Linear-attention models that compress the entire input sequence into a fixed-size recurrent state offer an efficient alternative to Transformers, but their finite memory induces forgetfulness that harms retrieval-intensive tasks. To mitigate the issue, we explore a series of hybrid models that restore direct access to past tokens. We interleave token mixers with intermediate time and space complexity between linear and full attention, including sparse attention with token eviction, and the query-aware native sparse attention. Particularly, we propose a novel learnable token eviction approach. Combined with sliding-window attention, an end-to-end trainable lightweight CNN aggregates information from both past and future adjacent tokens to adaptively retain a limited set of critical KV-pairs per head, maintaining linear attention's constant time and space complexity. Efficient Triton kernels for the sparse attention mechanisms are provided. Empirical evaluations on retrieval-intensive benchmarks support the effectiveness of our approaches.

Urban air pollution poses significant risks to public health, environmental sustainability, and policy planning. Effective air quality management requires predictive tools that can integrate diverse datasets and communicate complex spatial and temporal pollution patterns. There is a gap in interactive tools with seamless integration of forecasting and visualization of spatial distributions of air pollutant concentrations. We present CityAQVis, an interactive machine learning ML sandbox tool designed to predict and visualize pollutant concentrations at the ground level using multi-source data, which includes satellite observations, meteorological parameters, population density, elevation, and nighttime lights. While traditional air quality visualization tools often lack forecasting capabilities, CityAQVis enables users to build and compare predictive models, visualizing the model outputs and offering insights into pollution dynamics at the ground level. The pilot implementation of the tool is tested through case studies predicting nitrogen dioxide (NO2) concentrations in metropolitan regions, highlighting its adaptability to various pollutants. Through an intuitive graphical user interface (GUI), the user can perform comparative visualizations of the spatial distribution of surface-level pollutant concentration in two different urban scenarios. Our results highlight the potential of ML-driven visual analytics to improve situational awareness and support data-driven decision-making in air quality management.

Facial Landmark Detection (FLD) in thermal imagery is critical for applications in challenging lighting conditions, but it is hampered by the lack of rich visual cues. Conventional cross-modal solutions, like feature fusion or image translation from RGB data, are often computationally expensive or introduce structural artifacts, limiting their practical deployment. To address this, we propose Multi-Level Cross-Modal Knowledge Distillation (MLCM-KD), a novel framework that decouples high-fidelity RGB-to-thermal knowledge transfer from model compression to create both accurate and efficient thermal FLD models. A central challenge during knowledge transfer is the profound modality gap between RGB and thermal data, where traditional unidirectional distillation fails to enforce semantic consistency across disparate feature spaces. To overcome this, we introduce Dual-Injected Knowledge Distillation (DIKD), a bidirectional mechanism designed specifically for this task. DIKD establishes a connection between modalities: it not only guides the thermal student with rich RGB features but also validates the student's learned representations by feeding them back into the frozen teacher's prediction head. This closed-loop supervision forces the student to learn modality-invariant features that are semantically aligned with the teacher, ensuring a robust and profound knowledge transfer. Experiments show that our approach sets a new state-of-the-art on public thermal FLD benchmarks, notably outperforming previous methods while drastically reducing computational overhead.

Large language models (LLMs) have been increasingly adopted across the scientific research pipeline, assisting tasks from ideation to writing (Zhang et al., 2025; Si et al., 2024). However, generating correct and executable research code remains a difficult problem, not only because it requires long-range reasoning and robust verification (Padigela et al., 2025; Starace et al., 2025; Zhu et al., 2025), but also because the input contexts in research settings are often complex, indirect, and noisy. Research papers describe methods through high-level narratives, mathematical formulas, and domain-specific conventions, with many implementation details left implicit. As a result, translating these fragmented and underspecified descriptions into functional code remains a fundamental challenge for current LLMs (Li et al., 2025b;a). Existing benchmarks for research code generation (Zheng et al., 2023; Sun et al., 2023; Toledo et al., 2025; Hua et al., 2025) primarily evaluate models in a non-interactive setting, where they are expected to produce correct code in a single response. This design neglects the crucial role of human feedback in realistic workflows: on the one hand, users often cannot fully specify their requirements in one shot.

Building general-purpose graphical user interface (GUI) agents has become increasingly promising with the progress in vision language models. However, developing effective mobile GUI agents with reinforcement learning (RL) remains challenging due to the heavy-tailed distribution of task difficulty and the inefficiency of large-scale environment sampling. We present an online agentic reinforcement learning framework MobileRL to enhance GUI agents in mobile environments. Its core component is the Difficulty-ADAptive GRPO (ADAGRPO) algorithm. In ADAGRPO, we design difficulty-adaptive positive replay and failure curriculum filtering to adapt the model to different task difficulties. We introduce the shortest-path reward adjustment strategy to reshape rewards concerning the task length in multi-turn agentic tasks. Those strategies jointly stabilize RL training, improve sample efficiency, and generate strong performance across diverse mobile apps and tasks. We apply MOBILERL to two open models (Qwen2.5-VL-7B-Instruct and GLM-4.1V-9B-Base). The resultant MOBILERL-9B model achieves state-of-the-art results in terms of success rates on both AndroidWorld (80.2%) and AndroidLab (53.6%). The MOBILERL framework is open-sourced at: https://github.com/THUDM/MobileRL.

Neural collapse, i.e., the emergence of highly symmetric, class-wise clustered representations, is frequently observed in deep networks and is often assumed to reflect or enable generalization. In parallel, flatness of the loss landscape has been theoretically and empirically linked to generalization. Yet, the causal role of either phenomenon remains unclear: Are they prerequisites for generalization, or merely by-products of training dynamics? We disentangle these questions using grokking, a training regime in which memorization precedes generalization, allowing us to temporally separate generalization from training dynamics and we find that while both neural collapse and relative flatness emerge near the onset of generalization, only flatness consistently predicts it. Models encouraged to collapse or prevented from collapsing generalize equally well, whereas models regularized away from flat solutions exhibit delayed generalization, resembling grokking, even in architectures and datasets where it does not typically occur. Furthermore, we show theoretically that neural collapse leads to relative flatness under classical assumptions, explaining their empirical co-occurrence. Our results support the view that relative flatness is a potentially necessary and more fundamental property for generalization, and demonstrate how grokking can serve as a powerful probe for isolating its geometric underpinnings.

Entity resolution plays a significant role in enterprise systems where data integrity must be rigorously maintained. Traditional methods often struggle with handling noisy data or semantic understanding, while modern methods suffer from computational costs or the excessive need for parallel computation. In this study, we introduce a scalable hybrid framework, which is designed to address several important problems, including scalability, noise robustness, and reliable results. We utilized a pre-trained language model to encode each structured data into corresponding semantic embedding vectors. Subsequently, after retrieving a semantically relevant subset of candidates, we apply a syntactic verification stage using fuzzy string matching techniques to refine classification on the unlabeled data. This approach was applied to a real-world entity resolution task, which exposed a linkage between a central user management database and numerous shared hosting server records. Compared to other methods, this approach exhibits an outstanding performance in terms of both processing time and robustness, making it a reliable solution for a server-side product. Crucially, this efficiency does not compromise results, as the system maintains a high retrieval recall of approximately 0.97. The scalability of the framework makes it deployable on standard CPU-based infrastructure, offering a practical and effective solution for enterprise-level data integrity auditing.

Autoregressive (AR) language models generate text one token at a time, which limits their inference speed. Diffusion-based language models offer a promising alternative, as they can decode multiple tokens in parallel. However, we identify a key bottleneck in current diffusion LMs: the long decoding-window problem, where tokens generated far from the input context often become irrelevant or repetitive. Previous solutions like semi-autoregressive address this issue by splitting windows into blocks (sacrificing bidirectionality), but we find that this also leads to time-interval expansion problem, sacrificing the speed. Therefore, semi-AR eliminates the main advantages of diffusion models. To overcome this, we propose Convolutional decoding (Conv), a normalization-based method that narrows the decoding window without hard segmentation, leading to better fluency and flexibility. Additionally, we introduce Rejecting Rule-based Fine-Tuning (R2FT), a post-hoc training scheme that better aligns tokens at positions far from context. Our methods achieve state-of-the-art results on open-ended generation benchmarks (e.g., AlpacaEval) among diffusion LM baselines, with significantly lower step size than previous works, demonstrating both speed and quality improvements.

Cardiovascular disease (CVD) is the leading cause of death globally, accounting for 32% of all deaths according to The W orld Health Organization (WHO) statistics in 2019 [1]. With its non-invasive nature and ability to reflect the heart's electrical activity, the electrocardiogram is a key diagnostic tool in clinical practice [2], [3]. However, traditional ECG analysis is mainly based on human experts prone to errors and delays.

This study investigates whether second-order geometric cues - planar curvature magnitude, curvature sign, and gradient orientation - are sufficient on their own to drive a multilayer perceptron (MLP) classifier for handwritten character recognition (HCR), offering an alternative to convolutional neural networks (CNNs). Using these three handcrafted feature maps as inputs, our curvature-orientation MLP achieves 97 percent accuracy on MNIST digits and 89 percent on EMNIST letters. These results underscore the discriminative power of curvature-based representations for handwritten character images and demonstrate that the advantages of deep learning can be realized even with interpretable, hand-engineered features.