DocR1: Evidence Page-Guided GRPO for Multi-Page Document Understanding

Understanding multi-page documents poses a significant challenge for multimodal large language models (MLLMs), as it requires fine-grained visual comprehension and multi-hop reasoning across pages. While prior work has explored reinforcement learning (RL) for enhancing advanced reasoning in MLLMs, its application to multi-page document understanding remains underexplored. In this paper, we introduce DocR1, an MLLM trained with a novel RL framework, Evidence Page-Guided GRPO (EviGRPO). EviGRPO incorporates an evidence-aware reward mechanism that promotes a coarse-to-fine reasoning strategy, guiding the model to first retrieve relevant pages before generating answers. This training paradigm enables us to build high-quality models with limited supervision. To support this, we design a two-stage annotation pipeline and a curriculum learning strategy, based on which we construct two datasets: EviBench, a high-quality training set with 4.8k examples, and ArxivFullQA, an evaluation benchmark with 8.6k QA pairs based on scientific papers. Extensive experiments across a wide range of benchmarks demonstrate that DocR1 achieves state-of-the-art performance on multi-page tasks, while consistently maintaining strong results on single-page benchmarks.

Kimi K2: Open Agentic Intelligence

We introduce Kimi K2, a Mixture-of-Experts (MoE) large language model with 32 billion activated parameters and 1 trillion total parameters. We propose the MuonClip optimizer, which improves upon Muon with a novel QK-clip technique to address training instability while enjoying the advanced token efficiency of Muon. Based on MuonClip, K2 was pre-trained on 15.5 trillion tokens with zero loss spike. During post-training, K2 undergoes a multi-stage post-training process, highlighted by a large-scale agentic data synthesis pipeline and a joint reinforcement learning (RL) stage, where the model improves its capabilities through interactions with real and synthetic environments. Kimi K2 achieves state-of-the-art performance among open-source non-thinking models, with strengths in agentic capabilities. Notably, K2 obtains 66.1 on Tau2-Bench, 76.5 on ACEBench (En), 65.8 on SWE-Bench Verified, and 47.3 on SWE-Bench Multilingual -- surpassing most open and closed-sourced baselines in non-thinking settings. It also exhibits strong capabilities in coding, mathematics, and reasoning tasks, with a score of 53.7 on LiveCodeBench v6, 49.5 on AIME 2025, 75.1 on GPQA-Diamond, and 27.1 on OJBench, all without extended thinking. These results position Kimi K2 as one of the most capable open-source large language models to date, particularly in software engineering and agentic tasks. We release our base and post-trained model checkpoints to facilitate future research and applications of agentic intelligence.

CodeBrain: Bridging Decoupled Tokenizer and Multi-Scale Architecture for EEG Foundation Model

Electroencephalography (EEG) provides real-time insights into brain activity and is widely used in neuroscience. However, variations in channel configurations, sequence lengths, and task objectives limit the transferability of traditional task-specific models. Although recent EEG foundation models (EFMs) aim to learn generalizable representations, they struggle with limited heterogeneous representation capacity and inefficiency in capturing multi-scale brain dependencies. To address these challenges, we propose CodeBrain, an efficient EFM structurally aligned with brain organization, trained in two stages. (1) We introduce a TFDual-Tokenizer that independently tokenizes heterogeneous temporal and frequency components, enabling a quadratic expansion of the discrete representation space. This also offers a degree of interpretability through cross-domain token analysis. (2) We propose the EEGSSM, which combines a structured global convolution architecture and a sliding window attention mechanism to jointly model sparse long-range and local dependencies. Unlike fully connected Transformer models, EEGSSM better reflects the brain's small-world topology and efficiently captures EEG's inherent multi-scale structure. EEGSSM is trained with a masked self-supervised learning objective to predict token indices obtained in TFDual-Tokenizer. Comprehensive experiments on 10 public EEG datasets demonstrate the generalizability of CodeBrain with linear probing. By offering biologically informed and interpretable EEG modeling, CodeBrain lays the foundation for future neuroscience research. Both code and pretraining weights will be released in the future version.

Impromptu VLA: Open Weights and Open Data for Driving Vision-Language-Action Models

Vision-Language-Action (VLA) models for autonomous driving show promise but falter in unstructured corner case scenarios, largely due to a scarcity of targeted benchmarks. To address this, we introduce Impromptu VLA. Our core contribution is the Impromptu VLA Dataset: over 80,000 meticulously curated video clips, distilled from over 2M source clips sourced from 8 open-source large-scale datasets. This dataset is built upon our novel taxonomy of four challenging unstructured categories and features rich, planning-oriented question-answering annotations and action trajectories. Crucially, experiments demonstrate that VLAs trained with our dataset achieve substantial performance gains on established benchmarks--improving closed-loop NeuroNCAP scores and collision rates, and reaching near state-of-the-art L2 accuracy in open-loop nuScenes trajectory prediction. Furthermore, our Q&A suite serves as an effective diagnostic, revealing clear VLM improvements in perception, prediction, and planning. Our code, data and models are available at https://github.com/ahydchh/Impromptu-VLA.

X-ARES: A Comprehensive Framework for Assessing Audio Encoder Performance

We introduces X-ARES (eXtensive Audio Representation and Evaluation Suite), a novel open-source benchmark designed to systematically assess audio encoder performance across diverse domains. By encompassing tasks spanning speech, environmental sounds, and music, X-ARES provides two evaluation approaches for evaluating audio representations: linear fine-tuning and unparameterized evaluation. The framework includes 22 distinct tasks that cover essential aspects of audio processing, from speech recognition and emotion detection to sound event classification and music genre identification. Our extensive evaluation of state-of-the-art audio encoders reveals significant performance variations across different tasks and domains, highlighting the complexity of general audio representation learning.

FB-4D: Spatial-Temporal Coherent Dynamic 3D Content Generation with Feature Banks

With the rapid advancements in diffusion models and 3D generation techniques, dynamic 3D content generation has become a crucial research area. However, achieving high-fidelity 4D (dynamic 3D) generation with strong spatial-temporal consistency remains a challenging task. Inspired by recent findings that pretrained diffusion features capture rich correspondences, we propose FB-4D, a novel 4D generation framework that integrates a Feature Bank mechanism to enhance both spatial and temporal consistency in generated frames. In FB-4D, we store features extracted from previous frames and fuse them into the process of generating subsequent frames, ensuring consistent characteristics across both time and multiple views. To ensure a compact representation, the Feature Bank is updated by a proposed dynamic merging mechanism. Leveraging this Feature Bank, we demonstrate for the first time that generating additional reference sequences through multiple autoregressive iterations can continuously improve generation performance. Experimental results show that FB-4D significantly outperforms existing methods in terms of rendering quality, spatial-temporal consistency, and robustness. It surpasses all multi-view generation tuning-free approaches by a large margin and achieves performance on par with training-based methods.

PHGNN: A Novel Prompted Hypergraph Neural Network to Diagnose Alzheimer's Disease

The accurate diagnosis of Alzheimer's disease (AD) and prognosis of mild cognitive impairment (MCI) conversion are crucial for early intervention. However, existing multimodal methods face several challenges, from the heterogeneity of input data, to underexplored modality interactions, missing data due to patient dropouts, and limited data caused by the time-consuming and costly data collection process. In this paper, we propose a novel Prompted Hypergraph Neural Network (PHGNN) framework that addresses these limitations by integrating hypergraph based learning with prompt learning. Hypergraphs capture higher-order relationships between different modalities, while our prompt learning approach for hypergraphs, adapted from NLP, enables efficient training with limited data. Our model is validated through extensive experiments on the ADNI dataset, outperforming SOTA methods in both AD diagnosis and the prediction of MCI conversion.

Col-OLHTR: A Novel Framework for Multimodal Online Handwritten Text Recognition

Online Handwritten Text Recognition (OLHTR) has gained considerable attention for its diverse range of applications. Current approaches usually treat OLHTR as a sequence recognition task, employing either a single trajectory or image encoder, or multi-stream encoders, combined with a CTC or attention-based recognition decoder. However, these approaches face several drawbacks: 1) single encoders typically focus on either local trajectories or visual regions, lacking the ability to dynamically capture relevant global features in challenging cases; 2) multi-stream encoders, while more comprehensive, suffer from complex structures and increased inference costs. To tackle this, we propose a Collaborative learning-based OLHTR framework, called Col-OLHTR, that learns multimodal features during training while maintaining a single-stream inference process. Col-OLHTR consists of a trajectory encoder, a Point-to-Spatial Alignment (P2SA) module, and an attention-based decoder. The P2SA module is designed to learn image-level spatial features through trajectory-encoded features and 2D rotary position embeddings. During training, an additional image-stream encoder-decoder is collaboratively trained to provide supervision for P2SA features. At inference, the extra streams are discarded, and only the P2SA module is used and merged before the decoder, simplifying the process while preserving high performance. Extensive experimental results on several OLHTR benchmarks demonstrate the state-of-the-art (SOTA) performance, proving the effectiveness and robustness of our design.

SelectiveFinetuning: Enhancing Transfer Learning in Sleep Staging through Selective Domain Alignment

In practical sleep stage classification, a key challenge is the variability of EEG data across different subjects and environments. Differences in physiology, age, health status, and recording conditions can lead to domain shifts between data. These domain shifts often result in decreased model accuracy and reliability, particularly when the model is applied to new data with characteristics different from those it was originally trained on, which is a typical manifestation of negative transfer. To address this, we propose SelectiveFinetuning in this paper. Our method utilizes a pretrained Multi Resolution Convolutional Neural Network (MRCNN) to extract EEG features, capturing the distinctive characteristics of different sleep stages. To mitigate the effect of domain shifts, we introduce a domain aligning mechanism that employs Earth Mover Distance (EMD) to evaluate and select source domain data closely matching the target domain. By finetuning the model with selective source data, our SelectiveFinetuning enhances the model's performance on target domain that exhibits domain shifts compared to the data used for training. Experimental results show that our method outperforms existing baselines, offering greater robustness and adaptability in practical scenarios where data distributions are often unpredictable.

Towards Wireless-Native Big AI Model: Insights into Its Ambitions, Peculiarities and Methodologies

Researches on leveraging big artificial intelligence model (BAIM) technology to drive the intelligent evolution of wireless networks are emerging. However, since the breakthrough in generalization brought about by BAIM techniques mainly occurs in natural language processing, there is still a lack of a clear technical roadmap on how to efficiently apply BAIM techniques to wireless systems with many additional peculiarities. To this end, this paper first reviews recent research works on BAIM for wireless and assesses the current research situation. Then, this paper analyzes and compares the differences between language intelligence and wireless intelligence on multiple levels, including scientific foundations, core usages, and technical details. It highlights the necessity and scientific significance of developing BAIM technology in a wireless-native way, as well as new issues that need to be considered in specific technical implementation. Finally, by synthesizing the evolutionary laws of language models with the particularities of wireless system, this paper provides several instructive methodologies for how to develop wireless-native BAIM.