MobileDreamer: Generative Sketch World Model for GUI Agent

Mobile GUI agents have shown strong potential in real-world automation and practical applications. However, most existing agents remain reactive, making decisions mainly from current screen, which limits their performance on long-horizon tasks. Building a world model from repeated interactions enables forecasting action outcomes and supports better decision making for mobile GUI agents. This is challenging because the model must predict post-action states with spatial awareness while remaining efficient enough for practical deployment. In this paper, we propose MobileDreamer, an efficient world-model-based lookahead framework to equip the GUI agents based on the future imagination provided by the world model. It consists of textual sketch world model and rollout imagination for GUI agent. Textual sketch world model forecasts post-action states through a learning process to transform digital images into key task-related sketches, and designs a novel order-invariant learning strategy to preserve the spatial information of GUI elements. The rollout imagination strategy for GUI agent optimizes the action-selection process by leveraging the prediction capability of world model. Experiments on Android World show that MobileDreamer achieves state-of-the-art performance and improves task success by 5.25%. World model evaluations further verify that our textual sketch modeling accurately forecasts key GUI elements.

Learning When to Look: A Disentangled Curriculum for Strategic Perception in Multimodal Reasoning

Multimodal Large Language Models (MLLMs) demonstrate significant potential but remain brittle in complex, long-chain visual reasoning tasks. A critical failure mode is "visual forgetting", where models progressively lose visual grounding as reasoning extends, a phenomenon aptly described as "think longer, see less". We posit this failure stems from current training paradigms prematurely entangling two distinct cognitive skills: (1) abstract logical reasoning "how-to-think") and (2) strategic visual perception ("when-to-look"). This creates a foundational cold-start deficiency -- weakening abstract reasoning -- and a strategic perception deficit, as models lack a policy for when to perceive. In this paper, we propose a novel curriculum-based framework to disentangle these skills. First, we introduce a disentangled Supervised Fine-Tuning (SFT) curriculum that builds a robust abstract reasoning backbone on text-only data before anchoring it to vision with a novel Perception-Grounded Chain-of-Thought (PG-CoT) paradigm. Second, we resolve the strategic perception deficit by formulating timing as a reinforcement learning problem. We design a Pivotal Perception Reward that teaches the model when to look by coupling perceptual actions to linguistic markers of cognitive uncertainty (e.g., "wait", "verify"), thereby learning an autonomous grounding policy. Our contributions include the formalization of these two deficiencies and the development of a principled, two-stage framework to address them, transforming the model from a heuristic-driven observer to a strategic, grounded reasoner. \textbf{Code}: \url{https://github.com/gaozilve-max/learning-when-to-look}.

Context Cascade Compression: Exploring the Upper Limits of Text Compression

Million-level token inputs in long-context tasks pose significant computational and memory challenges for Large Language Models (LLMs). Recently, DeepSeek-OCR conducted research into the feasibility of Contexts Optical Compression and achieved preliminary results. Inspired by this, we introduce Context Cascade Compression C3 to explore the upper limits of text compression. Our method cascades two LLMs of different sizes to handle the compression and decoding tasks. Specifically, a small LLM, acting as the first stage, performs text compression by condensing a long context into a set of latent tokens (e.g., 32 or 64 in length), achieving a high ratio of text tokens to latent tokens. A large LLM, as the second stage, then executes the decoding task on this compressed context. Experiments show that at a 20x compression ratio (where the number of text tokens is 20 times the number of latent tokens), our model achieves 98% decoding accuracy, compared to approximately 60% for DeepSeek-OCR. When we further increase the compression ratio to 40x, the accuracy is maintained at around 93%. This indicates that in the domain of context compression, C3 Compression demonstrates superior performance and feasibility over optical character compression. C3 uses a simpler, pure-text pipeline that ignores factors like layout, color, and information loss from a visual encoder. This also suggests a potential upper bound for compression ratios in future work on optical character compression, OCR, and related fields. Codes and model weights are publicly accessible at https://github.com/liufanfanlff/C3-Context-Cascade-Compression

Metis-HOME: Hybrid Optimized Mixture-of-Experts for Multimodal Reasoning

Inspired by recent advancements in LLM reasoning, the field of multimodal reasoning has seen remarkable progress, achieving significant performance gains on intricate tasks such as mathematical problem-solving. Despite this progress, current multimodal large reasoning models exhibit two key limitations. They tend to employ computationally expensive reasoning even for simple queries, leading to inefficiency. Furthermore, this focus on specialized reasoning often impairs their broader, more general understanding capabilities. In this paper, we propose Metis-HOME: a Hybrid Optimized Mixture-of-Experts framework designed to address this trade-off. Metis-HOME enables a ''Hybrid Thinking'' paradigm by structuring the original dense model into two distinct expert branches: a thinking branch tailored for complex, multi-step reasoning, and a non-thinking branch optimized for rapid, direct inference on tasks like general VQA and OCR. A lightweight, trainable router dynamically allocates queries to the most suitable expert. We instantiate Metis-HOME by adapting the Qwen2.5-VL-7B into an MoE architecture. Comprehensive evaluations reveal that our approach not only substantially enhances complex reasoning abilities but also improves the model's general capabilities, reversing the degradation trend observed in other reasoning-specialized models. Our work establishes a new paradigm for building powerful and versatile MLLMs, effectively resolving the prevalent reasoning-vs-generalization dilemma.

Metis-RISE: RL Incentivizes and SFT Enhances Multimodal Reasoning Model Learning

Recent advancements in large language models (LLMs) have witnessed a surge in the development of advanced reasoning paradigms, which are now being integrated into multimodal large language models (MLLMs). However, existing approaches often fall short: methods solely employing reinforcement learning (RL) can struggle with sample inefficiency and activating entirely absent reasoning capabilities, while conventional pipelines that initiate with a cold-start supervised fine-tuning (SFT) phase before RL may restrict the model's exploratory capacity and face suboptimal convergence. In this work, we introduce \textbf{Metis-RISE} (\textbf{R}L \textbf{I}ncentivizes and \textbf{S}FT \textbf{E}nhances) for multimodal reasoning model learning. Unlike conventional approaches, Metis-RISE distinctively omits an initial SFT stage, beginning instead with an RL phase (e.g., using a Group Relative Policy Optimization variant) to incentivize and activate the model's latent reasoning capacity. Subsequently, the targeted SFT stage addresses two key challenges identified during RL: (1) \textit{inefficient trajectory sampling} for tasks where the model possesses but inconsistently applies correct reasoning, which we tackle using self-distilled reasoning trajectories from the RL model itself; and (2) \textit{fundamental capability absence}, which we address by injecting expert-augmented knowledge for prompts where the model entirely fails. This strategic application of RL for incentivization followed by SFT for enhancement forms the core of Metis-RISE, leading to two versions of our MLLMs (7B and 72B parameters). Evaluations on the OpenCompass Multimodal Reasoning Leaderboard demonstrate that both models achieve state-of-the-art performance among similar-sized models, with the 72B version ranking fourth overall.

UniToken: Harmonizing Multimodal Understanding and Generation through Unified Visual Encoding

We introduce UniToken, an auto-regressive generation model that encodes visual inputs through a combination of discrete and continuous representations, enabling seamless integration of unified visual understanding and image generation tasks. Unlike previous approaches that rely on unilateral visual representations, our unified visual encoding framework captures both high-level semantics and low-level details, delivering multidimensional information that empowers heterogeneous tasks to selectively assimilate domain-specific knowledge based on their inherent characteristics. Through in-depth experiments, we uncover key principles for developing a unified model capable of both visual understanding and image generation. Extensive evaluations across a diverse range of prominent benchmarks demonstrate that UniToken achieves state-of-the-art performance, surpassing existing approaches. These results establish UniToken as a robust foundation for future research in this domain. The code and models are available at https://github.com/SxJyJay/UniToken.

PointHR: Exploring High-Resolution Architectures for 3D Point Cloud Segmentation

Significant progress has been made recently in point cloud segmentation utilizing an encoder-decoder framework, which initially encodes point clouds into low-resolution representations and subsequently decodes high-resolution predictions. Inspired by the success of high-resolution architectures in image dense prediction, which always maintains a high-resolution representation throughout the entire learning process, we consider it also highly important for 3D dense point cloud analysis. Therefore, in this paper, we explore high-resolution architectures for 3D point cloud segmentation. Specifically, we generalize high-resolution architectures using a unified pipeline named PointHR, which includes a knn-based sequence operator for feature extraction and a differential resampling operator to efficiently communicate different resolutions. Additionally, we propose to avoid numerous on-the-fly computations of high-resolution architectures by pre-computing the indices for both sequence and resampling operators. By doing so, we deliver highly competitive high-resolution architectures while capitalizing on the benefits of well-designed point cloud blocks without additional effort. To evaluate these architectures for dense point cloud analysis, we conduct thorough experiments using S3DIS and ScanNetV2 datasets, where the proposed PointHR outperforms recent state-of-the-art methods without any bells and whistles. The source code is available at \url{https://github.com/haibo-qiu/PointHR}.

Collect-and-Distribute Transformer for 3D Point Cloud Analysis

Although remarkable advancements have been made recently in point cloud analysis through the exploration of transformer architecture, it remains challenging to effectively learn local and global structures within point clouds. In this paper, we propose a new transformer architecture equipped with a collect-and-distribute mechanism to communicate short- and long-range contexts of point clouds, which we refer to as CDFormer. Specifically, we first utilize self-attention to capture short-range interactions within each local patch, and the updated local features are then collected into a set of proxy reference points from which we can extract long-range contexts. Afterward, we distribute the learned long-range contexts back to local points via cross-attention. To address the position clues for short- and long-range contexts, we also introduce context-aware position encoding to facilitate position-aware communications between points. We perform experiments on four popular point cloud datasets, namely ModelNet40, ScanObjectNN, S3DIS, and ShapeNetPart, for classification and segmentation. Results show the effectiveness of the proposed CDFormer, delivering several new state-of-the-art performances on point cloud classification and segmentation tasks. The code is available at \url{https://github.com/haibo-qiu/CDFormer}.

GFNet: Geometric Flow Network for 3D Point Cloud Semantic Segmentation

Point cloud semantic segmentation from projected views, such as range-view (RV) and bird's-eye-view (BEV), has been intensively investigated. Different views capture different information of point clouds and thus are complementary to each other. However, recent projection-based methods for point cloud semantic segmentation usually utilize a vanilla late fusion strategy for the predictions of different views, failing to explore the complementary information from a geometric perspective during the representation learning. In this paper, we introduce a geometric flow network (GFNet) to explore the geometric correspondence between different views in an align-before-fuse manner. Specifically, we devise a novel geometric flow module (GFM) to bidirectionally align and propagate the complementary information across different views according to geometric relationships under the end-to-end learning scheme. We perform extensive experiments on two widely used benchmark datasets, SemanticKITTI and nuScenes, to demonstrate the effectiveness of our GFNet for project-based point cloud semantic segmentation. Concretely, GFNet not only significantly boosts the performance of each individual view but also achieves state-of-the-art results over all existing projection-based models. Code is available at \url{https://github.com/haibo-qiu/GFNet}.

End2End Occluded Face Recognition by Masking Corrupted Features

With the recent advancement of deep convolutional neural networks, significant progress has been made in general face recognition. However, the state-of-the-art general face recognition models do not generalize well to occluded face images, which are exactly the common cases in real-world scenarios. The potential reasons are the absences of large-scale occluded face data for training and specific designs for tackling corrupted features brought by occlusions. This paper presents a novel face recognition method that is robust to occlusions based on a single end-to-end deep neural network. Our approach, named FROM (Face Recognition with Occlusion Masks), learns to discover the corrupted features from the deep convolutional neural networks, and clean them by the dynamically learned masks. In addition, we construct massive occluded face images to train FROM effectively and efficiently. FROM is simple yet powerful compared to the existing methods that either rely on external detectors to discover the occlusions or employ shallow models which are less discriminative. Experimental results on the LFW, Megaface challenge 1, RMF2, AR dataset and other simulated occluded/masked datasets confirm that FROM dramatically improves the accuracy under occlusions, and generalizes well on general face recognition.