The computational imaging technique of Fourier Ptychographic Microscopy (FPM) enables high-resolution imaging with a wide field of view and can serve as an extremely valuable tool, e.g. in the classification of cells in medical applications. However, reconstructing a high-resolution image from tens or even hundreds of measurements is computationally expensive, particularly for a wide field of view. Therefore, in this paper, we investigate the idea of classifying the image content in the FPM measurements directly without performing a reconstruction step first. We show that Convolutional Neural Networks (CNN) can extract meaningful information from measurement sequences, significantly outperforming the classification on a single band-limited image (up to 12 %) while being significantly more efficient than a reconstruction of a high-resolution image. Furthermore, we demonstrate that a learned multiplexing of several raw measurements allows maintaining the classification accuracy while reducing the amount of data (and consequently also the acquisition time) significantly.
Single-photon LiDAR achieves high-precision 3D imaging in extreme environments through quantum-level photon detection technology. Current research primarily focuses on reconstructing 3D scenes from sparse photon events, whereas the semantic interpretation of single-photon images remains underexplored, due to high annotation costs and inefficient labeling strategies. This paper presents the first active learning framework for single-photon image classification. The core contribution is an imaging condition-aware sampling strategy that integrates synthetic augmentation to model variability across imaging conditions. By identifying samples where the model is both uncertain and sensitive to these conditions, the proposed method selectively annotates only the most informative examples. Experiments on both synthetic and real-world datasets show that our approach outperforms all baselines and achieves high classification accuracy with significantly fewer labeled samples. Specifically, our approach achieves 97% accuracy on synthetic single-photon data using only 1.5% labeled samples. On real-world data, we maintain 90.63% accuracy with just 8% labeled samples, which is 4.51% higher than the best-performing baseline. This illustrates that active learning enables the same level of classification performance on single-photon images as on classical images, opening doors to large-scale integration of single-photon data in real-world applications.
Federated learning (FL) enables collaborative model training across distributed clients while preserving data locality. Although FedAvg pioneered synchronous rounds for global model averaging, slower devices can delay collective progress. Asynchronous FL (e.g., FedAsync) addresses stragglers by continuously integrating client updates, yet naive implementations risk client drift due to non-IID data and stale contributions. Some Blockchain-based FL approaches (e.g., BRAIN) employ robust weighting or scoring of updates to resist malicious or misaligned proposals. However, performance drops can still persist under severe data heterogeneity or high staleness, and synchronization overhead has emerged as a new concern due to its aggregator-free architectures. We introduce Fast-and-Reliable AI Network, FRAIN, a new asynchronous FL method that mitigates these limitations by incorporating two key ideas. First, our FastSync strategy eliminates the need to replay past model versions, enabling newcomers and infrequent participants to efficiently approximate the global model. Second, we adopt spherical linear interpolation (SLERP) when merging parameters, preserving models' directions and alleviating destructive interference from divergent local training. Experiments with a CNN image-classification model and a Transformer-based language model demonstrate that FRAIN achieves more stable and robust convergence than FedAvg, FedAsync, and BRAIN, especially under harsh environments: non-IID data distributions, networks that experience delays and require frequent re-synchronization, and the presence of malicious nodes.
First-order optimization methods remain the standard for training deep neural networks (DNNs). Optimizers like Adam incorporate limited curvature information by preconditioning the stochastic gradient with a diagonal matrix. Despite the widespread adoption of first-order methods, second-order optimization algorithms often exhibit superior convergence compared to methods like Adam and SGD. However, their practicality in training DNNs is still limited by a significantly higher per-iteration computational cost compared to first-order methods. In this thesis, we present AdaFisher, a novel adaptive second-order optimizer that leverages a diagonal block-Kronecker approximation of the Fisher information matrix to adaptively precondition gradients. AdaFisher aims to bridge the gap between the improved convergence and generalization of second-order methods and the computational efficiency needed for training DNNs. Despite the traditionally slower speed of second-order optimizers, AdaFisher is effective for tasks such as image classification and language modeling, exhibiting remarkable stability and robustness during hyperparameter tuning. We demonstrate that AdaFisher outperforms state-of-the-art optimizers in both accuracy and convergence speed. The code is available from https://github.com/AtlasAnalyticsLab/AdaFisher.
Semantic broadcast communications (Semantic BC) for image transmission have achieved significant performance gains for single-task scenarios. Nevertheless, extending these methods to multi-task scenarios remains challenging, as different tasks typically require distinct objective functions, leading to potential conflicts within the shared encoder. In this paper, we propose a tri-level reinforcement learning (RL)-based multi-task Semantic BC framework, termed SemanticBC-TriRL, which effectively resolves such conflicts and enables the simultaneous support of multiple downstream tasks at the receiver side, including image classification and content reconstruction tasks. Specifically, the proposed framework employs a bottom-up tri-level alternating learning strategy, formulated as a constrained multi-objective optimization problem. At the first level, task-specific decoders are locally optimized using supervised learning. At the second level, the shared encoder is updated via proximal policy optimization (PPO), guided by task-oriented rewards. At the third level, a multi-gradient aggregation-based task weighting module adaptively adjusts task priorities and steers the encoder optimization. Through this hierarchical learning process, the encoder and decoders are alternately trained, and the three levels are cohesively integrated via constrained learning objective. Besides, the convergence of SemanticBC-TriRL is also theoretically established. Extensive simulation results demonstrate the superior performance of the proposed framework across diverse channel conditions, particularly in low SNR regimes, and confirm its scalability with increasing numbers of receivers.
Knowledge distillation (KD) compresses the network capacity by transferring knowledge from a large (teacher) network to a smaller one (student). It has been mainstream that the teacher directly transfers knowledge to the student with its original distribution, which can possibly lead to incorrect predictions. In this article, we propose a logit-based distillation via swapped logit processing, namely Swapped Logit Distillation (SLD). SLD is proposed under two assumptions: (1) the wrong prediction occurs when the prediction label confidence is not the maximum; (2) the "natural" limit of probability remains uncertain as the best value addition to the target cannot be determined. To address these issues, we propose a swapped logit processing scheme. Through this approach, we find that the swap method can be effectively extended to teacher and student outputs, transforming into two teachers. We further introduce loss scheduling to boost the performance of two teachers' alignment. Extensive experiments on image classification tasks demonstrate that SLD consistently performs best among previous state-of-the-art methods.
The Computer_Aided Diagnosis (CAD) systems facilitate accurate diagnosis of diseases. The development of CADs by leveraging third generation neural network, namely, Spiking Neural Network (SNN), is essential to utilize of the benefits of SNNs, such as their event_driven processing, parallelism, low power consumption, and the ability to process sparse temporal_spatial information. However, Deep SNN as a deep learning model faces challenges with unreliability. To deal with unreliability challenges due to inability to quantify the uncertainty of the predictions, we proposed a deep Bayesian Convolutional Spiking Neural Network based_CADs with uncertainty_aware module. In this study, the Monte Carlo Dropout method as Bayesian approximation is used as an uncertainty quantification method. This method was applied to several medical image classification tasks. Our experimental results demonstrate that our proposed model is accurate and reliable and will be a proper alternative to conventional deep learning for medical image classification.
This paper presents a comprehensive empirical analysis of conformal prediction methods on a challenging aerial image dataset featuring diverse events in unconstrained environments. Conformal prediction is a powerful post-hoc technique that takes the output of any classifier and transforms it into a set of likely labels, providing a statistical guarantee on the coverage of the true label. Unlike evaluations on standard benchmarks, our study addresses the complexities of data-scarce and highly variable real-world settings. We investigate the effectiveness of leveraging pretrained models (MobileNet, DenseNet, and ResNet), fine-tuned with limited labeled data, to generate informative prediction sets. To further evaluate the impact of calibration, we consider two parallel pipelines (with and without temperature scaling) and assess performance using two key metrics: empirical coverage and average prediction set size. This setup allows us to systematically examine how calibration choices influence the trade-off between reliability and efficiency. Our findings demonstrate that even with relatively small labeled samples and simple nonconformity scores, conformal prediction can yield valuable uncertainty estimates for complex tasks. Moreover, our analysis reveals that while temperature scaling is often employed for calibration, it does not consistently lead to smaller prediction sets, underscoring the importance of careful consideration in its application. Furthermore, our results highlight the significant potential of model compression techniques within the conformal prediction pipeline for deployment in resource-constrained environments. Based on our observations, we advocate for future research to delve into the impact of noisy or ambiguous labels on conformal prediction performance and to explore effective model reduction strategies.
We derive a differential equation that governs the evolution of the generalization gap when a deep network is trained by gradient descent. This differential equation is controlled by two quantities, a contraction factor that brings together trajectories corresponding to slightly different datasets, and a perturbation factor that accounts for them training on different datasets. We analyze this differential equation to compute an ``effective Gram matrix'' that characterizes the generalization gap after training in terms of the alignment between this Gram matrix and a certain initial ``residual''. Empirical evaluations on image classification datasets indicate that this analysis can predict the test loss accurately. Further, at any point during training, the residual predominantly lies in the subspace of the effective Gram matrix with the smallest eigenvalues. This indicates that the training process is benign, i.e., it does not lead to significant deterioration of the generalization gap (which is zero at initialization). The alignment between the effective Gram matrix and the residual is different for different datasets and architectures. The match/mismatch of the data and the architecture is primarily responsible for good/bad generalization.
Federated Learning enables collaborative training of machine learning models on decentralized data. This scheme, however, is vulnerable to adversarial attacks, when some of the clients submit corrupted model updates. In real-world scenarios, the total number of compromised clients is typically unknown, with the extent of attacks potentially varying over time. To address these challenges, we propose an adaptive approach for robust aggregation of model updates based on Bayesian inference. The mean update is defined by the maximum of the likelihood marginalized over probabilities of each client to be `honest'. As a result, the method shares the simplicity of the classical average estimators (e.g., sample mean or geometric median), being independent of the number of compromised clients. At the same time, it is as effective against attacks as methods specifically tailored to Federated Learning, such as Krum. We compare our approach with other aggregation schemes in federated setting on three benchmark image classification data sets. The proposed method consistently achieves state-of-the-art performance across various attack types with static and varying number of malicious clients.
This paper introduces an enhanced Genetic Algorithm technique called Dual-Individual Genetic Algorithm (Dual-Individual GA), which optimizes neural networks for binary image classification tasks, such as cat vs. non-cat classification. The proposed method employs only two individuals for crossover, represented by two parameter sets: Leader and Follower. The Leader focuses on exploitation, representing the primary optimal solution at even-indexed positions (0, 2, 4, ...), while the Follower promotes exploration by preserving diversity and avoiding premature convergence, operating at odd-indexed positions (1, 3, 5, ...). Leader and Follower are modeled as two phases or roles. The key contributions of this work are threefold: (1) a self-adaptive layer dimension mechanism that eliminates the need for manual tuning of layer architectures; (2) generates two parameter sets, leader and follower parameter sets, with 10 layer architecture configurations (5 for each set), ranked by Pareto dominance and cost. post-optimization; and (3) demonstrated superior performance compared to traditional gradient-based methods. Experimental results show that the Dual-Individual GA achieves 99.04% training accuracy and 80% testing accuracy (cost = 0.034) on a three-layer network with architecture [12288, 17, 4, 1], outperforming a gradient-based approach that achieves 98% training accuracy and 80% testing accuracy (cost = 0.092) on a four-layer network with architecture [12288, 20, 7, 5, 1]. These findings highlight the efficiency and effectiveness of the proposed method in optimizing neural networks.
Curriculum learning strategies have been proven to be effective in various applications and have gained significant interest in the field of machine learning. It has the ability to improve the final model's performance and accelerate the training process. However, in the medical imaging domain, data irregularities can make the recognition task more challenging and usually result in misclassification between the different classes in the dataset. Class-decomposition approaches have shown promising results in solving such a problem by learning the boundaries within the classes of the data set. In this paper, we present a novel convolutional neural network (CNN) training method based on the curriculum learning strategy and the class decomposition approach, which we call CLOG-CD, to improve the performance of medical image classification. We evaluated our method on four different imbalanced medical image datasets, such as Chest X-ray (CXR), brain tumour, digital knee X-ray, and histopathology colorectal cancer (CRC). CLOG-CD utilises the learnt weights from the decomposition granularity of the classes, and the training is accomplished from descending to ascending order (i.e., anti-curriculum technique). We also investigated the classification performance of our proposed method based on different acceleration factors and pace function curricula. We used two pre-trained networks, ResNet-50 and DenseNet-121, as the backbone for CLOG-CD. The results with ResNet-50 show that CLOG-CD has the ability to improve classification performance with an accuracy of 96.08% for the CXR dataset, 96.91% for the brain tumour dataset, 79.76% for the digital knee X-ray, and 99.17% for the CRC dataset, compared to other training strategies. In addition, with DenseNet-121, CLOG-CD has achieved 94.86%, 94.63%, 76.19%, and 99.45% for CXR, brain tumour, digital knee X-ray, and CRC datasets, respectively
Visual defect detection in industrial glass manufacturing remains a critical challenge due to the low frequency of defective products, leading to imbalanced datasets that limit the performance of deep learning models and computer vision systems. This paper presents a novel approach using Denoising Diffusion Probabilistic Models (DDPMs) to generate synthetic defective glass product images for data augmentation, effectively addressing class imbalance issues in manufacturing quality control and automated visual inspection. The methodology significantly enhances image classification performance of standard CNN architectures (ResNet50V2, EfficientNetB0, and MobileNetV2) in detecting anomalies by increasing the minority class representation. Experimental results demonstrate substantial improvements in key machine learning metrics, particularly in recall for defective samples across all tested deep neural network architectures while maintaining perfect precision. The most dramatic improvement was observed in ResNet50V2's overall classification accuracy, which increased from 78 percent to 93 percent when trained with the augmented data. This work provides a scalable, cost-effective approach to enhancing automated defect detection in glass manufacturing that can potentially be extended to other industrial quality assurance systems and industries with similar class imbalance challenges.
Modern edge devices, such as cameras, drones, and Internet-of-Things nodes, rely on deep learning to enable a wide range of intelligent applications, including object recognition, environment perception, and autonomous navigation. However, deploying deep learning models directly on the often resource-constrained edge devices demands significant memory footprints and computational power for real-time inference using traditional digital computing architectures. In this paper, we present WISE, a novel computing architecture for wireless edge networks designed to overcome energy constraints in deep learning inference. WISE achieves this goal through two key innovations: disaggregated model access via wireless broadcasting and in-physics computation of general complex-valued matrix-vector multiplications directly at radio frequency. Using a software-defined radio platform with wirelessly broadcast model weights over the air, we demonstrate that WISE achieves 95.7% image classification accuracy with ultra-low operation power of 6.0 fJ/MAC per client, corresponding to a computation efficiency of 165.8 TOPS/W. This approach enables energy-efficient deep learning inference on wirelessly connected edge devices, achieving more than two orders of magnitude improvement in efficiency compared to traditional digital computing.
Recent advancement in large-scale Artificial Intelligence (AI) models offering multimodal services have become foundational in AI systems, making them prime targets for model theft. Existing methods select Out-of-Distribution (OoD) data as backdoor watermarks and retrain the original model for copyright protection. However, existing methods are susceptible to malicious detection and forgery by adversaries, resulting in watermark evasion. In this work, we propose Model-\underline{ag}nostic Black-box Backdoor W\underline{ate}rmarking Framework (AGATE) to address stealthiness and robustness challenges in multimodal model copyright protection. Specifically, we propose an adversarial trigger generation method to generate stealthy adversarial triggers from ordinary dataset, providing visual fidelity while inducing semantic shifts. To alleviate the issue of anomaly detection among model outputs, we propose a post-transform module to correct the model output by narrowing the distance between adversarial trigger image embedding and text embedding. Subsequently, a two-phase watermark verification is proposed to judge whether the current model infringes by comparing the two results with and without the transform module. Consequently, we consistently outperform state-of-the-art methods across five datasets in the downstream tasks of multimodal image-text retrieval and image classification. Additionally, we validated the robustness of AGATE under two adversarial attack scenarios.
Current self-supervised algorithms mostly rely on transformations such as data augmentation and masking to learn visual representations. This is achieved by inducing invariance or equivariance with respect to these transformations after encoding two views of an image. This dominant two-view paradigm can limit the flexibility of learned representations for downstream adaptation by creating performance trade-offs between invariance-related tasks such as image classification and more fine-grained equivariance-related tasks. In this work, we introduce \emph{seq-JEPA}, a world modeling paradigm based on joint-embedding predictive architecture that leverages architectural inductive biases to resolve this trade-off. Without requiring an additional equivariance predictor or loss term, seq-JEPA simultaneously learns two architecturally segregated representations: one equivariant to the specified transformations and another invariant to them and suited for tasks such as classification. To do so, our model processes a short sequence of different views (observations) of an input image. Each encoded view is concatenated with embeddings corresponding to the relative transformation (action) producing the next observation in the sequence. A transformer encoder outputs an aggregate representation of this sequence, which is subsequently conditioned on the action leading to the next observation to predict its representation. Empirically, seq-JEPA achieves strong performance on equivariant benchmarks and image classification without sacrificing one for the other. Additionally, our framework excels at tasks that inherently require aggregating a sequence of observations, such as path integration across actions and predictive learning across eye movements.
This paper presents a comprehensive evaluation of lightweight deep learning models for image classification, emphasizing their suitability for deployment in resource-constrained environments such as low-memory devices. Five state-of-the-art architectures - MobileNetV3 Small, ResNet18, SqueezeNet, EfficientNetV2-S, and ShuffleNetV2 - are benchmarked across three diverse datasets: CIFAR-10, CIFAR-100, and Tiny ImageNet. The models are assessed using four key performance metrics: classification accuracy, inference time, floating-point operations (FLOPs), and model size. Additionally, we investigate the impact of hyperparameter tuning, data augmentation, and training paradigms by comparing pretrained models with scratch-trained counterparts, focusing on MobileNetV3 Small. Our findings reveal that transfer learning significantly enhances model accuracy and computational efficiency, particularly for complex datasets like Tiny ImageNet. EfficientNetV2 consistently achieves the highest accuracy, while MobileNetV3 offers the best balance between accuracy and efficiency, and SqueezeNet excels in inference speed and compactness. This study highlights critical trade-offs between accuracy and efficiency, offering actionable insights for deploying lightweight models in real-world applications where computational resources are limited. By addressing these challenges, this research contributes to optimizing deep learning systems for edge computing and mobile platforms.
The customizability of RISC-V makes it an attractive choice for accelerating deep neural networks (DNNs). It can be achieved through instruction set extensions and corresponding custom functional units. Yet, efficiently exploiting these opportunities requires a hardware/software co-design approach in which the DNN model, software, and hardware are designed together. In this paper, we propose novel RISC-V extensions for accelerating DNN models containing semi-structured and unstructured sparsity. While the idea of accelerating structured and unstructured pruning is not new, our novel design offers various advantages over other designs. To exploit semi-structured sparsity, we take advantage of the fine-grained (bit-level) configurability of FPGAs and suggest reserving a few bits in a block of DNN weights to encode the information about sparsity in the succeeding blocks. The proposed custom functional unit utilizes this information to skip computations. To exploit unstructured sparsity, we propose a variable cycle sequential multiply-and-accumulate unit that performs only as many multiplications as the non-zero weights. Our implementation of unstructured and semi-structured pruning accelerators can provide speedups of up to a factor of 3 and 4, respectively. We then propose a combined design that can accelerate both types of sparsities, providing speedups of up to a factor of 5. Our designs consume a small amount of additional FPGA resources such that the resulting co-designs enable the acceleration of DNNs even on small FPGAs. We benchmark our designs on standard TinyML applications such as keyword spotting, image classification, and person detection.
Given the severe challenges confronting the global growth security of economic crops, precise identification and prevention of plant diseases has emerged as a critical issue in artificial intelligence-enabled agricultural technology. To address the technical challenges in plant disease recognition, including small-sample learning, leaf occlusion, illumination variations, and high inter-class similarity, this study innovatively proposes a Dynamic Dual-Stream Fusion Network (DS_FusionNet). The network integrates a dual-backbone architecture, deformable dynamic fusion modules, and bidirectional knowledge distillation strategy, significantly enhancing recognition accuracy. Experimental results demonstrate that DS_FusionNet achieves classification accuracies exceeding 90% using only 10% of the PlantDisease and CIFAR-10 datasets, while maintaining 85% accuracy on the complex PlantWild dataset, exhibiting exceptional generalization capabilities. This research not only provides novel technical insights for fine-grained image classification but also establishes a robust foundation for precise identification and management of agricultural diseases.
Contrastive Language-Image Pre-training (CLIP) has shown impressive zero-shot transfer capabilities, but its potential for specific downstream tasks is not fully utilized. To further enhance CLIP’s few-shot capability for specific datasets, some subsequent works have been proposed, such as methods based on lightweight adapters and prompt learning. However, since CLIP is pretrained on a diverse collection of image and text pairs sourced from the internet, it is difficult to sufficiently tune models to specific datasets using only lightweight adaptions. In this paper, we argue that largely modifying the internal representations within CLIP’s encoders can yield better results on downstream datasets. In this work, we introduce Ta-Adapter, a method that equips both the visual and textual encoders of CLIP with task-specific prompts. These prompts are generated using a collaborative prompt learning approach, which allows the encoders to produce representations that are better aligned with specific downstream datasets. Then, we initialize an adapter module using the optimized features generated by the task-aware visual encoder for further feature alignment, and this module can also be further fine-tuned. Our extensive experiments on image classification datasets show that compared to the state-of-theart few-shot methods Tip-Adapter-F and MaPLe, our model exhibits good performance and obtains an average absolute gain of 2.04% and 1.62% on 11 different image recognition datasets, respectively. In conclusion, this work presents a unique and effective approach to unlocking the full potential of CLIP’s few-shot learning capabilities.