Hanxin Zhu

Cong Wang, Hanxin Zhu, Jiayi Luo et al.

Large-scale video generation models have made remarkable progress in semantic consistency and visual quality, producing videos that are increasingly coherent and visually convincing. Nevertheless, the dynamics induced by pixel-level fitting do not naturally accommodate the regularities that govern real-world motion and interaction, resulting in persistent shortcomings in physical plausibility. To address this limitation, we propose \textbf{PILA} (Physics-Informed Latent Alignment), a framework that injects physics-structured latent guidance into the frozen flow-matching dynamics of pretrained video models. Specifically, PILA first employs anchored field estimation to map frozen-generator latents into an operational physical attribute bank organized by field-proxy slots, using observable motion as a kinematic anchor for constructing less directly observed proxies. To handle the heterogeneity of real-world dynamics, PILA adopts a mixture-of-experts design over physical categories. Label-prior masked expert routing selects category-specific operator experts, whose refinements are regularized by operational residuals abstracted from physical relations. Finally, the refined proxies are fused into the physical attribute bank and decoded into a correction to the flow-matching vector field, injecting physics-aware guidance while preserving the visual prior of the pretrained backbone. With staged adapter training on Wan 2.1-1.3B and direct transfer of the learned adapter to Wan 2.2-14B, PILA achieves state-of-the-art results on VBench-2.0, VideoPhy-2, and PhyGenBench in both visual quality and benchmark-measured physical plausibility.

Xin Li, Bingchen Li, Yeying Jin et al.

Compressed Image Super-resolution (CSR) aims to simultaneously super-resolve the compressed images and tackle the challenging hybrid distortions caused by compression. However, existing works on CSR usually focuses on a single compression codec, i.e., JPEG, ignoring the diverse traditional or learning-based codecs in the practical application, e.g., HEVC, VVC, HIFIC, etc. In this work, we propose the first universal CSR framework, dubbed UCIP, with dynamic prompt learning, intending to jointly support the CSR distortions of any compression codecs/modes. Particularly, an efficient dynamic prompt strategy is proposed to mine the content/spatial-aware task-adaptive contextual information for the universal CSR task, using only a small amount of prompts with spatial size 1x1. To simplify contextual information mining, we introduce the novel MLP-like framework backbone for our UCIP by adapting the Active Token Mixer (ATM) to CSR tasks for the first time, where the global information modeling is only taken in horizontal and vertical directions with offset prediction. We also build an all-in-one benchmark dataset for the CSR task by collecting the datasets with the popular 6 diverse traditional and learning-based codecs, including JPEG, HEVC, VVC, HIFIC, etc., resulting in 23 common degradations. Extensive experiments have shown the consistent and excellent performance of our UCIP on universal CSR tasks. The project can be found in https://lixinustc.github.io/UCIP.github.io

Jiayi Luo, Qiyan Liu, Tengyang Wang et al.

Autoregressive (AR) video generation has emerged as a promising paradigm for long-horizon video synthesis, where each frame is generated conditioned on previously generated tokens. To accelerate inference, the KV cache is used to avoid redundant recomputation across generation steps. Nevertheless, its growth with generation length introduces increasing memory and error accumulation, limiting the scalability of AR models to even longer sequences. Existing KV cache compression methods mitigate this issue by selectively retaining only video tokens deemed important. However, most existing methods assess token importance using short-horizon signals derived from the current or historical generation context, making these methods prone to overlooking tokens that appear unimportant at early steps but later become critical for future frames. In this work, we identify an important property of trained AR video models: although RoPE-modulated queries evolve across autoregressive steps, the underlying canonical pre-RoPE query distribution remains remarkably stable throughout the video generation process. This approximate stationarity implies that future query distributions are estimable from historical statistics, enabling principled future-aware cache decisions without any additional training. Building on this insight, we propose Future Forcing, a training-free future-aware KV cache policy for AR video generation. Specifically, Future Forcing first constructs a future query proxy from historical statistics, then scores KV cache tokens by their importance under this proxy, and finally merges redundant token pairs within the affine subspace induced by the future query. Extensive experiments show that Future Forcing improves long-horizon consistency under limited KV caches, achieving up to 1.49 improvement in subject consistency on VBench-Long for 60s generation over existing AR video KV cache policies.

Hanxin Zhu, Henan Wang, Zhibo Chen

Traditional representations for light fields can be separated into two types: explicit representation and implicit representation. Unlike explicit representation that represents light fields as Sub-Aperture Images (SAIs) based arrays or Micro-Images (MIs) based lenslet images, implicit representation treats light fields as neural networks, which is inherently a continuous representation in contrast to discrete explicit representation. However, at present almost all the implicit representations for light fields utilize SAIs to train an MLP to learn a pixel-wise mapping from 4D spatial-angular coordinate to pixel colors, which is neither compact nor of low complexity. Instead, in this paper we propose MiNL, a novel MI-wise implicit neural representation for light fields that train an MLP + CNN to learn a mapping from 2D MI coordinates to MI colors. Given the micro-image's coordinate, MiNL outputs the corresponding micro-image's RGB values. Light field encoding in MiNL is just training a neural network to regress the micro-images and the decoding process is a simple feedforward operation. Compared with common pixel-wise implicit representation, MiNL is more compact and efficient that has faster decoding speed (\textbf{$\times$80$\sim$180} speed-up) as well as better visual quality (\textbf{1$\sim$4dB} PSNR improvement on average).

Hanxin Zhu, Tianyu He, Anni Tang et al.

Significant progress has been made in text-to-video generation through the use of powerful generative models and large-scale internet data. However, substantial challenges remain in precisely controlling individual concepts within the generated video, such as the motion and appearance of specific characters and the movement of viewpoints. In this work, we propose a novel paradigm that generates each concept in 3D representation separately and then composes them with priors from Large Language Models (LLM) and 2D diffusion models. Specifically, given an input textual prompt, our scheme consists of three stages: 1) We leverage LLM as the director to first decompose the complex query into several sub-prompts that indicate individual concepts within the video~(\textit{e.g.}, scene, objects, motions), then we let LLM to invoke pre-trained expert models to obtain corresponding 3D representations of concepts. 2) To compose these representations, we prompt multi-modal LLM to produce coarse guidance on the scales and coordinates of trajectories for the objects. 3) To make the generated frames adhere to natural image distribution, we further leverage 2D diffusion priors and use Score Distillation Sampling to refine the composition. Extensive experiments demonstrate that our method can generate high-fidelity videos from text with diverse motion and flexible control over each concept. Project page: \url{https://aka.ms/c3v}.

Xin Ding, Xinrui Wang, Yifan Yang et al.

Existing memory systems for embodied agents typically inject retrieved memory as static context at episode start, a paradigm we term Ahead-of-time Monolithic Memory Injection (AMMI). However, this static design quickly becomes misaligned with the agent's evolving state and may degrade lightweight executors below the no-memory baseline. To address this, we propose MemCompiler, which reframes memory utilization as State-Conditioned Memory Compilation. A learned Memory Compiler reads a structured Brief State capturing the agent's current execution state and dynamically selects and compiles only relevant memory into executable guidance. This guidance is delivered through a text channel and a latent Soft-Mem channel that preserves perceptual information not expressible in text. Across Alf World, EmbodiedBench, and ScienceWorld, MemCompiler consistently improves over no-memory across open-source backbones (up to +129%), matches or approaches frontier closed-source systems, and reduces per-step latency by 60%, demonstrating that state-aware memory compilation improves both effectiveness and efficiency.

Jiayi Luo, Jiayu Chen, Jiankun Wang et al.

Diffusion Transformers (DiTs) achieve strong video generation quality but suffer from high inference cost due to dense 3D attention, leading to the development of sparse attention technologies to improve efficiency. However, existing training-free sparse attention methods in video generation still face two unresolved limitations: ignoring layer heterogeneity in attention pruning and ignoring query-key coupling in block partitioning, which hinder a better quality-speedup trade-off. In this work, we uncover a critical insight that the attention sparsity of each layer is its intrinsic property, with minor effects across different inputs. Motivated by this, we propose SVOO, a training-free Sparse attention framework for fast Video generation via Offline layer-wise sparsity profiling and Online bidirectional co-clustering. Specifically, SVOO adopts a two-stage paradigm: (i) offline layer-wise sensitivity profiling to derive intrinsic per-layer pruning levels, and (ii) online block-wise sparse attention via a novel bidirectional co-clustering algorithm. Extensive experiments on seven widely used video generation models demonstrate that SVOO achieves a superior quality-speedup trade-off over state-of-the-art methods, delivering up to $1.93\times$ speedup while maintaining a PSNR of up to 29 dB on Wan2.1.

Jingwen Sun, Wenyao Zhang, Zekun Qi et al.

Pretraining Vision-Language-Action (VLA) policies on internet-scale video is appealing, yet current latent-action objectives often learn the wrong thing: they remain anchored to pixel variation rather than action-relevant state transitions, making them vulnerable to appearance bias, nuisance motion, and information leakage. We introduce VLA-JEPA, a JEPA-style pretraining framework that sidesteps these pitfalls by design. The key idea is \emph{leakage-free state prediction}: a target encoder produces latent representations from future frames, while the student pathway sees only the current observation -- future information is used solely as supervision targets, never as input. By predicting in latent space rather than pixel space, VLA-JEPA learns dynamics abstractions that are robust to camera motion and irrelevant background changes. This yields a simple two-stage recipe -- JEPA pretraining followed by action-head fine-tuning -- without the multi-stage complexity of prior latent-action pipelines. Experiments on LIBERO, LIBERO-Plus, SimplerEnv and real-world manipulation tasks show that VLA-JEPA achieves consistent gains in generalization and robustness over existing methods.

Cong Wang, Hanxin Zhu, Xiao Tang et al.

Recent progress in video generation has led to substantial improvements in visual fidelity, yet ensuring physically consistent motion remains a fundamental challenge. Intuitively, this limitation can be attributed to the fact that real-world object motion unfolds in three-dimensional space, while video observations provide only partial, view-dependent projections of such dynamics. To address these issues, we propose PhysVideo, a two-stage framework that first generates physics-aware orthogonal foreground videos and then synthesizes full videos with background. In the first stage, Phys4View leverages physics-aware attention to capture the influence of physical attributes on motion dynamics, and enhances spatio-temporal consistency by incorporating geometry-enhanced cross-view attention and temporal attention. In the second stage, VideoSyn uses the generated foreground videos as guidance and learns the interactions between foreground dynamics and background context for controllable video synthesis. To support training, we construct PhysMV, a dataset containing 40K scenes, each consisting of four orthogonal viewpoints, resulting in a total of 160K video sequences. Extensive experiments demonstrate that PhysVideo significantly improves physical realism and spatial-temporal coherence over existing video generation methods. Home page: https://anonymous.4open.science/w/Phys4D/.

Hanxin Zhu, Cong Wang, Peiyan Tu et al.

Recent developments in generative models and large-scale datasets have substantially advanced 3D world generation, facilitating a broad range of domains including spatial intelligence, embodied intelligence, and autonomous driving. While achieving remarkable progress, existing approaches to 3D world generation typically prioritize appearance prediction with limited modeling of the underlying geometry, leading to issues such as unreliable scene structure estimation and degraded cross-view consistency. To address these limitations, motivated by the coarse-to-fine nature of human visual perception, we propose GTA, a novel image-to-3D world generation method following a Geometry-Then-Appearance paradigm. Specifically, given a single input image, to improve the structural fidelity of synthesized 3D scenes, GTA adopts a two-stage framework with two dedicated video diffusion models, which first generate coarse geometric structure from novel viewpoints and then synthesize fine-grained appearance conditioned on the predicted geometry. To further enhance cross-view appearance consistency, we introduce a random latent shuffle strategy during the training process, along with a test-time scaling scheme that improves perceptual quality without compromising quantitative performance. Extensive experiments have demonstrated that our proposed method consistently outperforms existing approaches in terms of fidelity, visual quality, and geometric accuracy. Moreover, GTA is shown to be effective as a general enhancement module that further improves the generation quality of existing image-to-3D world pipelines, as well as supporting multiple downstream applications and exhibiting favorable data efficiency during model training, highlighting its versatility and broad applicability. Project page: https://hanxinzhu-lab.github.io/GTA/.

Henan Wang, Hanxin Zhu, Tianyu He et al.

3D Gaussian Splatting (3DGS) has become an emerging technique with remarkable potential in 3D representation and image rendering. However, the substantial storage overhead of 3DGS significantly impedes its practical applications. In this work, we formulate the compact 3D Gaussian learning as an end-to-end Rate-Distortion Optimization (RDO) problem and propose RDO-Gaussian that can achieve flexible and continuous rate control. RDO-Gaussian addresses two main issues that exist in current schemes: 1) Different from prior endeavors that minimize the rate under the fixed distortion, we introduce dynamic pruning and entropy-constrained vector quantization (ECVQ) that optimize the rate and distortion at the same time. 2) Previous works treat the colors of each Gaussian equally, while we model the colors of different regions and materials with learnable numbers of parameters. We verify our method on both real and synthetic scenes, showcasing that RDO-Gaussian greatly reduces the size of 3D Gaussian over 40x, and surpasses existing methods in rate-distortion performance.

Bingchen Li, Xin Li, Hanxin Zhu et al.

Generative Adversarial Networks (GANs) have been widely used to recover vivid textures in image super-resolution (SR) tasks. In particular, one discriminator is utilized to enable the SR network to learn the distribution of real-world high-quality images in an adversarial training manner. However, the distribution learning is overly coarse-grained, which is susceptible to virtual textures and causes counter-intuitive generation results. To mitigate this, we propose the simple and effective Semantic-aware Discriminator (denoted as SeD), which encourages the SR network to learn the fine-grained distributions by introducing the semantics of images as a condition. Concretely, we aim to excavate the semantics of images from a well-trained semantic extractor. Under different semantics, the discriminator is able to distinguish the real-fake images individually and adaptively, which guides the SR network to learn the more fine-grained semantic-aware textures. To obtain accurate and abundant semantics, we take full advantage of recently popular pretrained vision models (PVMs) with extensive datasets, and then incorporate its semantic features into the discriminator through a well-designed spatial cross-attention module. In this way, our proposed semantic-aware discriminator empowered the SR network to produce more photo-realistic and pleasing images. Extensive experiments on two typical tasks, i.e., SR and Real SR have demonstrated the effectiveness of our proposed methods.

Peiyan Tu, Hanxin Zhu, Jingwen Sun et al.

World models have made significant progress in modeling dynamic environments; however, most embodied world models are still restricted to 2D representations, lacking the comprehensive multi-view information essential for embodied spatial reasoning. Bridging this gap is non-trivial, primarily due to challenges from severe scarcity of paired multi-view data, the difficulty of maintaining spatiotemporal consistency in generated 3D geometries, and the tendency to hallucinate manipulation details. To address these challenges, we propose Embody4D, a dedicated video-to-video world model for embodied scenarios, capable of synthesizing arbitrary novel views from a monocular video. First, to tackle data scarcity, we introduce a 3D-aware compositional synthesis pipeline to curate a heterogeneous dataset compositing cross-embodiment robotic arms with diverse backgrounds, guaranteeing broad generalization. Second, to enforce geometric stability, we devise an adaptive noise injection strategy; by leveraging confidence disparities across image regions, this method selectively regularizes the diffusion process to ensure strict spatiotemporal consistency. Finally, to guarantee manipulation fidelity, we incorporate an interaction-aware attention mechanism that explicitly attends to the robotic interaction regions. Extensive experiments demonstrate that Embody4D achieves state-of-the-art performance, serving as a robust world model that synthesizes high-fidelity, view-consistent videos to empower downstream robotic planning and learning.

Hanxin Zhu, Tianyu He, Xin Li et al. · microsoft-research

Neural Radiance Field (NeRF) has achieved superior performance for novel view synthesis by modeling the scene with a Multi-Layer Perception (MLP) and a volume rendering procedure, however, when fewer known views are given (i.e., few-shot view synthesis), the model is prone to overfit the given views. To handle this issue, previous efforts have been made towards leveraging learned priors or introducing additional regularizations. In contrast, in this paper, we for the first time provide an orthogonal method from the perspective of network structure. Given the observation that trivially reducing the number of model parameters alleviates the overfitting issue, but at the cost of missing details, we propose the multi-input MLP (mi-MLP) that incorporates the inputs (i.e., location and viewing direction) of the vanilla MLP into each layer to prevent the overfitting issue without harming detailed synthesis. To further reduce the artifacts, we propose to model colors and volume density separately and present two regularization terms. Extensive experiments on multiple datasets demonstrate that: 1) although the proposed mi-MLP is easy to implement, it is surprisingly effective as it boosts the PSNR of the baseline from $14.73$ to $24.23$. 2) the overall framework achieves state-of-the-art results on a wide range of benchmarks. We will release the code upon publication.

Hanxin Zhu, Tianyu He, Xiqian Yu et al.

Recent advancements in generative models have ignited substantial interest in dynamic 3D content creation (\ie, 4D generation). Existing approaches primarily rely on Score Distillation Sampling (SDS) to infer novel-view videos, typically leading to issues such as limited diversity, spatial-temporal inconsistency and poor prompt alignment, due to the inherent randomness of SDS. To tackle these problems, we propose AR4D, a novel paradigm for SDS-free 4D generation. Specifically, our paradigm consists of three stages. To begin with, for a monocular video that is either generated or captured, we first utilize pre-trained expert models to create a 3D representation of the first frame, which is further fine-tuned to serve as the canonical space. Subsequently, motivated by the fact that videos happen naturally in an autoregressive manner, we propose to generate each frame's 3D representation based on its previous frame's representation, as this autoregressive generation manner can facilitate more accurate geometry and motion estimation. Meanwhile, to prevent overfitting during this process, we introduce a progressive view sampling strategy, utilizing priors from pre-trained large-scale 3D reconstruction models. To avoid appearance drift introduced by autoregressive generation, we further incorporate a refinement stage based on a global deformation field and the geometry of each frame's 3D representation. Extensive experiments have demonstrated that AR4D can achieve state-of-the-art 4D generation without SDS, delivering greater diversity, improved spatial-temporal consistency, and better alignment with input prompts.

Xingrui Wang, Cuiling Lan, Hanxin Zhu et al.

Modeling and understanding the 3D world is crucial for various applications, from augmented reality to robotic navigation. Recent advancements based on 3D Gaussian Splatting have integrated semantic information from multi-view images into Gaussian primitives. However, these methods typically require costly per-scene optimization from dense calibrated images, limiting their practicality. In this paper, we consider the new task of generalizable 3D semantic field modeling from sparse, uncalibrated image pairs. Building upon the Splatt3R architecture, we introduce GSemSplat, a framework that learns open-vocabulary semantic representations linked to 3D Gaussians without the need for per-scene optimization, dense image collections or calibration. To ensure effective and reliable learning of semantic features in 3D space, we employ a dual-feature approach that leverages both region-specific and context-aware semantic features as supervision in the 2D space. This allows us to capitalize on their complementary strengths. Experimental results on the ScanNet++ dataset demonstrate the effectiveness and superiority of our approach compared to the traditional scene-specific method. We hope our work will inspire more research into generalizable 3D understanding.

Hanxin Zhu, Tianyu He, Zhibo Chen · microsoft-research

Neural Radiance Field (NeRF) has shown impressive results in novel view synthesis, particularly in Virtual Reality (VR) and Augmented Reality (AR), thanks to its ability to represent scenes continuously. However, when just a few input view images are available, NeRF tends to overfit the given views and thus make the estimated depths of pixels share almost the same value. Unlike previous methods that conduct regularization by introducing complex priors or additional supervisions, we propose a simple yet effective method that explicitly builds depth-aware consistency across input views to tackle this challenge. Our key insight is that by forcing the same spatial points to be sampled repeatedly in different input views, we are able to strengthen the interactions between views and therefore alleviate the overfitting problem. To achieve this, we build the neural networks on layered representations (\textit{i.e.}, multiplane images), and the sampling point can thus be resampled on multiple discrete planes. Furthermore, to regularize the unseen target views, we constrain the rendered colors and depths from different input views to be the same. Although simple, extensive experiments demonstrate that our proposed method can achieve better synthesis quality over state-of-the-art methods.

Xingrui Wang, Xin Li, Yaosi Hu et al.

Text-driven Image to Video Generation (TI2V) aims to generate controllable video given the first frame and corresponding textual description. The primary challenges of this task lie in two parts: (i) how to identify the target objects and ensure the consistency between the movement trajectory and the textual description. (ii) how to improve the subjective quality of generated videos. To tackle the above challenges, we propose a new diffusion-based TI2V framework, termed TIV-Diffusion, via object-centric textual-visual alignment, intending to achieve precise control and high-quality video generation based on textual-described motion for different objects. Concretely, we enable our TIV-Diffuion model to perceive the textual-described objects and their motion trajectory by incorporating the fused textual and visual knowledge through scale-offset modulation. Moreover, to mitigate the problems of object disappearance and misaligned objects and motion, we introduce an object-centric textual-visual alignment module, which reduces the risk of misaligned objects/motion by decoupling the objects in the reference image and aligning textual features with each object individually. Based on the above innovations, our TIV-Diffusion achieves state-of-the-art high-quality video generation compared with existing TI2V methods.

Henan Wang, Hanxin Zhu, Zhibo Chen

Light field, as a new data representation format in multimedia, has the ability to capture both intensity and direction of light rays. However, the additional angular information also brings a large volume of data. Classical coding methods are not effective to describe the relationship between different views, leading to redundancy left. To address this problem, we propose a novel light field compression scheme based on implicit neural representation to reduce redundancies between views. We store the information of a light field image implicitly in an neural network and adopt model compression methods to further compress the implicit representation. Extensive experiments have demonstrated the effectiveness of our proposed method, which achieves comparable rate-distortion performance as well as superior perceptual quality over traditional methods.

Yiting Lu, Wei Luo, Peiyan Tu et al.

World Generation Models are emerging as a cornerstone of next-generation multimodal intelligence systems. Unlike traditional 2D visual generation, World Models aim to construct realistic, dynamic, and physically consistent 3D/4D worlds from images, videos, or text. These models not only need to produce high-fidelity visual content but also maintain coherence across space, time, physics, and instruction control, enabling applications in virtual reality, autonomous driving, embodied intelligence, and content creation. However, prior benchmarks emphasize different evaluation dimensions and lack a unified assessment of world-realism capability. To systematically evaluate World Models, we introduce the 4DWorldBench, which measures models across four key dimensions: Perceptual Quality, Condition-4D Alignment, Physical Realism, and 4D Consistency. The benchmark covers tasks such as Image-to-3D/4D, Video-to-4D, Text-to-3D/4D. Beyond these, we innovatively introduce adaptive conditioning across multiple modalities, which not only integrates but also extends traditional evaluation paradigms. To accommodate different modality-conditioned inputs, we map all modality conditions into a unified textual space during evaluation, and further integrate LLM-as-judge, MLLM-as-judge, and traditional network-based methods. This unified and adaptive design enables more comprehensive and consistent evaluation of alignment, physical realism, and cross-modal coherence. Preliminary human studies further demonstrate that our adaptive tool selection achieves closer agreement with subjective human judgments. We hope this benchmark will serve as a foundation for objective comparisons and improvements, accelerating the transition from "visual generation" to "world generation." Our project can be found at https://yeppp27.github.io/4DWorldBench.github.io/.

Henan Wang, Hanxin Zhu, Xinliang Gong et al. · microsoft-research

3D Gaussian Splatting (3DGS) has garnered significant attention due to its superior scene representation fidelity and real-time rendering performance, especially for dynamic 3D scene reconstruction (\textit{i.e.}, 4D reconstruction). However, despite achieving promising results, most existing algorithms overlook the substantial temporal and spatial redundancies inherent in dynamic scenes, leading to prohibitive memory consumption. To address this, we propose P-4DGS, a novel dynamic 3DGS representation for compact 4D scene modeling. Inspired by intra- and inter-frame prediction techniques commonly used in video compression, we first design a 3D anchor point-based spatial-temporal prediction module to fully exploit the spatial-temporal correlations across different 3D Gaussian primitives. Subsequently, we employ an adaptive quantization strategy combined with context-based entropy coding to further reduce the size of the 3D anchor points, thereby achieving enhanced compression efficiency. To evaluate the rate-distortion performance of our proposed P-4DGS in comparison with other dynamic 3DGS representations, we conduct extensive experiments on both synthetic and real-world datasets. Experimental results demonstrate that our approach achieves state-of-the-art reconstruction quality and the fastest rendering speed, with a remarkably low storage footprint (around \textbf{1MB} on average), achieving up to \textbf{40$\times$} and \textbf{90$\times$} compression on synthetic and real-world scenes, respectively.

Xiqian Yu, Hanxin Zhu, Tianyu He et al.

Achieving high-resolution novel view synthesis (HRNVS) from low-resolution input views is a challenging task due to the lack of high-resolution data. Previous methods optimize high-resolution Neural Radiance Field (NeRF) from low-resolution input views but suffer from slow rendering speed. In this work, we base our method on 3D Gaussian Splatting (3DGS) due to its capability of producing high-quality images at a faster rendering speed. To alleviate the shortage of data for higher-resolution synthesis, we propose to leverage off-the-shelf 2D diffusion priors by distilling the 2D knowledge into 3D with Score Distillation Sampling (SDS). Nevertheless, applying SDS directly to Gaussian-based 3D super-resolution leads to undesirable and redundant 3D Gaussian primitives, due to the randomness brought by generative priors. To mitigate this issue, we introduce two simple yet effective techniques to reduce stochastic disturbances introduced by SDS. Specifically, we 1) shrink the range of diffusion timestep in SDS with an annealing strategy; 2) randomly discard redundant Gaussian primitives during densification. Extensive experiments have demonstrated that our proposed GaussainSR can attain high-quality results for HRNVS with only low-resolution inputs on both synthetic and real-world datasets. Project page: https://chchnii.github.io/GaussianSR/