Test-time adaptation (TTA) has proven effective in mitigating performance drops under single-domain distribution shifts by updating model parameters during inference. However, real-world deployments often involve mixed distribution shifts---where test samples are affected by diverse and potentially conflicting domain factors---posing significant challenges even for state-of-the-art TTA methods. A key limitation in existing approaches is their reliance on a unified adaptation path, which fails to account for the fact that optimal gradient directions can vary significantly across different domains. Moreover, current benchmarks focus only on synthetic or homogeneous shifts, failing to capture the complexity of real-world heterogeneous mixed distribution shifts. To address this, we propose MoETTA, a novel entropy-based TTA framework that integrates the Mixture-of-Experts (MoE) architecture. Rather than enforcing a single parameter update rule for all test samples, MoETTA introduces a set of structurally decoupled experts, enabling specialization along diverse gradient directions. This design allows the model to better accommodate heterogeneous shifts through flexible and disentangled parameter updates. To simulate realistic deployment conditions, we introduce two new benchmarks: potpourri and potpourri+. While classical settings focus solely on synthetic corruptions (i.e., ImageNet-C), potpourri encompasses a broader range of domain shifts—including natural, artistic, and adversarial distortions—capturing more realistic deployment challenges. On top of that, potpourri+ further includes source-domain samples to evaluate robustness against catastrophic forgetting. Extensive experiments across three mixed distribution shifts settings show that MoETTA consistently outperforms strong baselines, establishing new state-of-the-art performance and highlighting the benefit of modeling multiple adaptation directions via expert-level diversity.

Recent vision-language models (VLMs) face significant challenges in test-time adaptation to novel domains. While cache-based methods show promise by leveraging historical information, they struggle with both caching unreliable feature-label pairs and indiscriminately using single-class information during querying, significantly compromising adaptation accuracy. To address these limitations, we propose COSMIC (Clique-Oriented Semantic Multi-space Integration for CLIP), a robust test-time adaptation framework that enhances adaptability through multi-granular, cross-modal semantic caching and graph-based querying mechanisms. Our framework introduces two key innovations: Dual Semantics Graph (DSG) and Clique Guided Hyper-class (CGH). The Dual Semantics Graph constructs complementary semantic spaces by incorporating textual features, coarse-grained CLIP features, and fine-grained DINOv2 features to capture rich semantic relationships. Building upon these dual graphs, the Clique Guided Hyper-class component leverages structured class relationships to enhance prediction robustness through correlated class selection. Extensive experiments demonstrate COSMIC’s superior performance across multiple benchmarks, achieving significant improvements over state-of-the-art methods: 15.81% gain on out-of-distribution tasks and 5.33% on cross-domain generation with CLIP RN-50. Code is available at github.com/hf618/COSMIC.

Deep Neural Networks suffer significant performance degradation when faced with distribution shifts between training and test data. Test-time adaptation (TTA) has emerged as a practical solution that enables models to adapt to the shifted test distribution. Currently, most existing TTA methods are designed around a single model, which incorporate limited information from a singular data distribution. In practice, pre-trained models derived from diverse source domains are readily accessible, each capturing a distinct data distribution and containing complementary information. To exploit this diversity, we propose Model Fusion-based multi-source Test-Time Adaptation (MFTTA), which constructs a target model by fusing the parameters of multiple source models. Drawing inspiration from deep model fusion, we introduce a fine-grained fusion mechanism governed by an off-policy reinforcement learning agent, which dynamically assigns fusion weights based on the current data distribution. Furthermore, we design a correlation-aware model update strategy that prioritizes the source model most relevant to the incoming test data. Extensive experiments on standard out-of-distribution benchmarks demonstrate that our method effectively integrates knowledge from multiple source models, adapts robustly to dynamic distribution shifts, and alleviates the problem of forgetting in long-term adaptation.

Test-Time Adaptation (TTA) addresses domain shifts between training and testing. However, existing methods assume a homogeneous target domain (e.g., single domain) at any given time. They fail to handle the dynamic nature of real-world data, where single-domain and multiple-domain distributions change over time. We identify that performance drops in multiple-domain scenarios are caused by batch normalization errors and gradient conflicts, which hinder adaptation. To solve these challenges, we propose Domain Diversity Adaptive Test-Time Adaptation (DATTA), the first approach to handle TTA under dynamic domain shift data streams. It is guided by a novel domain-diversity score. DATTA has three key components: a domain-diversity discriminator to recognize single- and multiple-domain patterns, domain-diversity adaptive batch normalization to combine source and test-time statistics, and domain-diversity adaptive fine-tuning to resolve gradient conflicts. Extensive experiments show that DATTA significantly outperforms state-of-the-art methods by up to 13%. Code is available at https://github.com/DYW77/DATTA.

Despite progress, deep neural networks still suffer performance declines under distribution shifts between training and test domains, leading to a substantial decrease in Quality of Experience (QoE) for applications. Existing test-time adaptation (TTA) methods are challenged by dynamic, multiple test distributions within batches. We observe that feature distributions across different domains inherently cluster into distinct groups with varying means and variances. This divergence reveals a critical limitation of previous global normalization strategies in TTA, which inevitably distort the original data characteristics. Based on this insight, we propose Feature-based Instance Neighbor Discovery (FIND), which comprises three key components: Layer-Wise Feature Disentanglement (LFD), Feature-Aware Batch Normalization (FABN) and Selective FABN (S-FABN). LFD stably captures features with similar distributions at each layer by constructing graph structures; while FABN optimally combines source statistics with test-time distribution-specific statistics for robust feature representation. Finally, S-FABN determines which layers require feature partitioning and which can remain unified, thus enhancing the efficiency of inference. Extensive experiments demonstrate that FIND significantly outperforms existing methods, achieving up to approximately 30\% accuracy improvement in dynamic scenarios while maintaining computational efficiency. The source code is available at https://github.com/Peanut-255/FIND.

Vehicular services aim to provide smart and timely services (e.g., collision warning) by taking the advantage of recent advances in artificial intelligence and employing task offloading techniques in mobile edge computing. In practice, the volume of vehicles in the Internet of Vehicles (IoV) often surges at a single location and renders the edge servers (ESs) severely overloaded, resulting in a very high delay in delivering the services. Therefore, it is of practical importance and urgency to coordinate the resources of ESs with bandwidth allocation for mitigating the occurrence of a spike traffic flow. For this challenge, existing work sought the periodicities of traffic flow by analyzing historical traffic data. However, the changes in traffic flow caused by sudden traffic conditions cannot be obtained from these periodicities. In this paper, we propose a distributed traffic flow forecasting and task offloading approach named TFFTO to optimize the execution time and power consumption in service processing. Specifically, graph attention networks (GATs) are leveraged to forecast future traffic flow in short-term and the traffic volume is utilized to estimate the number of services offloaded to the ESs in the subsequent period. With the estimate, the current load of the ESs is adjusted to ensure that the services can be handled in a timely manner. Potential game theory is adopted to determine the optimal service offloading strategy. Extensive experiments are conducted to evaluate our approach and the results validate our robust performance.

The development of the Internet of vehicles (IoV) has spawned a series of driving assistance services (e.g., collision warning), which improves the safety and intelligence of transportation. In IoV, the driving assistance services need to be met in time due to the rapid speed of vehicles. By introducing edge computing into the IoV, the insufficiency of local computation resources in vehicles is improved, providing high quality services for users. Nevertheless, the resources provided by edge servers are often limited, which fail to meet all the needs of users in IoV simultaneously. Thereby, how to minimize the tasks processing latency of users in the case of limited edge server resources is still a challenge. To handle the above problem, a task offloading scheme fuzzy-task-offloading-and-resource-allocation (F-TORA) based on Takagi–Sugeno fuzzy neural network (T–S FNN) and game theory is designed. Primarily, the cloud server predicts the future traffic flow of each section through T–S FNN and transmits the prediction results to the roadside units (RSUs). Then, the RSU adjusts the current load based on the captured future traffic flow data. After the load balancing of each RSU, the optimal task offloading strategy is determined for the users by game theory. Following, the edge server acts as an agent to allocate computing resources for the offloaded tasks by$Q$-learning algorithm. Finally, the robust performance of the proposed method is validated by comparative experiments.

With the development of the vehicular network (VN), emerging driver assistance applications are adhibited in daily life. Commonly, edge computing is adopted to satisfy the timeliness requirements of these applications, as the vehicular devices are usually insufficient in computation resources. Nevertheless, the increasing volume of service requests (SRs) are potential to overload the edge servers (ESs), thus increasing the task execution time. Besides, the randomness and the diversity of the SRs also challenge the dynamic resource allocation for the users. To deal with these challenges, a task offloading and resource allocation scheme based on game theory and reinforcement learning (RL) named TORA is proposed. Specifically, game theory is leveraged to determine the optimal task offloading strategy for improving the quality of service (QoS). Meanwhile, RL is applied to implement the dynamic resource allocation of the ES. Finally, the robust performance of the proposed method is validated by comparative experiments.

With the trend of low emissions and sustainable development, the demand for hybrid electric vehicles (HEVs) has increased rapidly. By combining a conventional internal combustion engine with one or more electric motors powered by a battery, HEVs have the advantages over traditional vehicles in better fuel economy and lower tailpipe emissions. Nevertheless, the power management strategies (PMSs) for conventional vehicles which mainly focus on the efficiency of internal combustion engine are no longer applicable due to the complex internal structure of HEVs. Hence, a large number of novel strategies appropriate for HEVs have been surveyed, but most of the researches concentrate on discussing the classifications of PMSs and comparing their cons and pros. This paper presents a comprehensive review of power management strategies adopted in HEVs aiming at specific challenges for the first time. The categories of the existing PMSs are presented based on the different algorithms, followed by a brief study of each type including the analysis of its pros and cons. Afterwards, the implementation and optimization of power management strategies aiming at proposed challenges are introduced in detail with the description of their optimization objectives and optimized results. Finally, future directions and open issues of PMSs in HEVs are discussed.