Parameter-efficient fine-tuning for continual learning (PEFT-CL) has shown promise in adapting pre-trained models to sequential tasks while mitigating catastrophic forgetting problem. However, understanding the mechanisms that dictate continual performance in this paradigm remains elusive. To unravel this mystery, we undertake a rigorous analysis of PEFT-CL dynamics to derive relevant metrics for continual scenarios using Neural Tangent Kernel (NTK) theory. With the aid of NTK as a mathematical analysis tool, we recast the challenge of test-time forgetting into the quantifiable generalization gaps during training, identifying three key factors that influence these gaps and the performance of PEFT-CL: training sample size, task-level feature orthogonality, and regularization. To address these challenges, we introduce NTK-CL, a novel framework that eliminates task-specific parameter storage while adaptively generating task-relevant features. Aligning with theoretical guidance, NTK-CL triples the feature representation of each sample, theoretically and empirically reducing the magnitude of both task-interplay and task-specific generalization gaps. Grounded in NTK analysis, our framework imposes an adaptive exponential moving average mechanism and constraints on task-level feature orthogonality, maintaining intra-task NTK forms while attenuating inter-task NTK forms. Ultimately, by fine-tuning optimizable parameters with appropriate regularization, NTK-CL achieves state-of-the-art performance on established PEFT-CL benchmarks. This work provides a theoretical foundation for understanding and improving PEFT-CL models, offering insights into the interplay between feature representation, task orthogonality, and generalization, contributing to the development of more efficient continual learning systems.

Generalized Category Discovery (GCD) aims to classify both base and novel images using labeled base data. However, current approaches inadequately address the intrinsic optimization of the co-occurrence matrix A¯ based on cosine similarity, failing to achieve zero base-novel regions and adequate sparsity in base and novel domains. To address these deficiencies, we propose a Non-Negative Generalized Category Discovery (NN-GCD) framework. By establishing within the Symmetric Non-negative Matrix Factorization (SNMF) framework: (i) the equivalence between ideal k-means clustering and ideal SNMF, and (ii) the equivalence between SNMF solvers and Non-negative Contrastive Learning (NCL) optimization, we reformulate both the optimization of A¯ and k-means clustering as an NCL optimization problem. Moreover, to satisfy the non-negative constraints and make a GCD model converge to a near-ideal region, we propose a GELU activation function and an NMF NCE loss. To transition A¯ from a near-ideal state to the desired A¯∗, we introduce a hybrid sparse regularization approach to impose sparsity constraints. Experimental results show NN-GCD outperforms state-of-the-art methods on GCD benchmarks, achieving an average accuracy of 66.9% on the Semantic Shift Benchmark, surpassing prior counterparts by 2.5%.

Self-Explainable Models (SEMs) rely on Prototypical Concept Learning (PCL) to enable their visual recognition processes more interpretable, but they often struggle in data-scarce settings where insufficient training samples lead to suboptimal performance. To address this limitation, we propose a Few-Shot Prototypical Concept Classification (FSPCC) framework that systematically mitigates two key challenges under low-data regimes: parametric imbalance and representation misalignment. Specifically, our approach leverages a Mixture of LoRA Experts (MoLE) for parameter-efficient adaptation, ensuring a balanced allocation of trainable parameters between the backbone and the PCL module. Meanwhile, cross-module concept guidance enforces tight alignment between the backbone's feature representations and the prototypical concept activation patterns. In addition, we incorporate a multi-level feature preservation strategy that fuses spatial and semantic cues across various layers, thereby enriching the learned representations and mitigating the challenges posed by limited data availability. Finally, to enhance interpretability and minimize concept overlap, we introduce a geometry-aware concept discrimination loss that enforces orthogonality among concepts, encouraging more disentangled and transparent decision boundaries. Experimental results on six popular benchmarks (CUB-200-2011, mini-ImageNet, CIFAR-FS, Stanford Cars, FGVC-Aircraft, and DTD) demonstrate that our approach consistently outperforms existing SEMs by a notable margin, with 4.2%-8.7% relative gains in 5-way 5-shot classification. These findings highlight the efficacy of coupling concept learning with few-shot adaptation to achieve both higher accuracy and clearer model interpretability, paving the way for more transparent visual recognition systems.

Vision Language Models (VLMs), pre-trained on large-scale image-text datasets, enable zero-shot predictions for unseen data but may underperform on specific unseen tasks. Continual learning (CL) can help VLMs effectively adapt to new data distributions without joint training, but faces challenges of catastrophic forgetting and generalization forgetting. Although significant progress has been achieved by distillation-based methods, they exhibit two severe limitations. One is the popularly adopted single-teacher paradigm fails to impart comprehensive knowledge, The other is the existing methods inadequately leverage the multimodal information in the original training dataset, instead they rely on additional data for distillation, which increases computational and storage overhead. To mitigate both limitations, by drawing on Knowledge Integration Theory (KIT), we propose a Multi-Stage Knowledge Integration network (MulKI) to emulate the human learning process in distillation methods. MulKI achieves this through four stages, including Eliciting Ideas, Adding New Ideas, Distinguishing Ideas, and Making Connections. During the four stages, we first leverage prototypes to align across modalities, eliciting cross-modal knowledge, then adding new knowledge by constructing fine-grained intra- and inter-modality relationships with prototypes. After that, knowledge from two teacher models is adaptively distinguished and re-weighted. Finally, we connect between models from intra- and inter-task, integrating preceding and new knowledge. Our method demonstrates significant improvements in maintaining zero-shot capabilities while supporting continual learning across diverse downstream tasks, showcasing its potential in adapting VLMs to evolving data distributions.

The proliferation of Few-Shot Class Incremental Learning (FSCIL) methodologies has highlighted the critical challenge of maintaining robust anti-amnesia capabilities in FSCIL learners. In this paper, we present a novel conceptualization of anti-amnesia in terms of mathematical generalization, leveraging the Neural Tangent Kernel (NTK) perspective. Our method focuses on two key aspects: ensuring optimal NTK convergence and minimizing NTK-related generalization loss, which serve as the theoretical foundation for cross-task generalization. To achieve global NTK convergence, we introduce a principled meta-learning mechanism that guides optimization within an expanded network architecture. Concurrently, to reduce the NTK-related generalization loss, we systematically optimize its constituent factors. Specifically, we initiate self-supervised pre-training on the base session to enhance NTK-related generalization potential. These self-supervised weights are then carefully refined through curricular alignment, followed by the application of dual NTK regularization tailored specifically for both convolutional and linear layers. Through the combined effects of these measures, our network acquires robust NTK properties, ensuring optimal convergence and stability of the NTK matrix and minimizing the NTK-related generalization loss, significantly enhancing its theoretical generalization. On popular FSCIL benchmark datasets, our NTK-FSCIL surpasses contemporary state-of-the-art approaches, elevating end-session accuracy by 2.9% to 9.3%.

Object detection, as one of the most important environment perception tasks for traffic safety in intelligent transportation systems, has been widely investigated recently. However, most of the researches focus on the fully supervised scenario, and inevitably lead to model failure. With the continuous development of Zero-Shot Learning (ZSL) models, Generalized Zero-Shot Detection (GZSD) has attracted great attention due to its ability of detecting unseen objects. Many researchers tend to map the detected visual features to semantic attributes and then separate seen and unseen domains during inference. But they have ignore that the generative methods generally have higher performance than these visual-semantic mapping methods, and they have been confirmed from previous GZSL methods. In order to make up for the vacancy of GZSD in the generative methods, we propose an idea of using curriculum learning to generate more precise unseen visual features. And with the excellent performance of WGAN-based method in sample synthesis, we realize the function of using semantics to generate visual features for unseen domains. In addition, we also adopt part of the idea of meta-learning to progressively correct the capability of the generator for better mitigating domain shift problem during the generation process. Through the above ideas, we can detect both seen and unseen bounding boxes and classify them accurately, by combining with the excellent detection ability of Faster-RCNN. Extensive experimental results on two popular datasets, i.e., MSCOCO and KITTI, show that our proposed method can outperform the state-of-the-art methods.

Due to the powerful feature synthesis ability, Generative Adversarial Networks (GAN) is well adapted to the Generalized Zero-Shot Learning (GZSL) task and has achieved great success. Most GAN models for GZSL usually employ random noise with normal distribution to synthesize unseen samples. However, the generated samples often have the same normal distribution as the input noise, which is unrealistic in most circumstances. Therefore, in this paper, we consider that the distribution of unseen classes should be follow that of seen classes and propose a near-real feature generative network (NereNet), which utilizes the most semantically similar seen samples to generate the noise for the unseen classes. Specifically, we first calculate the most similar seen classes for the unseen classes, and then train an encoder network to generate the corresponding noise, which is subsequently combined with the unseen classes attributes to generate unseen samples with GAN. Extensive experiments are conducted on four datasets, and the results demonstrate the effectiveness of our proposed method.