Yidan Zhang

Yidan Zhang, Ting Zhang, Dong Chen et al. · microsoft-research, pku

While generative modeling has become prevalent across numerous research fields, its integration into the realm of image retrieval remains largely unexplored and underjustified. In this paper, we present a novel methodology, reframing image retrieval as a variant of generative modeling and employing a sequence-to-sequence model. This approach is harmoniously aligned with the current trend towards unification in research, presenting a cohesive framework that allows for end-to-end differentiable searching. This, in turn, facilitates superior performance via direct optimization techniques. The development of our model, dubbed IRGen, addresses the critical technical challenge of converting an image into a concise sequence of semantic units, which is pivotal for enabling efficient and effective search. Extensive experiments demonstrate that our model achieves state-of-the-art performance on three widely-used image retrieval benchmarks as well as two million-scale datasets, yielding significant improvement compared to prior competitive retrieval methods. In addition, the notable surge in precision scores facilitated by generative modeling presents the potential to bypass the reranking phase, which is traditionally indispensable in practical retrieval workflows.

Xianggang Yu, Mutian Xu, Yidan Zhang et al.

Being data-driven is one of the most iconic properties of deep learning algorithms. The birth of ImageNet drives a remarkable trend of "learning from large-scale data" in computer vision. Pretraining on ImageNet to obtain rich universal representations has been manifested to benefit various 2D visual tasks, and becomes a standard in 2D vision. However, due to the laborious collection of real-world 3D data, there is yet no generic dataset serving as a counterpart of ImageNet in 3D vision, thus how such a dataset can impact the 3D community is unraveled. To remedy this defect, we introduce MVImgNet, a large-scale dataset of multi-view images, which is highly convenient to gain by shooting videos of real-world objects in human daily life. It contains 6.5 million frames from 219,188 videos crossing objects from 238 classes, with rich annotations of object masks, camera parameters, and point clouds. The multi-view attribute endows our dataset with 3D-aware signals, making it a soft bridge between 2D and 3D vision. We conduct pilot studies for probing the potential of MVImgNet on a variety of 3D and 2D visual tasks, including radiance field reconstruction, multi-view stereo, and view-consistent image understanding, where MVImgNet demonstrates promising performance, remaining lots of possibilities for future explorations. Besides, via dense reconstruction on MVImgNet, a 3D object point cloud dataset is derived, called MVPNet, covering 87,200 samples from 150 categories, with the class label on each point cloud. Experiments show that MVPNet can benefit the real-world 3D object classification while posing new challenges to point cloud understanding. MVImgNet and MVPNet will be publicly available, hoping to inspire the broader vision community.

Yidan Zhang, Yu Wan, Dayiheng Liu et al.

Beam search is the most widely used decoding method for neural machine translation (NMT). In practice, the top-1 candidate with the highest log-probability among the n candidates is selected as the preferred one. However, this top-1 candidate may not be the best overall translation among the n-best list. Recently, Minimum Bayes Risk (MBR) decoding has been proposed to improve the quality for NMT, which seeks for a consensus translation that is closest on average to other candidates from the n-best list. We argue that MBR still suffers from the following problems: The utility function only considers the lexical-level similarity between candidates; The expected utility considers the entire n-best list which is time-consuming and inadequate candidates in the tail list may hurt the performance; Only the relationship between candidates is considered. To solve these issues, we design a regularized MBR reranking framework (RMBR), which considers semantic-based similarity and computes the expected utility for each candidate by truncating the list. We expect the proposed framework to further consider the translation quality and model uncertainty of each candidate. Thus the proposed quality regularizer and uncertainty regularizer are incorporated into the framework. Extensive experiments on multiple translation tasks demonstrate the effectiveness of our method.

Hui Wu, Haoquan Zhai, Yuchen Li et al.

Retrieval-based multimodal document QA aims to identify and integrate relevant information from visually rich documents with complex multimodal structures. While retrieval-augmented generation (RAG) has shown strong performance in text-based QA, its extensions to multimodal documents remain underexplored and face significant limitations. Specifically, current approaches rely on query-agnostic document representations that overlook salient content and use static top-k evidence selection, which fails to adapt to the uncertain distribution of relevant information. To address these limitations, we propose the Multimodal Adaptive Retrieval-Augmented (MARA) framework, which introduces query-adaptive mechanisms to both retrieval and generation. MARA consists of two components: a Query-Aligned Region Encoder that builds multi-level document representations and reweights them based on query relevance to improve retrieval precision; and a Self-Reflective Evidence Controller that monitors evidence sufficiency during generation and adaptively incorporates content from lower-ranked sources using a sliding-window strategy. Experiments on six multimodal QA benchmarks demonstrate that MARA consistently improves retrieval relevance and answer quality over existing SOTA method.

Yidan Zhang, Yu Wan, Boyi Deng et al.

Recent advancements in large language models (LLMs) showcase varied multilingual capabilities across tasks like translation, code generation, and reasoning. Previous assessments often limited their scope to fundamental natural language processing (NLP) or isolated capability-specific tasks. To alleviate this drawback, we aim to present a comprehensive multilingual multitask benchmark. First, we introduce P-MMEval, a large-scale benchmark covering effective fundamental and capability-specialized datasets. Furthermore, P-MMEval delivers consistent language coverage across various datasets and provides parallel samples. Finally, we conduct extensive experiments on representative multilingual model series to compare performances across models and tasks, explore the relationship between multilingual performances and factors such as tasks, model sizes, languages, and prompts, and examine the effectiveness of knowledge transfer from English to other languages. The resulting insights are intended to offer valuable guidance for future research. The dataset is available at https://huggingface.co/datasets/Qwen/P-MMEval.

Xiyu Qi, Yifan Wu, Yongqiang Mao et al.

The Segment Anything Model (SAM) exhibits remarkable versatility and zero-shot learning abilities, owing largely to its extensive training data (SA-1B). Recognizing SAM's dependency on manual guidance given its category-agnostic nature, we identified unexplored potential within few-shot semantic segmentation tasks for remote sensing imagery. This research introduces a structured framework designed for the automation of few-shot semantic segmentation. It utilizes the SAM model and facilitates a more efficient generation of semantically discernible segmentation outcomes. Central to our methodology is a novel automatic prompt learning approach, leveraging prior guided masks to produce coarse pixel-wise prompts for SAM. Extensive experiments on the DLRSD datasets underline the superiority of our approach, outperforming other available few-shot methodologies.

Zhan Chen, Yidan Zhang, Xiyu Qi et al.

Height estimation has long been a pivotal topic within measurement and remote sensing disciplines, proving critical for endeavours such as 3D urban modelling, MR and autonomous driving. Traditional methods utilise stereo matching or multisensor fusion, both well-established techniques that typically necessitate multiple images from varying perspectives and adjunct sensors like SAR, leading to substantial deployment costs. Single image height estimation has emerged as an attractive alternative, boasting a larger data source variety and simpler deployment. However, current methods suffer from limitations such as fixed receptive fields, a lack of global information interaction, leading to noticeable instance-level height deviations. The inherent complexity of height prediction can result in a blurry estimation of object edge depth when using mainstream regression methods based on fixed height division. This paper presents a comprehensive solution for monocular height estimation in remote sensing, termed HeightFormer, combining multilevel interactions and image-adaptive classification-regression. It features the Multilevel Interaction Backbone (MIB) and Image-adaptive Classification-regression Height Generator (ICG). MIB supplements the fixed sample grid in CNN of the conventional backbone network with tokens of different interaction ranges. It is complemented by a pixel-, patch-, and feature map-level hierarchical interaction mechanism, designed to relay spatial geometry information across different scales and introducing a global receptive field to enhance the quality of instance-level height estimation. The ICG dynamically generates height partition for each image and reframes the traditional regression task, using a refinement from coarse to fine classification-regression that significantly mitigates the innate ill-posedness issue and drastically improves edge sharpness.

Shichu Sun, Yichen Zhang, Haolin Song et al.

Visual encoding followed by token condensing has become the standard architectural paradigm in multi-modal large language models (MLLMs). Many recent MLLMs increasingly favor global native- resolution visual encoding over slice-based methods. To investigate this trend, we systematically compare their behavior on vision-language understanding and attention patterns, revealing that global encoding enhances overall capability but at the expense of greater computational overhead. To address this issue, we present LLaVA-UHD v3, an MLLM centered upon our proposed Progressive Visual Compression (PVC) method, which can be seamlessly integrated into standard Vision Transformer (ViT) to enable efficient native-resolution encoding. The PVC approach consists of two key modules: (i) refined patch embedding, which supports flexible patch-size scaling for fine-grained visual model- ing, (ii) windowed token compression, hierarchically deployed across ViT layers to progressively aggregate local token representations. Jointly modulated by these two modules, a widely pretrained ViT can be reconfigured into an efficient architecture while largely preserving generality. Evaluated across extensive benchmarks, the transformed ViT, termed ViT-UHD, demonstrates competitive performance with MoonViT while reducing TTFT (time-to-first-token) by 2.4x, when developed within an identical MLLM architecture. Building upon ViT-UHD, LLaVA-UHD v3 also achieves competitive performance to Qwen2-VL, while further reducing TTFT by 1.9x. We will release all code and checkpoints to support future research on efficient MLLMs.

Boyi Deng, Yu Wan, Yidan Zhang et al.

The mechanisms behind multilingual capabilities in Large Language Models (LLMs) have been examined using neuron-based or internal-activation-based methods. However, these methods often face challenges such as superposition and layer-wise activation variance, which limit their reliability. Sparse Autoencoders (SAEs) offer a more nuanced analysis by decomposing the activations of LLMs into a sparse linear combination of SAE features. We introduce a novel metric to assess the monolinguality of features obtained from SAEs, discovering that some features are strongly related to specific languages. Additionally, we show that ablating these SAE features only significantly reduces abilities in one language of LLMs, leaving others almost unaffected. Interestingly, we find some languages have multiple synergistic SAE features, and ablating them together yields greater improvement than ablating individually. Moreover, we leverage these SAE-derived language-specific features to enhance steering vectors, achieving control over the language generated by LLMs. The code is publicly available at https://github.com/Aatrox103/multilingual-llm-features.

Zhan Chen, Zile Guo, Enze Zhu et al.

Video prediction is plagued by a fundamental trilemma: achieving high-resolution and perceptual quality typically comes at the cost of real-time speed, hindering its use in latency-critical applications. This challenge is most acute for autonomous UAVs in dense urban environments, where foreseeing events from high-resolution imagery is non-negotiable for safety. Existing methods, reliant on iterative generation (diffusion, autoregressive models) or quadratic-complexity attention, fail to meet these stringent demands on edge hardware. To break this long-standing trade-off, we introduce RAPTOR, a video prediction architecture that achieves real-time, high-resolution performance. RAPTOR's single-pass design avoids the error accumulation and latency of iterative approaches. Its core innovation is Efficient Video Attention (EVA), a novel translator module that factorizes spatiotemporal modeling. Instead of processing flattened spacetime tokens with $O((ST)^2)$ or $O(ST)$ complexity, EVA alternates operations along the spatial (S) and temporal (T) axes. This factorization reduces the time complexity to $O(S + T)$ and memory complexity to $O(max(S, T))$, enabling global context modeling at $512^2$ resolution and beyond, operating directly on dense feature maps with a patch-free design. Complementing this architecture is a 3-stage training curriculum that progressively refines predictions from coarse structure to sharp, temporally coherent details. Experiments show RAPTOR is the first predictor to exceed 30 FPS on a Jetson AGX Orin for $512^2$ video, setting a new state-of-the-art on UAVid, KTH, and a custom high-resolution dataset in PSNR, SSIM, and LPIPS. Critically, RAPTOR boosts the mission success rate in a real-world UAV navigation task by 18%, paving the way for safer and more anticipatory embodied agents.

Peirong Zhang, Yidan Zhang, Luxiao Xu et al.

Recent advances in multimodal large language models(MLLMs) have led to remarkable progress in visual grounding, enabling fine-grained cross-modal alignment between textual queries and image regions. However, transferring such capabilities to remote sensing imagery remains challenging, as targets are often extremely small within kilometer-scale scenes, and queries typically involve intricate geospatial relations such as relative positions, spatial hierarchies, or contextual dependencies across distant objects. To address these challenges, we propose GeoViS, a Geospatially Rewarded Visual Search framework that reformulates remote sensing visual grounding as a progressive search-and-reasoning process. Rather than directly predicting the target location in a single step, GeoViS actively explores the global image through a tree-structured sequence of visual cues, integrating multimodal perception, spatial reasoning, and reward-guided exploration to refine geospatial hypotheses iteratively. This design enables the model to detect subtle small-scale targets while maintaining holistic scene awareness. Extensive experiments on five remote sensing grounding benchmarks demonstrate that GeoViS achieves precise geospatial understanding and consistently surpasses existing methods across key visual grounding metrics, highlighting its strong cross-domain generalization and interpretability.

Yichen Zhang, Da Peng, Zonghao Guo et al.

A recent cutting-edge topic in multimodal modeling is to unify visual comprehension and generation within a single model. However, the two tasks demand mismatched decoding regimes and visual representations, making it non-trivial to jointly optimize within a shared feature space. In this work, we present Cheers, a unified multimodal model that decouples patch-level details from semantic representations, thereby stabilizing semantics for multimodal understanding and improving fidelity for image generation via gated detail residuals. Cheers includes three key components: (i) a unified vision tokenizer that encodes and compresses image latent states into semantic tokens for efficient LLM conditioning, (ii) an LLM-based Transformer that unifies autoregressive decoding for text generation and diffusion decoding for image generation, and (iii) a cascaded flow matching head that decodes visual semantics first and then injects semantically gated detail residuals from the vision tokenizer to refine high-frequency content. Experiments on popular benchmarks demonstrate that Cheers matches or surpasses advanced UMMs in both visual understanding and generation. Cheers also achieves 4x token compression, enabling more efficient high-resolution image encoding and generation. Notably, Cheers outperforms the Tar-1.5B on the popular benchmarks GenEval and MMBench, while requiring only 20% of the training cost, indicating effective and efficient (i.e., 4x token compression) unified multimodal modeling. We will release all code and data for future research.

Zheng-Hui Huang, Zhixiang Wang, Jiaming Tan et al.

Scaling generative inverse and forward rendering to real-world scenarios is bottlenecked by the limited realism and temporal coherence of existing synthetic datasets. To bridge this persistent domain gap, we introduce a large-scale, dynamic dataset curated from visually complex AAA games. Using a novel dual-screen stitched capture method, we extracted 4M continuous frames (720p/30 FPS) of synchronized RGB and five G-buffer channels across diverse scenes, visual effects, and environments, including adverse weather and motion-blur variants. This dataset uniquely advances bidirectional rendering: enabling robust in-the-wild geometry and material decomposition, and facilitating high-fidelity G-buffer-guided video generation. Furthermore, to evaluate the real-world performance of inverse rendering without ground truth, we propose a novel VLM-based assessment protocol measuring semantic, spatial, and temporal consistency. Experiments demonstrate that inverse renderers fine-tuned on our data achieve superior cross-dataset generalization and controllable generation, while our VLM evaluation strongly correlates with human judgment. Combined with our toolkit, our forward renderer enables users to edit styles of AAA games from G-buffers using text prompts.

Ziling Huang, Yidan Zhang, Shin'ichi Satoh

Large-scale visual search engines are expected to solve a dual problem at once: (i) locate every image that truly contains the object described by a sentence and (ii) identify the object's bounding box or exact pixels within each hit. Existing techniques address only one side of this challenge. Visual grounding yields tight boxes and masks but rests on the unrealistic assumption that the object is present in every test image, producing a flood of false alarms when applied to web-scale collections. Text-to-image retrieval excels at sifting through massive databases to rank relevant images, yet it stops at whole-image matches and offers no fine-grained localization. We introduce Referring Search and Discovery (ReSeDis), the first task that unifies corpus-level retrieval with pixel-level grounding. Given a free-form description, a ReSeDis model must decide whether the queried object appears in each image and, if so, where it is, returning bounding boxes or segmentation masks. To enable rigorous study, we curate a benchmark in which every description maps uniquely to object instances scattered across a large, diverse corpus, eliminating unintended matches. We further design a task-specific metric that jointly scores retrieval recall and localization precision. Finally, we provide a straightforward zero-shot baseline using a frozen vision-language model, revealing significant headroom for future study. ReSeDis offers a realistic, end-to-end testbed for building the next generation of robust and scalable multimodal search systems.

Jiahao Chang, Chongjie Ye, Yushuang Wu et al.

Existing multi-view 3D object reconstruction methods heavily rely on sufficient overlap between input views, where occlusions and sparse coverage in practice frequently yield severe reconstruction incompleteness. Recent advancements in diffusion-based 3D generative techniques offer the potential to address these limitations by leveraging learned generative priors to hallucinate invisible parts of objects, thereby generating plausible 3D structures. However, the stochastic nature of the inference process limits the accuracy and reliability of generation results, preventing existing reconstruction frameworks from integrating such 3D generative priors. In this work, we comprehensively analyze the reasons why diffusion-based 3D generative methods fail to achieve high consistency, including (a) the insufficiency in constructing and leveraging cross-view connections when extracting multi-view image features as conditions, and (b) the poor controllability of iterative denoising during local detail generation, which easily leads to plausible but inconsistent fine geometric and texture details with inputs. Accordingly, we propose ReconViaGen to innovatively integrate reconstruction priors into the generative framework and devise several strategies that effectively address these issues. Extensive experiments demonstrate that our ReconViaGen can reconstruct complete and accurate 3D models consistent with input views in both global structure and local details.Project page: https://jiahao620.github.io/reconviagen.

Kaiwen Wei, Xiao Liu, Jie Zhang et al.

Multimodal Retrieval-Augmented Generation (MRAG) enables Multimodal Large Language Models (MLLMs) to generate responses with external multimodal evidence, and numerous video-based MRAG benchmarks have been proposed to evaluate model capabilities across retrieval and generation stages. However, existing benchmarks remain limited in modality coverage and format diversity, often focusing on single- or limited-modality tasks, or coarse-grained scene understanding. To address these gaps, we introduce CFVBench, a large-scale, manually verified benchmark constructed from 599 publicly available videos, yielding 5,360 open-ended QA pairs. CFVBench spans high-density formats and domains such as chart-heavy reports, news broadcasts, and software tutorials, requiring models to retrieve and reason over long temporal video spans while maintaining fine-grained multimodal information. Using CFVBench, we systematically evaluate 7 retrieval methods and 14 widely-used MLLMs, revealing a critical bottleneck: current models (even GPT5 or Gemini) struggle to capture transient yet essential fine-grained multimodal details. To mitigate this, we propose Adaptive Visual Refinement (AVR), a simple yet effective framework that adaptively increases frame sampling density and selectively invokes external tools when necessary. Experiments show that AVR consistently enhances fine-grained multimodal comprehension and improves performance across all evaluated MLLMs

Yidan Zhang, Mutian Xu, Yiming Hao et al.

Facing scaling laws, video data from the internet becomes increasingly important. However, collecting extensive videos that meet specific needs is extremely labor-intensive and time-consuming. In this work, we study the way to expedite this collection process and propose VC-Agent, the first interactive agent that is able to understand users' queries and feedback, and accordingly retrieve/scale up relevant video clips with minimal user input. Specifically, considering the user interface, our agent defines various user-friendly ways for the user to specify requirements based on textual descriptions and confirmations. As for agent functions, we leverage existing multi-modal large language models to connect the user's requirements with the video content. More importantly, we propose two novel filtering policies that can be updated when user interaction is continually performed. Finally, we provide a new benchmark for personalized video dataset collection, and carefully conduct the user study to verify our agent's usage in various real scenarios. Extensive experiments demonstrate the effectiveness and efficiency of our agent for customized video dataset collection. Project page: https://allenyidan.github.io/vcagent_page/.

Yipeng Zhang, Yifan Liu, Zonghao Guo et al.

Vision transformers (ViTs) are widely employed in multimodal large language models (MLLMs) for visual encoding. However, they exhibit inferior performance on tasks regarding fine-grained visual perception. We attribute this to the limitations of ViTs in capturing diverse multi-modal visual levels, such as low-level details. To address this issue, we present LLaVA-UHD v2, an MLLM with advanced perception abilities by introducing a well-designed vision-language projector, the Hierarchical window (Hiwin) transformer. Hiwin transformer enhances MLLM's ability to capture diverse multi-modal visual granularities, by incorporating our constructed high-resolution semantic pyramid. Specifically, Hiwin transformer comprises two key modules: (i) a visual detail injection module, which progressively injects low-level visual details into high-level language-aligned semantics features, thereby forming an inverse semantic pyramid (ISP), and (ii) a hierarchical window attention module, which leverages cross-scale windows to condense multi-level semantics from the ISP. Extensive experiments show that LLaVA-UHD v2 outperforms compared MLLMs on a wide range of benchmarks. Notably, our design achieves an average boost of 3.7% across 14 benchmarks compared with the baseline method, 9.3% on DocVQA for instance. All the data and code will be publicly available to facilitate future research.

Jerone T. A. Andrews, Yidan Zhang, Lewis D. Griffin

The model of camera that was used to capture a particular photographic image (model attribution) is typically inferred from high-frequency model-specific artifacts present within the image. Model anonymization is the process of transforming these artifacts such that the apparent capture model is changed. We propose a conditional adversarial approach for learning such transformations. In contrast to previous works, we cast model anonymization as the process of transforming both high and low spatial frequency information. We augment the objective with the loss from a pre-trained dual-stream model attribution classifier, which constrains the generative network to transform the full range of artifacts. Quantitative comparisons demonstrate the efficacy of our framework in a restrictive non-interactive black-box setting.

Dayiheng Liu, Jie Fu, Yidan Zhang et al.

Typical methods for unsupervised text style transfer often rely on two key ingredients: 1) seeking the explicit disentanglement of the content and the attributes, and 2) troublesome adversarial learning. In this paper, we show that neither of these components is indispensable. We propose a new framework that utilizes the gradients to revise the sentence in a continuous space during inference to achieve text style transfer. Our method consists of three key components: a variational auto-encoder (VAE), some attribute predictors (one for each attribute), and a content predictor. The VAE and the two types of predictors enable us to perform gradient-based optimization in the continuous space, which is mapped from sentences in a discrete space, to find the representation of a target sentence with the desired attributes and preserved content. Moreover, the proposed method naturally has the ability to simultaneously manipulate multiple fine-grained attributes, such as sentence length and the presence of specific words, when performing text style transfer tasks. Compared with previous adversarial learning based methods, the proposed method is more interpretable, controllable and easier to train. Extensive experimental studies on three popular text style transfer tasks show that the proposed method significantly outperforms five state-of-the-art methods.