attack and the more complex DFST attack.

arXiv:2208.06163공식 코드 · Papers with Code

Dropout is NOT All You Need to Prevent Gradient Leakage

Gradient inversion attacks on federated learning systems reconstruct client training data from exchanged gradient information. To defend against such attacks, a variety of defense mechanisms were proposed. However, they usually lead to an unacceptable trade-off between privacy and model utility. Recent observations suggest that dropout could mitigate gradient leakage and improve model utility if added to neural networks. Unfortunately, this phenomenon has not been systematically researched yet. In this work, we thoroughly analyze the effect of dropout on iterative gradient inversion attacks. We find that state of the art attacks are not able to reconstruct the client data due to the stochasticity induced by dropout during model training. Nonetheless, we argue that dropout does not offer reliable protection if the dropout induced stochasticity is adequately modeled during attack optimization. Consequently, we propose a novel Dropout Inversion Attack (DIA) that jointly optimizes for client data and dropout masks to approximate the stochastic client model. We conduct an extensive systematic evaluation of our attack on four seminal model architectures and three image classification datasets of increasing complexity. We find that our proposed attack bypasses the protection seemingly induced by dropout and reconstructs client data with high fidelity. Our work demonstrates that privacy inducing changes to model architectures alone cannot be assumed to reliably protect from gradient leakage and therefore should be combined with complementary defense mechanisms.

arXiv:2304.03646공식 코드 · Papers with Code

Fairness through Aleatoric Uncertainty

We propose a simple yet effective solution to tackle the often-competing goals of fairness and utility in classification tasks. While fairness ensures that the model's predictions are unbiased and do not discriminate against any particular group or individual, utility focuses on maximizing the model's predictive performance. This work introduces the idea of leveraging aleatoric uncertainty (e.g., data ambiguity) to improve the fairness-utility trade-off. Our central hypothesis is that aleatoric uncertainty is a key factor for algorithmic fairness and samples with low aleatoric uncertainty are modeled more accurately and fairly than those with high aleatoric uncertainty. We then propose a principled model to improve fairness when aleatoric uncertainty is high and improve utility elsewhere. Our approach first intervenes in the data distribution to better decouple aleatoric uncertainty and epistemic uncertainty. It then introduces a fairness-utility bi-objective loss defined based on the estimated aleatoric uncertainty. Our approach is theoretically guaranteed to improve the fairness-utility trade-off. Experimental results on both tabular and image datasets show that the proposed approach outperforms state-of-the-art methods w.r.t. the fairness-utility trade-off and w.r.t. both group and individual fairness metrics. This work presents a fresh perspective on the trade-off between utility and algorithmic fairness and opens a key avenue for the potential of using prediction uncertainty in fair machine learning.

arXiv:2407.07841공식 코드 · Papers with Code

Benchmarking Embedding Aggregation Methods in Computational Pathology: A Clinical Data Perspective

Recent advances in artificial intelligence (AI), in particular self-supervised learning of foundation models (FMs), are revolutionizing medical imaging and computational pathology (CPath). A constant challenge in the analysis of digital Whole Slide Images (WSIs) is the problem of aggregating tens of thousands of tile-level image embeddings to a slide-level representation. Due to the prevalent use of datasets created for genomic research, such as TCGA, for method development, the performance of these techniques on diagnostic slides from clinical practice has been inadequately explored. This study conducts a thorough benchmarking analysis of ten slide-level aggregation techniques across nine clinically relevant tasks, including diagnostic assessment, biomarker classification, and outcome prediction. The results yield following key insights: (1) Embeddings derived from domain-specific (histological images) FMs outperform those from generic ImageNet-based models across aggregation methods. (2) Spatial-aware aggregators enhance the performance significantly when using ImageNet pre-trained models but not when using FMs. (3) No single model excels in all tasks and spatially-aware models do not show general superiority as it would be expected. These findings underscore the need for more adaptable and universally applicable aggregation techniques, guiding future research towards tools that better meet the evolving needs of clinical-AI in pathology. The code used in this work is available at \url{https://github.com/fuchs-lab-public/CPath_SABenchmark}.

arXiv:1908.10468공식 코드 · Papers with Code

Adversarial regression training for visualizing the progression of chronic obstructive pulmonary disease with chest x-rays

Knowledge of what spatial elements of medical images deep learning methods use as evidence is important for model interpretability, trustiness, and validation. There is a lack of such techniques for models in regression tasks. We propose a method, called visualization for regression with a generative adversarial network (VR-GAN), for formulating adversarial training specifically for datasets containing regression target values characterizing disease severity. We use a conditional generative adversarial network where the generator attempts to learn to shift the output of a regressor through creating disease effect maps that are added to the original images. Meanwhile, the regressor is trained to predict the original regression value for the modified images. A model trained with this technique learns to provide visualization for how the image would appear at different stages of the disease. We analyze our method in a dataset of chest x-rays associated with pulmonary function tests, used for diagnosing chronic obstructive pulmonary disease (COPD). For validation, we compute the difference of two registered x-rays of the same patient at different time points and correlate it to the generated disease effect map. The proposed method outperforms a technique based on classification and provides realistic-looking images, making modifications to images following what radiologists usually observe for this disease. Implementation code is available at https://github.com/ricbl/vrgan.

arXiv:1912.09533공식 코드 · Papers with Code

Towards Verifying Robustness of Neural Networks Against Semantic Perturbations

Verifying robustness of neural networks given a specified threat model is a fundamental yet challenging task. While current verification methods mainly focus on the $\ell_p$-norm threat model of the input instances, robustness verification against semantic adversarial attacks inducing large $\ell_p$-norm perturbations, such as color shifting and lighting adjustment, are beyond their capacity. To bridge this gap, we propose \textit{Semantify-NN}, a model-agnostic and generic robustness verification approach against semantic perturbations for neural networks. By simply inserting our proposed \textit{semantic perturbation layers} (SP-layers) to the input layer of any given model, \textit{Semantify-NN} is model-agnostic, and any $\ell_p$-norm based verification tools can be used to verify the model robustness against semantic perturbations. We illustrate the principles of designing the SP-layers and provide examples including semantic perturbations to image classification in the space of hue, saturation, lightness, brightness, contrast and rotation, respectively. In addition, an efficient refinement technique is proposed to further significantly improve the semantic certificate. Experiments on various network architectures and different datasets demonstrate the superior verification performance of \textit{Semantify-NN} over $\ell_p$-norm-based verification frameworks that naively convert semantic perturbation to $\ell_p$-norm. The results show that \textit{Semantify-NN} can support robustness verification against a wide range of semantic perturbations. Code available https://github.com/JeetMo/Semantify-NN

arXiv:2206.01366공식 코드 · Papers with Code

Supernet Training for Federated Image Classification under System Heterogeneity

Efficient deployment of deep neural networks across many devices and resource constraints, particularly on edge devices, is one of the most challenging problems in the presence of data-privacy preservation issues. Conventional approaches have evolved to either improve a single global model while keeping each local heterogeneous training data decentralized (i.e. data heterogeneity; Federated Learning (FL)) or to train an overarching network that supports diverse architectural settings to address heterogeneous systems equipped with different computational capabilities (i.e. system heterogeneity; Neural Architecture Search). However, few studies have considered both directions simultaneously. This paper proposes the federation of supernet training (FedSup) framework to consider both scenarios simultaneously, i.e., where clients send and receive a supernet that contains all possible architectures sampled from itself. The approach is inspired by observing that averaging parameters during model aggregation for FL is similar to weight-sharing in supernet training. Thus, the proposed FedSup framework combines a weight-sharing approach widely used for training single shot models with FL averaging (FedAvg). Furthermore, we develop an efficient algorithm (E-FedSup) by sending the sub-model to clients on the broadcast stage to reduce communication costs and training overhead, including several strategies to enhance supernet training in the FL environment. We verify the proposed approach with extensive empirical evaluations. The resulting framework also ensures data and model heterogeneity robustness on several standard benchmarks.

arXiv:2206.07736공식 코드 · Papers with Code

Improving Diversity with Adversarially Learned Transformations for Domain Generalization

To be successful in single source domain generalization, maximizing diversity of synthesized domains has emerged as one of the most effective strategies. Many of the recent successes have come from methods that pre-specify the types of diversity that a model is exposed to during training, so that it can ultimately generalize well to new domains. However, na\"ive diversity based augmentations do not work effectively for domain generalization either because they cannot model large domain shift, or because the span of transforms that are pre-specified do not cover the types of shift commonly occurring in domain generalization. To address this issue, we present a novel framework that uses adversarially learned transformations (ALT) using a neural network to model plausible, yet hard image transformations that fool the classifier. This network is randomly initialized for each batch and trained for a fixed number of steps to maximize classification error. Further, we enforce consistency between the classifier's predictions on the clean and transformed images. With extensive empirical analysis, we find that this new form of adversarial transformations achieve both objectives of diversity and hardness simultaneously, outperforming all existing techniques on competitive benchmarks for single source domain generalization. We also show that ALT can naturally work with existing diversity modules to produce highly distinct, and large transformations of the source domain leading to state-of-the-art performance.

arXiv:2210.06311공식 코드 · Papers with Code

Semantic Cross Attention for Few-shot Learning

Few-shot learning (FSL) has attracted considerable attention recently. Among existing approaches, the metric-based method aims to train an embedding network that can make similar samples close while dissimilar samples as far as possible and achieves promising results. FSL is characterized by using only a few images to train a model that can generalize to novel classes in image classification problems, but this setting makes it difficult to learn the visual features that can identify the images' appearance variations. The model training is likely to move in the wrong direction, as the images in an identical semantic class may have dissimilar appearances, whereas the images in different semantic classes may share a similar appearance. We argue that FSL can benefit from additional semantic features to learn discriminative feature representations. Thus, this study proposes a multi-task learning approach to view semantic features of label text as an auxiliary task to help boost the performance of the FSL task. Our proposed model uses word-embedding representations as semantic features to help train the embedding network and a semantic cross-attention module to bridge the semantic features into the typical visual modal. The proposed approach is simple, but produces excellent results. We apply our proposed approach to two previous metric-based FSL methods, all of which can substantially improve performance. The source code for our model is accessible from github.

arXiv:2210.09224공식 코드 · Papers with Code

Self-Supervised Learning Through Efference Copies

Self-supervised learning (SSL) methods aim to exploit the abundance of unlabelled data for machine learning (ML), however the underlying principles are often method-specific. An SSL framework derived from biological first principles of embodied learning could unify the various SSL methods, help elucidate learning in the brain, and possibly improve ML. SSL commonly transforms each training datapoint into a pair of views, uses the knowledge of this pairing as a positive (i.e. non-contrastive) self-supervisory sign, and potentially opposes it to unrelated, (i.e. contrastive) negative examples. Here, we show that this type of self-supervision is an incomplete implementation of a concept from neuroscience, the Efference Copy (EC). Specifically, the brain also transforms the environment through efference, i.e. motor commands, however it sends to itself an EC of the full commands, i.e. more than a mere SSL sign. In addition, its action representations are likely egocentric. From such a principled foundation we formally recover and extend SSL methods such as SimCLR, BYOL, and ReLIC under a common theoretical framework, i.e. Self-supervision Through Efference Copies (S-TEC). Empirically, S-TEC restructures meaningfully the within- and between-class representations. This manifests as improvement in recent strong SSL baselines in image classification, segmentation, object detection, and in audio. These results hypothesize a testable positive influence from the brain's motor outputs onto its sensory representations.

arXiv:2310.09276공식 코드 · Papers with Code

Transformer-based Multimodal Change Detection with Multitask Consistency Constraints

Change detection plays a fundamental role in Earth observation for analyzing temporal iterations over time. However, recent studies have largely neglected the utilization of multimodal data that presents significant practical and technical advantages compared to single-modal approaches. This research focuses on leveraging {pre-event} digital surface model (DSM) data and {post-event} digital aerial images captured at different times for detecting change beyond 2D. We observe that the current change detection methods struggle with the multitask conflicts between semantic and height change detection tasks. To address this challenge, we propose an efficient Transformer-based network that learns shared representation between cross-dimensional inputs through cross-attention. {It adopts a consistency constraint to establish the multimodal relationship. Initially, pseudo-changes are derived by employing height change thresholding. Subsequently, the $L2$ distance between semantic and pseudo-changes within their overlapping regions is minimized. This explicitly endows the height change detection (regression task) and semantic change detection (classification task) with representation consistency.} A DSM-to-image multimodal dataset encompassing three cities in the Netherlands was constructed. It lays a new foundation for beyond-2D change detection from cross-dimensional inputs. Compared to five state-of-the-art change detection methods, our model demonstrates consistent multitask superiority in terms of semantic and height change detection. Furthermore, the consistency strategy can be seamlessly adapted to the other methods, yielding promising improvements.

arXiv:2311.15157공식 코드 · Papers with Code

Advancing Vision Transformers with Group-Mix Attention

Vision Transformers (ViTs) have been shown to enhance visual recognition through modeling long-range dependencies with multi-head self-attention (MHSA), which is typically formulated as Query-Key-Value computation. However, the attention map generated from the Query and Key captures only token-to-token correlations at one single granularity. In this paper, we argue that self-attention should have a more comprehensive mechanism to capture correlations among tokens and groups (i.e., multiple adjacent tokens) for higher representational capacity. Thereby, we propose Group-Mix Attention (GMA) as an advanced replacement for traditional self-attention, which can simultaneously capture token-to-token, token-to-group, and group-to-group correlations with various group sizes. To this end, GMA splits the Query, Key, and Value into segments uniformly and performs different group aggregations to generate group proxies. The attention map is computed based on the mixtures of tokens and group proxies and used to re-combine the tokens and groups in Value. Based on GMA, we introduce a powerful backbone, namely GroupMixFormer, which achieves state-of-the-art performance in image classification, object detection, and semantic segmentation with fewer parameters than existing models. For instance, GroupMixFormer-L (with 70.3M parameters and 384^2 input) attains 86.2% Top-1 accuracy on ImageNet-1K without external data, while GroupMixFormer-B (with 45.8M parameters) attains 51.2% mIoU on ADE20K.