Multimodal large language models (MLLMs) are increasingly deployed as the core reasoning engine for web-facing systems, powering GUI agents and front-end automation that must interpret page structure, select actionable widgets, and execute multi-step interactions reliably. However, existing benchmarks largely emphasize visual perception or UI code generation, showing insufficient evaluation on the reasoning, robustness and safety capability required for end-to-end web applications. To bridge the gap, we introduce a comprehensive web understanding benchmark, named WebRRSBench, that jointly evaluates Reasoning, Robustness, and Safety across eight tasks, such as position relationship reasoning, color robustness, and safety critical detection, etc. The benchmark is constructed from 729 websites and contains 3799 QA pairs that probe multi-step inference over page structure, text, widgets, and safety-critical interactions. To ensure reliable measurement, we adopt standardized prompts, a protocolized and deterministic evaluation pipeline, and multi-stage quality control combining automatic checks with targeted human verification. We evaluate 11 MLLMs on WebRRSBench. The results reveal significant gaps: models still struggle with compositional and cross-element reasoning over realistic layouts, show limited robustness when facing perturbations in user interfaces and content such as layout rearrangements or visual style shifts, and are rather conservative in recognizing and avoiding safety critical or irreversible actions. Our code and appendix are available at https://github.com/JunliangLiu-repo/WebRRSBench.

Diffusion models, known for their generative ability, have recently been adapted to time series analysis. Most pioneering works rely on the standard isotropic diffusion, treating each time step and the entire frequency spectrum identically. However, it may not be suitable for time series, which often have more informative low-frequency components. We empirically found that direct application of standard diffusion to time series may cause gradient contradiction during training, due to the rapid decrease of low-frequency information in the diffusion process. To this end, we proposed a novel time series diffusion model, MA-TSD, which utilizes the moving average, a natural low-frequency filter, as the forward transition. Its backward process is accelerable like DDIM and can be further considered a time series super-resolution. Our experiments on various datasets demonstrated MA-TSD’s superior performance in time series forecasting and super-resolution tasks.

Most machine learning algorithms typically assume that the data distribution of the training and test sets are consistent, but this assumption often fails to hold in practical applications. Domain generalization aims to train a model using only available source data so that the model can generalize to unseen domains. Data augmentation is an important technique in domain generalization, but there are few comprehensive reviews investigating and summarizing its use in domain generalization. This study provides a comprehensive literature review of data augmentation methods in domain generalization for the first time. First, we formalize the definition of domain generalization and analyze the role of data augmentation in domain generalization. Second, we propose a new taxonomy that categorizes methods into three classes based on the augmentation objectives: domain-level, image-level, and feature-level augmentation. Third, we compare the experimental results of some data augmentation methods on three popular domain generalization datasets and discuss the characteristics and advantages of the current best methods. Fourth, we analyze the shortcomings of each category, propose the suggestions for improvements, and summarize the challenges and the future directions of data augmentation for achieving cross-domain generalization from both theoretical and practical perspectives.

Backdoor attacks can exploit vulnerabilities in the training process of Deep Neural Networks (DNNs), introducing hidden malicious functionality that can be activated by a specific input pattern. Existing defenses typically rely on the assumption of a significant difference between poisoned and clean samples. However, subtle differences in dynamic and low poisoning ratio attacks can conceal poisoning features, thereby evading defenses. In this work, we propose a novel backdoor defense approach calledArtemis, which utilizes distribution shifts to eliminate the discrepancy between poisoned and benign samples in the feature space. Additionally,Artemislearns from domain-invariant features after the shift. To further enhance the purification of backdoors in DNNs, we incorporate soft knowledge distillation intoArtemisto guide the alignment of features between the source dataset domain and the generated distribution-shift dataset domain. We extensively compare our proposed method with 5 state-of-the-art (SOTA) defensive techniques under 9 SOTA attacks on 4 datasets to demonstrate its effectiveness and robustness. Our results show that our method,Artemis, can successfully purify dynamic and low poisoning ratio backdoor attacks and outperform existing defenses by a significant margin. We release the code athttps://github.com/xmyun/Artemis.

Protein-protein interactions (PPIs) are crucial for various cellular activities and disease development, and modulating PPIs using small molecule inhibitors (PPIIs) has gradually become a promising therapeutic strategy. Recently, researchers have proposed several machine learning methods to screen PPIIs, but most of the works focused on unimodal representations of molecules or combining multimodal features in a naive splicing manner. Meanwhile, current research progress is being slowed by the lack of large-scale PPII datasets. To address these issues, we propose MCLPPII, a unified multimodal contrastive learning framework for PPII prediction. MCLPPII extracts comprehensive molecular information from four modalities and effectively combines them through an adaptive feature fusion method. Furthermore, we propose a three-stage training strategy to enhance the PPII prediction capability of MCLPPII by leveraging self-supervised pre-training on a large unlabeled dataset. We evaluate MCLPPII on nine PPI targets and two downstream tasks, including the PPI inhibitor identification task and potency prediction task. Experimental results show that MCLPPII achieves competitive performance. The source code and datasets are freely available at https://github.com/1zzt/MCLPPII.

Time series analysis is widely used in many fields such as power energy, economics, and transportation, including different tasks such as forecasting, anomaly detection, classification, etc. Missing values are widely observed in these tasks, and often leading to unpredictable negative effects on existing methods, hindering their further application. In response to this situation, existing time series imputation methods mainly focus on restoring sequences based on their data characteristics, while ignoring the performance of the restored sequences in downstream tasks. Considering different requirements of downstream tasks (e.g., forecasting), this paper proposes an efficient downstream task-oriented time series imputation evaluation approach. By combining time series imputation with neural network models used for downstream tasks, the gain of different imputation strategies on downstream tasks is estimated without retraining, and the most favorable imputation value for downstream tasks is given by combining different imputation strategies according to the estimated gain.