Deep neural networks are typically trained using global error signals that backpropagate (BP) end-to-end, which is not only biologically implausible but also suffers from the update locking problem and requires huge memory consumption. Local learning, which updates each layer independently with a gradient-isolated auxiliary network, offers a promising alternative to address the above problems. However, existing local learning methods are confronted with a large accuracy gap with the BP counterpart, particularly for large-scale networks. This is due to the weak coupling between local layers and their subsequent network layers, as there is no gradient communication across layers. To tackle this issue, we put forward an augmented local learning method, dubbed AugLocal. AugLocal constructs each hidden layer's auxiliary network by uniformly selecting a small subset of layers from its subsequent network layers to enhance their synergy. We also propose to linearly reduce the depth of auxiliary networks as the hidden layer goes deeper, ensuring sufficient network capacity while reducing the computational cost of auxiliary networks. Our extensive experiments on four image classification datasets (i.e., CIFAR-10, SVHN, STL-10, and ImageNet) demonstrate that AugLocal can effectively scale up to tens of local layers with a comparable accuracy to BP-trained networks while reducing GPU memory usage by around 40%. The proposed AugLocal method, therefore, opens up a myriad of opportunities for training high-performance deep neural networks on resource-constrained platforms.Code is available at https://github.com/ChenxiangMA/AugLocal.
Polarimetric synthetic aperture radar (PolSAR) images encompass valuable information that can facilitate extensive land cover interpretation and generate diverse output products. Extracting meaningful features from PolSAR data poses challenges distinct from those encountered in optical imagery. Deep learning (DL) methods offer effective solutions for overcoming these challenges in PolSAR feature extraction. Convolutional neural networks (CNNs) play a crucial role in capturing PolSAR image characteristics by leveraging kernel capabilities to consider local information and the complex-valued nature of PolSAR data. In this study, a novel three-branch fusion of complex-valued CNN, named the Shallow to Deep Feature Fusion Network (SDF2Net), is proposed for PolSAR image classification. To validate the performance of the proposed method, classification results are compared against multiple state-of-the-art approaches using the airborne synthetic aperture radar (AIRSAR) datasets of Flevoland and San Francisco, as well as the ESAR Oberpfaffenhofen dataset. The results indicate that the proposed approach demonstrates improvements in overallaccuracy, with a 1.3% and 0.8% enhancement for the AIRSAR datasets and a 0.5% improvement for the ESAR dataset. Analyses conducted on the Flevoland data underscore the effectiveness of the SDF2Net model, revealing a promising overall accuracy of 96.01% even with only a 1% sampling ratio.
The training paradigm of DETRs is heavily contingent upon pre-training their backbone on the ImageNet dataset. However, the limited supervisory signals provided by the image classification task and one-to-one matching strategy result in an inadequately pre-trained neck for DETRs. Additionally, the instability of matching in the early stages of training engenders inconsistencies in the optimization objectives of DETRs. To address these issues, we have devised an innovative training methodology termed step-by-step training. Specifically, in the first stage of training, we employ a classic detector, pre-trained with a one-to-many matching strategy, to initialize the backbone and neck of the end-to-end detector. In the second stage of training, we froze the backbone and neck of the end-to-end detector, necessitating the training of the decoder from scratch. Through the application of step-by-step training, we have introduced the first real-time end-to-end object detection model that utilizes a purely convolutional structure encoder, DETR with YOLO (DEYO). Without reliance on any supplementary training data, DEYO surpasses all existing real-time object detectors in both speed and accuracy. Moreover, the comprehensive DEYO series can complete its second-phase training on the COCO dataset using a single 8GB RTX 4060 GPU, significantly reducing the training expenditure. Source code and pre-trained models are available at https://github.com/ouyanghaodong/DEYO.
Traditional deep learning approaches for breast cancer classification has predominantly concentrated on single-view analysis. In clinical practice, however, radiologists concurrently examine all views within a mammography exam, leveraging the inherent correlations in these views to effectively detect tumors. Acknowledging the significance of multi-view analysis, some studies have introduced methods that independently process mammogram views, either through distinct convolutional branches or simple fusion strategies, inadvertently leading to a loss of crucial inter-view correlations. In this paper, we propose an innovative multi-view network exclusively based on transformers to address challenges in mammographic image classification. Our approach introduces a novel shifted window-based dynamic attention block, facilitating the effective integration of multi-view information and promoting the coherent transfer of this information between views at the spatial feature map level. Furthermore, we conduct a comprehensive comparative analysis of the performance and effectiveness of transformer-based models under diverse settings, employing the CBIS-DDSM and Vin-Dr Mammo datasets. Our code is publicly available at https://github.com/prithuls/MV-Swin-T
Deep neural networks (DNNs) have proven to be effective models for accurate Memory Access Prediction (MAP), a critical task in mitigating memory latency through data prefetching. However, existing DNN-based MAP models suffer from the challenges such as significant physical storage space and poor inference latency, primarily due to their large number of parameters. These limitations render them impractical for deployment in real-world scenarios. In this paper, we propose PaCKD, a Pattern-Clustered Knowledge Distillation approach to compress MAP models while maintaining the prediction performance. The PaCKD approach encompasses three steps: clustering memory access sequences into distinct partitions involving similar patterns, training large pattern-specific teacher models for memory access prediction for each partition, and training a single lightweight student model by distilling the knowledge from the trained pattern-specific teachers. We evaluate our approach on LSTM, MLP-Mixer, and ResNet models, as they exhibit diverse structures and are widely used for image classification tasks in order to test their effectiveness in four widely used graph applications. Compared to the teacher models with 5.406M parameters and an F1-score of 0.4626, our student models achieve a 552$\times$ model size compression while maintaining an F1-score of 0.4538 (with a 1.92% performance drop). Our approach yields an 8.70% higher result compared to student models trained with standard knowledge distillation and an 8.88% higher result compared to student models trained without any form of knowledge distillation.
Traditional deep learning models implicity encode knowledge limiting their transparency and ability to adapt to data changes. Yet, this adaptability is vital for addressing user data privacy concerns. We address this limitation by storing embeddings of the underlying training data independently of the model weights, enabling dynamic data modifications without retraining. Specifically, our approach integrates the $k$-Nearest Neighbor ($k$-NN) classifier with a vision-based foundation model, pre-trained self-supervised on natural images, enhancing interpretability and adaptability. We share open-source implementations of a previously unpublished baseline method as well as our performance-improving contributions. Quantitative experiments confirm improved classification across established benchmark datasets and the method's applicability to distinct medical image classification tasks. Additionally, we assess the method's robustness in continual learning and data removal scenarios. The approach exhibits great promise for bridging the gap between foundation models' performance and challenges tied to data privacy. The source code is available at https://github.com/TobArc/privacy-aware-image-classification-with-kNN.
LangXAI is a framework that integrates Explainable Artificial Intelligence (XAI) with advanced vision models to generate textual explanations for visual recognition tasks. Despite XAI advancements, an understanding gap persists for end-users with limited domain knowledge in artificial intelligence and computer vision. LangXAI addresses this by furnishing text-based explanations for classification, object detection, and semantic segmentation model outputs to end-users. Preliminary results demonstrate LangXAI's enhanced plausibility, with high BERTScore across tasks, fostering a more transparent and reliable AI framework on vision tasks for end-users.
The attention mechanism, a cornerstone of state-of-the-art neural models, faces computational hurdles in processing long sequences due to its quadratic complexity. Consequently, research efforts in the last few years focused on finding more efficient alternatives. Among them, Hyena (Poli et al., 2023) stands out for achieving competitive results in both language modeling and image classification, while offering sub-quadratic memory and computational complexity. Building on these promising results, we propose ConfHyena, a Conformer whose encoder self-attentions are replaced with an adaptation of Hyena for speech processing, where the long input sequences cause high computational costs. Through experiments in automatic speech recognition (for English) and translation (from English into 8 target languages), we show that our best ConfHyena model significantly reduces the training time by 27%, at the cost of minimal quality degradation (~1%), which, in most cases, is not statistically significant.
Weakly supervised object detection (WSup-OD) increases the usefulness and interpretability of image classification algorithms without requiring additional supervision. The successes of multiple instance learning in this task for natural images, however, do not translate well to medical images due to the very different characteristics of their objects (i.e. pathologies). In this work, we propose Weakly Supervised ROI Proposal Networks (WSRPN), a new method for generating bounding box proposals on the fly using a specialized region of interest-attention (ROI-attention) module. WSRPN integrates well with classic backbone-head classification algorithms and is end-to-end trainable with only image-label supervision. We experimentally demonstrate that our new method outperforms existing methods in the challenging task of disease localization in chest X-ray images. Code: https://github.com/philip-mueller/wsrpn
Multimodal Large Language Models (MLLMs) have demonstrated notable capabilities in general visual understanding and reasoning tasks. However, their deployment is hindered by substantial computational costs in both training and inference, limiting accessibility to the broader research and user communities. A straightforward solution is to leverage smaller pre-trained vision and language models, which inevitably cause significant performance drops. In this paper, we demonstrate the possibility of training a smaller but better MLLM with high-quality training data. Specifically, we introduce Bunny, a family of lightweight MLLMs with flexible vision and language backbones for efficient multimodal learning from selected training data. Experiments show that our Bunny-4B/8B outperforms the state-of-the-art large MLLMs on multiple benchmarks. We expect that this work can provide the community with a clean and flexible open-source tool for further research and development. The code, models, and data can be found in https://github.com/BAAI-DCAI/Bunny.
Vision Transformer (ViT) self-attention mechanism is characterized by feature collapse in deeper layers, resulting in the vanishing of low-level visual features. However, such features can be helpful to accurately represent and identify elements within an image and increase the accuracy and robustness of vision-based recognition systems. Following this rationale, we propose a novel residual attention learning method for improving ViT-based architectures, increasing their visual feature diversity and model robustness. In this way, the proposed network can capture and preserve significant low-level features, providing more details about the elements within the scene being analyzed. The effectiveness and robustness of the presented method are evaluated on five image classification benchmarks, including ImageNet1k, CIFAR10, CIFAR100, Oxford Flowers-102, and Oxford-IIIT Pet, achieving improved performances. Additionally, experiments on the COCO2017 dataset show that the devised approach discovers and incorporates semantic and spatial relationships for object detection and instance segmentation when implemented into spatial-aware transformer models.
We present a novel bi-directional Transformer architecture (BiXT) which scales linearly with input size in terms of computational cost and memory consumption, but does not suffer the drop in performance or limitation to only one input modality seen with other efficient Transformer-based approaches. BiXT is inspired by the Perceiver architectures but replaces iterative attention with an efficient bi-directional cross-attention module in which input tokens and latent variables attend to each other simultaneously, leveraging a naturally emerging attention-symmetry between the two. This approach unlocks a key bottleneck experienced by Perceiver-like architectures and enables the processing and interpretation of both semantics ('what') and location ('where') to develop alongside each other over multiple layers -- allowing its direct application to dense and instance-based tasks alike. By combining efficiency with the generality and performance of a full Transformer architecture, BiXT can process longer sequences like point clouds, text or images at higher feature resolutions and achieves competitive performance across a range of tasks like point cloud part segmentation, semantic image segmentation, image classification, hierarchical sequence modeling and document retrieval. Our experiments demonstrate that BiXT models outperform larger competitors by leveraging longer sequences more efficiently on vision tasks like classification and segmentation, and perform on par with full Transformer variants on sequence modeling and document retrieval -- but require $28\%$ fewer FLOPs and are up to $8.4\times$ faster.
Whole-slide image (WSI) classification is a challenging task because 1) patches from WSI lack annotation, and 2) WSI possesses unnecessary variability, e.g., stain protocol. Recently, Multiple-Instance Learning (MIL) has made significant progress, allowing for classification based on slide-level, rather than patch-level, annotations. However, existing MIL methods ignore that all patches from normal slides are normal. Using this free annotation, we introduce a semi-supervision signal to de-bias the inter-slide variability and to capture the common factors of variation within normal patches. Because our method is orthogonal to the MIL algorithm, we evaluate our method on top of the recently proposed MIL algorithms and also compare the performance with other semi-supervised approaches. We evaluate our method on two public WSI datasets including Camelyon-16 and TCGA lung cancer and demonstrate that our approach significantly improves the predictive performance of existing MIL algorithms and outperforms other semi-supervised algorithms. We release our code at https://github.com/AITRICS/pathology_mil.
Oftentimes, machine learning applications using neural networks involve solving discrete optimization problems, such as in pruning, parameter-isolation-based continual learning and training of binary networks. Still, these discrete problems are combinatorial in nature and are also not amenable to gradient-based optimization. Additionally, classical approaches used in discrete settings do not scale well to large neural networks, forcing scientists and empiricists to rely on alternative methods. Among these, two main distinct sources of top-down information can be used to lead the model to good solutions: (1) extrapolating gradient information from points outside of the solution set (2) comparing evaluations between members of a subset of the valid solutions. We take continuation path (CP) methods to represent using purely the former and Monte Carlo (MC) methods to represent the latter, while also noting that some hybrid methods combine the two. The main goal of this work is to compare both approaches. For that purpose, we first overview the two classes while also discussing some of their drawbacks analytically. Then, on the experimental section, we compare their performance, starting with smaller microworld experiments, which allow more fine-grained control of problem variables, and gradually moving towards larger problems, including neural network regression and neural network pruning for image classification, where we additionally compare against magnitude-based pruning.
Informally, a model is calibrated if its predictions are correct with a probability that matches the confidence of the prediction. By far the most common method in the literature for measuring calibration is the expected calibration error (ECE). Recent work, however, has pointed out drawbacks of ECE, such as the fact that it is discontinuous in the space of predictors. In this work, we ask: how fundamental are these issues, and what are their impacts on existing results? Towards this end, we completely characterize the discontinuities of ECE with respect to general probability measures on Polish spaces. We then use the nature of these discontinuities to motivate a novel continuous, easily estimated miscalibration metric, which we term Logit-Smoothed ECE (LS-ECE). By comparing the ECE and LS-ECE of pre-trained image classification models, we show in initial experiments that binned ECE closely tracks LS-ECE, indicating that the theoretical pathologies of ECE may be avoidable in practice.
Purpose: We apply federated learning to train an OCT image classifier simulating a realistic scenario with multiple clients and statistical heterogeneous data distribution where data in the clients lack samples of some categories entirely. Methods: We investigate the effectiveness of FedAvg and FedProx to train an OCT image classification model in a decentralized fashion, addressing privacy concerns associated with centralizing data. We partitioned a publicly available OCT dataset across multiple clients under IID and Non-IID settings and conducted local training on the subsets for each client. We evaluated two federated learning methods, FedAvg and FedProx for these settings. Results: Our experiments on the dataset suggest that under IID settings, both methods perform on par with training on a central data pool. However, the performance of both algorithms declines as we increase the statistical heterogeneity across the client data, while FedProx consistently performs better than FedAvg in the increased heterogeneity settings. Conclusion: Despite the effectiveness of federated learning in the utilization of private data across multiple medical institutions, the large number of clients and heterogeneous distribution of labels deteriorate the performance of both algorithms. Notably, FedProx appears to be more robust to the increased heterogeneity.
Satellite missions and Earth Observation (EO) systems represent fundamental assets for environmental monitoring and the timely identification of catastrophic events, long-term monitoring of both natural resources and human-made assets, such as vegetation, water bodies, forests as well as buildings. Different EO missions enables the collection of information on several spectral bandwidths, such as MODIS, Sentinel-1 and Sentinel-2. Thus, given the recent advances of machine learning, computer vision and the availability of labeled data, researchers demonstrated the feasibility and the precision of land-use monitoring systems and remote sensing image classification through the use of deep neural networks. Such systems may help domain experts and governments in constant environmental monitoring, enabling timely intervention in case of catastrophic events (e.g., forest wildfire in a remote area). Despite the recent advances in the field of computer vision, many works limit their analysis on Convolutional Neural Networks (CNNs) and, more recently, to vision transformers (ViTs). Given the recent successes of Graph Neural Networks (GNNs) on non-graph data, such as time-series and images, we investigate the performances of a recent Vision GNN architecture (ViG) applied to the task of land cover classification. The experimental results show that ViG achieves state-of-the-art performances in multiclass and multilabel classification contexts, surpassing both ViT and ResNet on large-scale benchmarks.
Convolutional neural networks (CNNs) for image processing tend to focus on localized texture patterns, commonly referred to as texture bias. While most of the previous works in the literature focus on the task of image classification, we go beyond this and study the texture bias of CNNs in semantic segmentation. In this work, we propose to train CNNs on pre-processed images with less texture to reduce the texture bias. Therein, the challenge is to suppress image texture while preserving shape information. To this end, we utilize edge enhancing diffusion (EED), an anisotropic image diffusion method initially introduced for image compression, to create texture reduced duplicates of existing datasets. Extensive numerical studies are performed with both CNNs and vision transformer models trained on original data and EED-processed data from the Cityscapes dataset and the CARLA driving simulator. We observe strong texture-dependence of CNNs and moderate texture-dependence of transformers. Training CNNs on EED-processed images enables the models to become completely ignorant with respect to texture, demonstrating resilience with respect to texture re-introduction to any degree. Additionally we analyze the performance reduction in depth on a level of connected components in the semantic segmentation and study the influence of EED pre-processing on domain generalization as well as adversarial robustness.
Exponential Moving Average (EMA) is a widely used weight averaging (WA) regularization to learn flat optima for better generalizations without extra cost in deep neural network (DNN) optimization. Despite achieving better flatness, existing WA methods might fall into worse final performances or require extra test-time computations. This work unveils the full potential of EMA with a single line of modification, i.e., switching the EMA parameters to the original model after each epoch, dubbed as Switch EMA (SEMA). From both theoretical and empirical aspects, we demonstrate that SEMA can help DNNs to reach generalization optima that better trade-off between flatness and sharpness. To verify the effectiveness of SEMA, we conduct comparison experiments with discriminative, generative, and regression tasks on vision and language datasets, including image classification, self-supervised learning, object detection and segmentation, image generation, video prediction, attribute regression, and language modeling. Comprehensive results with popular optimizers and networks show that SEMA is a free lunch for DNN training by improving performances and boosting convergence speeds.
Recent research has seen many behavioral comparisons between humans and deep neural networks (DNNs) in the domain of image classification. Often, comparison studies focus on the end-result of the learning process by measuring and comparing the similarities in the representations of object categories once they have been formed. However, the process of how these representations emerge -- that is, the behavioral changes and intermediate stages observed during the acquisition -- is less often directly and empirically compared. Here we report a detailed investigation of the learning dynamics in human observers and various classic and state-of-the-art DNNs. We develop a constrained supervised learning environment to align learning-relevant conditions such as starting point, input modality, available input data and the feedback provided. Across the whole learning process we evaluate and compare how well learned representations can be generalized to previously unseen test data. Comparisons across the entire learning process indicate that DNNs demonstrate a level of data efficiency comparable to human learners, challenging some prevailing assumptions in the field. However, our results also reveal representational differences: while DNNs' learning is characterized by a pronounced generalisation lag, humans appear to immediately acquire generalizable representations without a preliminary phase of learning training set-specific information that is only later transferred to novel data.