Searcharxiv⌕ Search

arXiv subjects

Yihao Zhao

Publications and source records attributed to Yihao Zhao.

17 recordsLinked to original sources
arXiv2026

AI Infrastructure in Space: How Far Can We Go?

Satellites are becoming programmable computing platforms capable of running increasingly demanding AI workloads. This shift raises a systems problem: how can AI services remain deployable, manageable, and recoverable after launch when compute capacity, connectivity, energy, and thermal headroom vary over orbital time? This paper develops a systems vision for AI infrastructure in space. We define it as the systems layer that manages AI capabilities across spacecraft, orbital networks, ground stations, and cloud backends, while treating orbital and physical state as part of the resource model. We synthesize relevant foundations from terrestrial AI infrastructure, satellite networking, and satellite edge computing, and examine the physical constraints that directly shape system design. We further ground this vision in three in-orbit case studies spanning the node, platform, and service levels. Telemetry from BUPT-1 satellite shows that usable compute capacity is bounded by thermal and energy envelopes. SateLight on BUPT-2 satellite reduces application-update transmission latency by 56.54% on average and up to 91.18%, with 100% update correctness. A stateful VLM serving case further shows that thermal interruptions make execution-state recovery a first-class systems problem. These observations motivate a research agenda for space-native resource management, lifecycle support, and sustained AI service across space and ground.

cs.DC↗
arXiv2026

Improving Text-Instance Alignment Of Foreground Conditioned Out-Painting Via Customized Concept Embedding

To showcase products, merchants often incur substantial costs creating high-quality display images. Foreground Conditioned Outpainting (FCO) meets this demand, allowing users to create desired backgrounds for foreground instances at a low cost by adjusting the text prompt. However, existing text-driven FCO methods exhibit critical flaws in their outputs, most notably the presence of artifacts, which refer to regions in the synthesized background that share the same semantics as the foreground instance. Such artifacts diminish the object's prominence and degrade image quality. We attribute the issue to the misalignment between the given instance and text-derived concept embeddings. To address this, we propose the Customized Concept Embedding Diffusion (CCE-Diffusion) framework. Its core is a CCE-Module to customize concept embeddings, bridging the gap between generic noun semantics and a specific visual instance. An Instance-Aware Loss guides the module's optimization, while a Semantic-Preserving Prompt Template prevents customized embeddings from distorting other words in the prompt. Both qualitative and quantitative evaluations demonstrate that CCE-Diffusion significantly reduces artifacts in the outputs. As a plug-and-play component, the CCE-Module can integrate with various FCO methods, enhancing their performance.

cs.CV↗
arXiv2026

Pose-ICL: 3D-Aware In-Context Learning for Pose-Controllable Subject Customization

Subject Customization is a foundational task in modern image generation. By providing a few reference images and a text prompt, users can generate images of a specific object in any desired scene. However, existing methods still struggle to achieve effective pose control for customized subjects. In practice, they often exhibit inaccurate poses or inconsistent cross-pose appearances. These limitations suggest that understanding objects in a volumetric manner remains a significant challenge for 2D-native backbones. To address this challenge, we propose Pose-ICL, a tuning-free framework that leverages 3D-aware In-Context Learning (ICL) to directly adapt to new subjects through multiple paired image-pose references. Its core mechanism,Surface-Anchored Position Embedding (SAPE), equips the model with explicit 3D awareness by anchoring image tokens to the surface coordinates of a volumetric bounding box. Dedicated refinements ensure its seamless compatibility with existing DiT models. Extensive evaluations on both 3D assets and real-world subjects demonstrate that Pose-ICL significantly outperforms current methods in both pose accuracy and identity consistency.

cs.CV↗
arXiv2026

ARL-Tangram: Unleash the Resource Efficiency in Agentic Reinforcement Learning

Agentic reinforcement learning (RL) has emerged as a transformative workload in cloud clusters, enabling large language models (LLMs) to solve complex problems through interactions with real world. However, unlike traditional RL, agentic RL demands substantial external cloud resources, e.g., CPUs for code execution and GPUs for reward models, that exist outside the primary training cluster. Existing agentic RL framework typically rely on static over-provisioning, i.e., resources are often tied to long-lived trajectories or isolated by tasks, which leads to severe resource inefficiency. We propose the action-level orchestration, and incorporate it into ARL-Tangram, a unified resource management system that enables fine-grained external resource sharing and elasticity. ARL-Tangram utilizes a unified action-level formulation and an elastic scheduling algorithm to minimize action completion time (ACT) while satisfying heterogeneous resource constraints. Further, heterogeneous resource managers are tailored to efficiently support the action-level execution on resources with heterogeneous characteristics and topologies. Evaluation on real-world agentic RL tasks demonstrates that ARL-Tangram improves average ACT by up to 4.3$\times$, speeds up the step duration of RL training by up to 1.5$\times$, and saves the external resources by up to 71.2$\%$. This system has been deployed to support the training of the MiMo series models.

cs.DC↗
arXiv2026

MiMo-V2-Flash Technical Report

We present MiMo-V2-Flash, a Mixture-of-Experts (MoE) model with 309B total parameters and 15B active parameters, designed for fast, strong reasoning and agentic capabilities. MiMo-V2-Flash adopts a hybrid attention architecture that interleaves Sliding Window Attention (SWA) with global attention, with a 128-token sliding window under a 5:1 hybrid ratio. The model is pre-trained on 27 trillion tokens with Multi-Token Prediction (MTP), employing a native 32k context length and subsequently extended to 256k. To efficiently scale post-training compute, MiMo-V2-Flash introduces a novel Multi-Teacher On-Policy Distillation (MOPD) paradigm. In this framework, domain-specialized teachers (e.g., trained via large-scale reinforcement learning) provide dense and token-level reward, enabling the student model to perfectly master teacher expertise. MiMo-V2-Flash rivals top-tier open-weight models such as DeepSeek-V3.2 and Kimi-K2, despite using only 1/2 and 1/3 of their total parameters, respectively. During inference, by repurposing MTP as a draft model for speculative decoding, MiMo-V2-Flash achieves up to 3.6 acceptance length and 2.6x decoding speedup with three MTP layers. We open-source both the model weights and the three-layer MTP weights to foster open research and community collaboration.

cs.CL↗
arXiv2025

MiMo-Audio: Audio Language Models are Few-Shot Learners

Existing audio language models typically rely on task-specific fine-tuning to accomplish particular audio tasks. In contrast, humans are able to generalize to new audio tasks with only a few examples or simple instructions. GPT-3 has shown that scaling next-token prediction pretraining enables strong generalization capabilities in text, and we believe this paradigm is equally applicable to the audio domain. By scaling MiMo-Audio's pretraining data to over one hundred million of hours, we observe the emergence of few-shot learning capabilities across a diverse set of audio tasks. We develop a systematic evaluation of these capabilities and find that MiMo-Audio-7B-Base achieves SOTA performance on both speech intelligence and audio understanding benchmarks among open-source models. Beyond standard metrics, MiMo-Audio-7B-Base generalizes to tasks absent from its training data, such as voice conversion, style transfer, and speech editing. MiMo-Audio-7B-Base also demonstrates powerful speech continuation capabilities, capable of generating highly realistic talk shows, recitations, livestreaming and debates. At the post-training stage, we curate a diverse instruction-tuning corpus and introduce thinking mechanisms into both audio understanding and generation. MiMo-Audio-7B-Instruct achieves open-source SOTA on audio understanding benchmarks (MMSU, MMAU, MMAR, MMAU-Pro), spoken dialogue benchmarks (Big Bench Audio, MultiChallenge Audio) and instruct-TTS evaluations, approaching or surpassing closed-source models. Model checkpoints and full evaluation suite are available at https://github.com/XiaomiMiMo/MiMo-Audio.

cs.CL↗
arXiv2025

The Madelung Problem of Finite Crystals

The Coulomb potential at an interior ion in a finite crystal of size $p$ is given by a linear superposition of contributions from displacement vectors ${\mathbf r}=(x,y,z)$ to its neighbors. This additive structure underlies universal relationships among Madelung constants and applies to both standard periodic boundary conditions and alternative Clifford supercells. Each pairwise contribution decomposes into three physically distinct components: a periodic bulk term, a quadratic boundary term, and a finite-size correction whose leading order term is $[24r^4-40(x^4+y^4+z^4)]/[9\sqrt{3} (2p+1)^2]$ for cubic crystals with unit lattice constant. Combining this decomposition with linear superposition yields a rapidly convergent direct-summation scheme, accurate even at $p=1$ ($3^3$ unit cells), enabling hands-on calculations of Madelung constants for a wide range of ionic crystals.

cond-mat.mtrl-sci↗
arXiv2025

Towards Autonomous UAV Visual Object Search in City Space: Benchmark and Agentic Methodology

Aerial Visual Object Search (AVOS) tasks in urban environments require Unmanned Aerial Vehicles (UAVs) to autonomously search for and identify target objects using visual and textual cues without external guidance. Existing approaches struggle in complex urban environments due to redundant semantic processing, similar object distinction, and the exploration-exploitation dilemma. To bridge this gap and support the AVOS task, we introduce CityAVOS, the first benchmark dataset for autonomous search of common urban objects. This dataset comprises 2,420 tasks across six object categories with varying difficulty levels, enabling comprehensive evaluation of UAV agents' search capabilities. To solve the AVOS tasks, we also propose PRPSearcher (Perception-Reasoning-Planning Searcher), a novel agentic method powered by multi-modal large language models (MLLMs) that mimics human three-tier cognition. Specifically, PRPSearcher constructs three specialized maps: an object-centric dynamic semantic map enhancing spatial perception, a 3D cognitive map based on semantic attraction values for target reasoning, and a 3D uncertainty map for balanced exploration-exploitation search. Also, our approach incorporates a denoising mechanism to mitigate interference from similar objects and utilizes an Inspiration Promote Thought (IPT) prompting mechanism for adaptive action planning. Experimental results on CityAVOS demonstrate that PRPSearcher surpasses existing baselines in both success rate and search efficiency (on average: +37.69% SR, +28.96% SPL, -30.69% MSS, and -46.40% NE). While promising, the performance gap compared to humans highlights the need for better semantic reasoning and spatial exploration capabilities in AVOS tasks. This work establishes a foundation for future advances in embodied target search. Dataset and source code are available at https://anonymous.4open.science/r/CityAVOS-3DF8.

cs.CV↗
arXiv2025

SeaLLM: Service-Aware and Latency-Optimized Resource Sharing for Large Language Model Inference

Large language models (LLMs) with different architectures and sizes have been developed. Serving each LLM with dedicated GPUs leads to resource waste and service inefficiency due to the varying demand of LLM requests. A common practice is to share multiple LLMs. However, existing sharing systems either do not consider the autoregressive pattern of LLM services, or only focus on improving the throughput, which impairs the sharing performance, especially the serving latency. We present SeaLLM, which enables service-aware and latency-optimized LLM sharing. SeaLLM improves the overall sharing performance by (1) a latency-optimized scheduling algorithm utilizing the characteristics of LLM services, (2) a placement algorithm to determine the placement plan and an adaptive replacement algorithm to decide the replacement interval, and (3) a unified key-value cache to share GPU memory among LLM services efficiently. Our evaluation under real-world traces and LLM services demonstrates that SeaLLM improves the normalized latency by up to $13.60\times$, the tail latency by up to $18.69\times$, and the SLO attainment by up to $3.64\times$ compared to existing solutions.

cs.DC↗
arXiv2025

Infinite Boundary Terms and Pairwise Interactions: A Unified Framework for Periodic Coulomb Systems

The introduction of the infinite boundary terms and the pairwise interactions [J. Chem. Theory Comput., 10, 5254, (2014)] enables a physically intuitive approach for deriving electrostatic energy and pressure for both neutral and non-neutral systems under the periodic boundary condition (PBC). For a periodic system consisting of $N$ point charges (with charge $q_j$ located at ${\mathbf r}_j$ where $j=1,2,\cdots N$) and one charge distribution of density $\rho({\mathbf r})$ within a primary cell of volume $V$, the derived electrostatic energy can be expressed as, \[ {\mathcal U} = \sum_{i<j}^N q_iq_j\nu({\mathbf r}_{ij} ) + \sum_{j=1}^N q_j \int_V d{\mathbf r}_0\,\rho({\mathbf r}_0) \nu({\mathbf r}_{0j} ) + \frac{1}{2}\int_V d{\mathbf r}_0 \int_V d{\mathbf r}_1\,\rho({\mathbf r}_0)\rho({\mathbf r}_1) \nu({\mathbf r}_{01}), \] where ${\mathbf r}_{ij}={\mathbf r}_i - {\mathbf r}_j$ is the relative vector and $\nu({\mathbf r})$ represents the effective pairwise interaction under PBC. The charge density $\rho({\mathbf r})$ is free of Delta-function-like divergence throughout the volume but may exhibit discontinuity. This unified formulation directly follows that of the isolated system by replacing the Coulomb interaction $1/\lvert {\mathbf r} \rvert$ or other modified Coulomb interactions with $\nu({\mathbf r})$. For a particular system of one-component plasma with a uniform neutralizing background, the implementation of various pairwise formulations clarifies the contribution of the background and subsequently reveals criteria for designing volume-dependent potentials that preserve the simple relation between energy and pressure.

physics.comp-ph↗
arXiv2024

A Survey of Resource-efficient LLM and Multimodal Foundation Models

Large foundation models, including large language models (LLMs), vision transformers (ViTs), diffusion, and LLM-based multimodal models, are revolutionizing the entire machine learning lifecycle, from training to deployment. However, the substantial advancements in versatility and performance these models offer come at a significant cost in terms of hardware resources. To support the growth of these large models in a scalable and environmentally sustainable way, there has been a considerable focus on developing resource-efficient strategies. This survey delves into the critical importance of such research, examining both algorithmic and systemic aspects. It offers a comprehensive analysis and valuable insights gleaned from existing literature, encompassing a broad array of topics from cutting-edge model architectures and training/serving algorithms to practical system designs and implementations. The goal of this survey is to provide an overarching understanding of how current approaches are tackling the resource challenges posed by large foundation models and to potentially inspire future breakthroughs in this field.

cs.LG↗
arXiv2024

TG-LMM: Enhancing Medical Image Segmentation Accuracy through Text-Guided Large Multi-Modal Model

We propose TG-LMM (Text-Guided Large Multi-Modal Model), a novel approach that leverages textual descriptions of organs to enhance segmentation accuracy in medical images. Existing medical image segmentation methods face several challenges: current medical automatic segmentation models do not effectively utilize prior knowledge, such as descriptions of organ locations; previous text-visual models focus on identifying the target rather than improving the segmentation accuracy; prior models attempt to use prior knowledge to enhance accuracy but do not incorporate pre-trained models. To address these issues, TG-LMM integrates prior knowledge, specifically expert descriptions of the spatial locations of organs, into the segmentation process. Our model utilizes pre-trained image and text encoders to reduce the number of training parameters and accelerate the training process. Additionally, we designed a comprehensive image-text information fusion structure to ensure thorough integration of the two modalities of data. We evaluated TG-LMM on three authoritative medical image datasets, encompassing the segmentation of various parts of the human body. Our method demonstrated superior performance compared to existing approaches, such as MedSAM, SAM and nnUnet.

cs.CV↗
arXiv2024

Non-unique Hamiltonians for Discrete Symplectic Dynamics

An outstanding property of any Hamiltonian system is the symplecticity of its flow, namely, the continuous trajectory preserves volume in phase space. Given a symplectic but discrete trajectory generated by a transition matrix applied at a fixed time-increment ($τ> 0$), it was generally believed that there exists a unique Hamiltonian producing a continuous trajectory that coincides at all discrete times ($t = nτ$ with $n$ integers) as long as $τ$ is small enough. However, it is now exactly demonstrated that, for any given discrete symplectic dynamics of a harmonic oscillator, there exist an infinite number of real-valued Hamiltonians for any small value of $τ$ and an infinite number of complex-valued Hamiltonians for any large value of $τ$. In addition, when the transition matrix is similar to a Jordan normal form with the supradiagonal element of $1$ and the two identical diagonal elements of either $1$ or $-1$, only one solution to the Hamiltonian is found for the case with the diagonal elements of $1$, but no solution can be found for the other case.

math-ph↗
arXiv2024

Quality assurance of organs-at-risk delineation in radiotherapy

The delineation of tumor target and organs-at-risk is critical in the radiotherapy treatment planning. Automatic segmentation can be used to reduce the physician workload and improve the consistency. However, the quality assurance of the automatic segmentation is still an unmet need in clinical practice. The patient data used in our study was a standardized dataset from AAPM Thoracic Auto-Segmentation Challenge. The OARs included were left and right lungs, heart, esophagus, and spinal cord. Two groups of OARs were generated, the benchmark dataset manually contoured by experienced physicians and the test dataset automatically created using a software AccuContour. A resnet-152 network was performed as feature extractor, and one-class support vector classifier was used to determine the high or low quality. We evaluate the model performance with balanced accuracy, F-score, sensitivity, specificity and the area under the receiving operator characteristic curve. We randomly generated contour errors to assess the generalization of our method, explored the detection limit, and evaluated the correlations between detection limit and various metrics such as volume, Dice similarity coefficient, Hausdorff distance, and mean surface distance. The proposed one-class classifier outperformed in metrics such as balanced accuracy, AUC, and others. The proposed method showed significant improvement over binary classifiers in handling various types of errors. Our proposed model, which introduces residual network and attention mechanism in the one-class classification framework, was able to detect the various types of OAR contour errors with high accuracy. The proposed method can significantly reduce the burden of physician review for contour delineation.

cs.CV↗
arXiv2023

MuxFlow: Efficient and Safe GPU Sharing in Large-Scale Production Deep Learning Clusters

Large-scale GPU clusters are widely-used to speed up both latency-critical (online) and best-effort (offline) deep learning (DL) workloads. However, most DL clusters either dedicate each GPU to one workload or share workloads in time, leading to very low GPU resource utilization. We present MuxFlow, the first production cluster system that supports efficient and safe space-sharing for DL workloads. NVIDIA MPS provides an opportunity to share multiple workloads in space on widely-deployed NVIDIA GPUs, but it cannot guarantee the performance and safety of online workloads. MuxFlow introduces a two-level protection mechanism for memory and computation to guarantee the performance of online workloads. Based on our practical error analysis, we design a mixed error-handling mechanism to guarantee the safety of online workloads. MuxFlow further proposes dynamic streaming multiprocessor (SM) allocation and matching-based scheduling to improve the efficiency of offline workloads. MuxFlow has been deployed at CompanyX's clusters with more than 20,000 GPUs. The deployment results indicate that MuxFlow substantially improves the GPU utilization from 26$\%$ to 76$\%$, SM activity from 16$\%$ to 33$\%$, and GPU memory from 42$\%$ to 48$\%$.

cs.DC↗
arXiv2021

Unpaired Image-to-Image Translation using Adversarial Consistency Loss

Unpaired image-to-image translation is a class of vision problems whose goal is to find the mapping between different image domains using unpaired training data. Cycle-consistency loss is a widely used constraint for such problems. However, due to the strict pixel-level constraint, it cannot perform geometric changes, remove large objects, or ignore irrelevant texture. In this paper, we propose a novel adversarial-consistency loss for image-to-image translation. This loss does not require the translated image to be translated back to be a specific source image but can encourage the translated images to retain important features of the source images and overcome the drawbacks of cycle-consistency loss noted above. Our method achieves state-of-the-art results on three challenging tasks: glasses removal, male-to-female translation, and selfie-to-anime translation.

cs.CV↗
arXiv2020

DLGAN: Disentangling Label-Specific Fine-Grained Features for Image Manipulation

Recent studies have shown how disentangling images into content and feature spaces can provide controllable image translation/ manipulation. In this paper, we propose a framework to enable utilizing discrete multi-labels to control which features to be disentangled, i.e., disentangling label-specific fine-grained features for image manipulation (dubbed DLGAN). By mapping the discrete label-specific attribute features into a continuous prior distribution, we leverage the advantages of both discrete labels and reference images to achieve image manipulation in a hybrid fashion. For example, given a face image dataset (e.g., CelebA) with multiple discrete fine-grained labels, we can learn to smoothly interpolate a face image between black hair and blond hair through reference images while immediately controlling the gender and age through discrete input labels. To the best of our knowledge, this is the first work that realizes such a hybrid manipulation within a single model. More importantly, it is the first work to achieve image interpolation between two different domains without requiring continuous labels as the supervision. Qualitative and quantitative experiments demonstrate the effectiveness of the proposed method.

cs.CV↗