arXiv subjects

Qing Liu

Publications and source records attributed to Qing Liu.

At least 55 records · Page 3Linked to original sources

arXiv2026

Controllable Localized Face Anonymization Via Diffusion Inpainting

The growing use of portrait images in computer vision highlights the need to protect personal identities. At the same time, anonymized images must remain useful for downstream computer vision tasks. In this work, we propose a unified framework that leverages the inpainting ability of latent diffusion models to generate realistic anonymized images. Unlike prior approaches, we have complete control over the anonymization process by designing an adaptive attribute-guidance module that applies gradient correction during the reverse denoising process, aligning the facial attributes of the generated image with those of the synthesized target image. Our framework also supports localized anonymization, allowing users to specify which facial regions are left unchanged. Extensive experiments conducted on the public CelebA-HQ and FFHQ datasets show that our method outperforms state-of-the-art approaches while requiring no additional model training. The source code is available on our page.

cs.CV↗

arXiv2026

PrivGemo: Privacy-Preserving Dual-Tower Graph Retrieval for Empowering LLM Reasoning with Memory Augmentation

Knowledge graphs (KGs) provide structured evidence that can ground large language model (LLM) reasoning for knowledge-intensive question answering. However, many practical KGs are private, and sending retrieved triples or exploration traces to closed-source LLM APIs introduces leakage risk. Existing privacy treatments focus on masking entity names, but they still face four limitations: structural leakage under semantic masking, uncontrollable remote interaction, fragile multi-hop and multi-entity reasoning, and limited experience reuse for stability and efficiency. To address these issues, we propose PrivGemo, a privacy-preserving retrieval-augmented framework for KG-grounded reasoning with memory-guided exposure control. PrivGemo uses a dual-tower design to keep raw KG knowledge local while enabling remote reasoning over an anonymized view that goes beyond name masking to limit both semantic and structural exposure. PrivGemo supports multi-hop, multi-entity reasoning by retrieving anonymized long-hop paths that connect all topic entities, while keeping grounding and verification on the local KG. A hierarchical controller and a privacy-aware experience memory further reduce unnecessary exploration and remote interactions. Comprehensive experiments on six benchmarks show that PrivGemo achieves overall state-of-the-art results, outperforming the strongest baseline by up to 17.1%. Furthermore, PrivGemo enables smaller models (e.g., Qwen3-4B) to achieve reasoning performance comparable to that of GPT-4-Turbo.

cs.CL↗

arXiv2026

On the cohomological representations of finite automorphism groups of singular curves and compact complex spaces

Let G be a finite group acting tamely on a proper reduced curve C over an algebraically closed field. We study the G-module structure on the cohomology groups of a G-equivariant locally free sheaf F on C, and give formulas of Chevalley--Weil type, with values in the Grothendieck ring R_k(G)_Q of finitely generated G-modules. We also give a similar formula for the singular cohomology of compact complex spaces. The focus is on the case where C is nodal. Using the Chevalley--Weil formula, we compute the G-invariant part of the global sections of the pluricanonical bundle ω_C^{\otimes m}. In turn, we use the formula for m=2 to compute the equivariant deformation space of a stable G-curve C. We also obtain numerical criteria for the presence of any given irreducible representation in space of the global sections of ω_C\otimes F, where F is an ample locally free G-sheaf on C. Some new phenomena, pathological compared to the smooth curve case, are discussed.

math.AG↗

arXiv2026

Unsupervised dense random survival forests identify interpretable patient profiles with heterogeneous treatment benefit

Precision oncology aims to prescribe the optimal cancer treatment to the right patients, maximizing therapeutic benefits. However, identifying patient subgroups that may benefit more from experimental cancer treatments based on randomized clinical trials presents a significant analytical challenge. To address this, we introduce a novel unsupervised machine learning approach based on very dense random survival forests (up to 100,000 trees), equipped with a new splitting rule that explicitly targets treatment-effect heterogeneity. This method is robust, interpretable, and effectively identifies responsive subgroups. Extensive simulations confirm its ability to detect heterogeneous patient responses and distinguish between datasets with and without heterogeneity, while maintaining a stringent Type I error rate of 1%. We further validate its performance using Phase III randomized clinical trial datasets, demonstrating significant patient heterogeneity in treatment response based on baseline characteristics.

stat.ME↗

arXiv2025

Both Semantics and Reconstruction Matter: Making Representation Encoders Ready for Text-to-Image Generation and Editing

Modern Latent Diffusion Models (LDMs) typically operate in low-level Variational Autoencoder (VAE) latent spaces that are primarily optimized for pixel-level reconstruction. To unify vision generation and understanding, a burgeoning trend is to adopt high-dimensional features from representation encoders as generative latents. However, we empirically identify two fundamental obstacles in this paradigm: (1) the discriminative feature space lacks compact regularization, making diffusion models prone to off-manifold latents that lead to inaccurate object structures; and (2) the encoder's inherently weak pixel-level reconstruction hinders the generator from learning accurate fine-grained geometry and texture. In this paper, we propose a systematic framework to adapt understanding-oriented encoder features for generative tasks. We introduce a semantic-pixel reconstruction objective to regularize the latent space, enabling the compression of both semantic information and fine-grained details into a highly compact representation (96 channels with 16x16 spatial downsampling). This design ensures that the latent space remains semantically rich and achieves state-of-the-art image reconstruction, while remaining compact enough for accurate generation. Leveraging this representation, we design a unified Text-to-Image (T2I) and image editing model. Benchmarking against various feature spaces, we demonstrate that our approach achieves state-of-the-art reconstruction, faster convergence, and substantial performance gains in both T2I and editing tasks, validating that representation encoders can be effectively adapted into robust generative components.

cs.CV↗

arXiv2025

HBridge: H-Shape Bridging of Heterogeneous Experts for Unified Multimodal Understanding and Generation

Recent unified models integrate understanding experts (e.g., LLMs) with generative experts (e.g., diffusion models), achieving strong multimodal performance. However, recent advanced methods such as BAGEL and LMFusion follow the Mixture-of-Transformers (MoT) paradigm, adopting a symmetric design that mirrors one expert to another for convenient initialization and fusion, which remains suboptimal due to inherent modality discrepancies. In this work, we propose HBridge, an asymmetric H-shaped architecture that enables heterogeneous experts to optimally leverage pretrained priors from their respective modality domains. Unlike prior dense fusion strategies that straightforwardly connect all layers between experts via shared attention, HBridge selectively bridges intermediate layers, reducing over 40% attention sharing, which improves efficiency and enhances generation quality. Shallow and deep layers, which capture modality-specific representations, are decoupled, while mid-layer bridging promotes semantic alignment. To further strengthen cross-modal coherence, we introduce semantic reconstruction tokens that explicitly guide the generative expert to reconstruct visual semantic tokens of the target image. Extensive experiments across multiple benchmarks demonstrate the effectiveness and superior performance of HBridge, establishing a new paradigm for unified multimodal generation.

cs.CV↗

arXiv2025

Quantitative Morphology of Galactic Cirrus in Deep Optical Imaging

Imaging of optical Galactic cirrus, the spatially resolved form of diffuse Galactic light, provides important insights into the properties of the diffuse interstellar medium (ISM) in the Milky Way. While previous investigations have focused mainly on the intensity characteristics of optical cirrus, their morphological properties remain largely unexplored. In this study, we employ several complementary statistical approaches -- local intensity statistics, angular power spectrum / $Δ$-variance analysis, and wavelet scattering transform analysis -- to characterize the morphology of cirrus in deep optical imaging data. We place our investigation of optical cirrus into a multi-wavelength context by comparing the morphology of cirrus seen with the Dragonfly Telephoto Array to that seen with space-based facilities working at longer wavelengths (Herschel 250 $μm$, WISE 12 $μm$, and Planck radiance), as well as with structures seen in the DHIGLS HI column density map. Our statistical methods quantify the similarities and the differences of cirrus morphology in all these datasets. The morphology of cirrus at visible wavelengths resembles that of far-infrared cirrus more closely than that of mid-infrared cirrus; on small scales, anisotropies in the cosmic infrared background and systematics may lead to differences. Across all dust tracers, cirrus morphology can be well described by a power spectrum with a common power-law index $γ\sim-2.9$. We demonstrate quantitatively that optical cirrus exhibits filamentary, coherent structures across a broad range of angular scales. Our results offer promising avenues for linking the analysis of coherent structures in optical cirrus to the underlying physical processes in the ISM that shape them. Furthermore, we demonstrate that these morphological signatures can be leveraged to distinguish and disentangle cirrus from extragalactic light.

astro-ph.GA↗

arXiv2025

Auxiliary-Free Replica Shadows: Efficient Estimation of Multiple Nonlinear Quantum Properties

Efficient estimation of nonlinear properties is a significant yet challenging task from quantum information processing to many-body physics. Current methodologies often suffer from an exponential sampling cost or require auxiliary qubits and deep quantum circuits. To address these limitations, we propose an efficient auxiliary-free replica shadow (AFRS) framework, which leverages the power of the joint entangling operation on a few input replicas while integrating the mindset of shadow estimation. We rigorously prove that AFRS can offer exponential improvements in estimation accuracy compared with the conventional shadow method, and facilitate the simultaneous estimation of various nonlinear properties, unlike the destructive swap test. Additionally, we introduce an advanced local-AFRS variant tailored to estimating local observables with constant-depth quantum circuits, significantly simplifying the experimental implementation. Our work paves the way for efficient and practical estimation of nonlinear properties on near-term quantum devices.

quant-ph↗

arXiv2025

PSScreen V2: Partially Supervised Multiple Retinal Disease Screening

In this work, we propose PSScreen V2, a partially supervised self-training framework for multiple retinal disease screening. Unlike previous methods that rely on fully labelled or single-domain datasets, PSScreen V2 is designed to learn from multiple partially labelled datasets with different distributions, addressing both label absence and domain shift challenges. To this end, PSScreen V2 adopts a three-branch architecture with one teacher and two student networks. The teacher branch generates pseudo labels from weakly augmented images to address missing labels, while the two student branches introduce novel feature augmentation strategies: Low-Frequency Dropout (LF-Dropout), which enhances domain robustness by randomly discarding domain-related low-frequency components, and Low-Frequency Uncertainty (LF-Uncert), which estimates uncertain domain variability via adversarially learned Gaussian perturbations of low-frequency statistics. Extensive experiments on multiple in-domain and out-of-domain fundus datasets demonstrate that PSScreen V2 achieves state-of-the-art performance and superior domain generalization ability. Furthermore, compatibility tests with diverse backbones, including the vision foundation model DINOv2, as well as evaluations on chest X-ray datasets, highlight the universality and adaptability of the proposed framework. The codes are available at https://github.com/boyiZheng99/PSScreen_V2.

cs.CV↗

arXiv2025

BrainMCLIP: Brain Image Decoding with Multi-Layer feature Fusion of CLIP

Decoding images from fMRI often involves mapping brain activity to CLIP's final semantic layer. To capture finer visual details, many approaches add a parameter-intensive VAE-based pipeline. However, these approaches overlook rich object information within CLIP's intermediate layers and contradicts the brain's functionally hierarchical. We introduce BrainMCLIP, which pioneers a parameter-efficient, multi-layer fusion approach guided by human visual system's functional hierarchy, eliminating the need for such a separate VAE pathway. BrainMCLIP aligns fMRI signals from functionally distinct visual areas (low-/high-level) to corresponding intermediate and final CLIP layers, respecting functional hierarchy. We further introduce a Cross-Reconstruction strategy and a novel multi-granularity loss. Results show BrainMCLIP achieves highly competitive performance, particularly excelling on high-level semantic metrics where it matches or surpasses SOTA(state-of-the-art) methods, including those using VAE pipelines. Crucially, it achieves this with substantially fewer parameters, demonstrating a reduction of 71.7\%(Table.\ref{tab:compare_clip_vae}) compared to top VAE-based SOTA methods, by avoiding the VAE pathway. By leveraging intermediate CLIP features, it effectively captures visual details often missed by CLIP-only approaches, striking a compelling balance between semantic accuracy and detail fidelity without requiring a separate VAE pipeline.

cs.CV↗

arXiv2025

Baking Gaussian Splatting into Diffusion Denoiser for Fast and Scalable Single-stage Image-to-3D Generation and Reconstruction

Existing feedforward image-to-3D methods mainly rely on 2D multi-view diffusion models that cannot guarantee 3D consistency. These methods easily collapse when changing the prompt view direction and mainly handle object-centric cases. In this paper, we propose a novel single-stage 3D diffusion model, DiffusionGS, for object generation and scene reconstruction from a single view. DiffusionGS directly outputs 3D Gaussian point clouds at each timestep to enforce view consistency and allow the model to generate robustly given prompt views of any directions, beyond object-centric inputs. Plus, to improve the capability and generality of DiffusionGS, we scale up 3D training data by developing a scene-object mixed training strategy. Experiments show that DiffusionGS yields improvements of 2.20 dB/23.25 and 1.34 dB/19.16 in PSNR/FID for objects and scenes than the state-of-the-art methods, without depth estimator. Plus, our method enjoys over 5$\times$ faster speed ($\sim$6s on an A100 GPU). Our Project page at https://caiyuanhao1998.github.io/project/DiffusionGS/ shows the video and interactive results. The code and models are publicly available at https://github.com/caiyuanhao1998/Open-DiffusionGS

cs.CV↗

arXiv2025

PSScreen: Partially Supervised Multiple Retinal Disease Screening

Leveraging multiple partially labeled datasets to train a model for multiple retinal disease screening reduces the reliance on fully annotated datasets, but remains challenging due to significant domain shifts across training datasets from various medical sites, and the label absent issue for partial classes. To solve these challenges, we propose PSScreen, a novel Partially Supervised multiple retinal disease Screening model. Our PSScreen consists of two streams and one learns deterministic features and the other learns probabilistic features via uncertainty injection. Then, we leverage the textual guidance to decouple two types of features into disease-wise features and align them via feature distillation to boost the domain generalization ability. Meanwhile, we employ pseudo label consistency between two streams to address the label absent issue and introduce a self-distillation to transfer task-relevant semantics about known classes from the deterministic to the probabilistic stream to further enhance the detection performances. Experiments show that our PSScreen significantly enhances the detection performances on six retinal diseases and the normal state averagely and achieves state-of-the-art results on both in-domain and out-of-domain datasets. Codes are available at https://github.com/boyiZheng99/PSScreen.

cs.CV↗

arXiv2025

HydraRAG: Structured Cross-Source Enhanced Large Language Model Reasoning

Retrieval-augmented generation (RAG) enhances large language models (LLMs) by incorporating external knowledge. Current hybrid RAG system retrieves evidence from both knowledge graphs (KGs) and text documents to support LLM reasoning. However, it faces challenges like handling multi-hop reasoning, multi-entity questions, multi-source verification, and effective graph utilization. To address these limitations, we present HydraRAG, a training-free framework that unifies graph topology, document semantics, and source reliability to support deep, faithful reasoning in LLMs. HydraRAG handles multi-hop and multi-entity problems through agent-driven exploration that combines structured and unstructured retrieval, increasing both diversity and precision of evidence. To tackle multi-source verification, HydraRAG uses a tri-factor cross-source verification (source trustworthiness assessment, cross-source corroboration, and entity-path alignment), to balance topic relevance with cross-modal agreement. By leveraging graph structure, HydraRAG fuses heterogeneous sources, guides efficient exploration, and prunes noise early. Comprehensive experiments on seven benchmark datasets show that HydraRAG achieves overall state-of-the-art results on all benchmarks with GPT-3.5-Turbo, outperforming the strong hybrid baseline ToG-2 by an average of 20.3% and up to 30.1%. Furthermore, HydraRAG enables smaller models (e.g., Llama-3.1-8B) to achieve reasoning performance comparable to that of GPT-4-Turbo. The source code is available on https://stevetantan.github.io/HydraRAG/.

cs.CL↗

arXiv2025

MMAPG: A Training-Free Framework for Multimodal Multi-hop Question Answering via Adaptive Planning Graphs

Multimodal Multi-hop question answering requires integrating information from diverse sources, such as images and texts, to derive answers. Existing methods typically rely on sequential retrieval and reasoning, where each step builds on the previous output. However, this single-path paradigm makes them vulnerable to errors due to misleading intermediate steps. Moreover, developing multimodal models can be computationally expensive, often requiring extensive training. To address these limitations, we propose a training-free framework guided by an Adaptive Planning Graph, which consists of planning, retrieval and reasoning modules. The planning module analyzes the current state of the Adaptive Planning Graph, determines the next action and where to expand the graph, which enables dynamic and flexible exploration of reasoning paths. To handle retrieval of text to unspecified target modalities, we devise modality-specific strategies that dynamically adapt to distinct data types. Our approach preserves the characteristics of multimodal information without costly task-specific training, enabling seamless integration with up-to-date models. Finally, the experiments on MultimodalQA and WebQA show that our approach matches or outperforms existing models that rely on training.

cs.AI↗

arXiv2025

Superposition Property in Disjoint Variables for the Infinity Laplace Equation

We establish a superposition principle in disjoint variables for the inhomogeneous infinity-Laplace equation. We show that the sum of viscosity solutions of the inhomogeneous infinity-Laplace equation in separate domains is a viscosity solution in the product domain. This result has been used in the literature with certain particular choices of solutions to simplify regularity analysis for a general inhomogeneous infinity-Laplace equation by reducing it to the case without sign-changing inhomogeneous terms and vanishing gradient singularities. We present a proof of this superposition principle for general viscosity solutions. We also explore generalization in metric spaces using cone comparison techniques and study related properties for general elliptic and convex equations.

math.AP↗

arXiv2025

M-learner:A Flexible And Powerful Framework To Study Heterogeneous Treatment Effect In Mediation Model

We propose a novel method, termed the M-learner, for estimating heterogeneous indirect and total treatment effects and identifying relevant subgroups within a mediation framework. The procedure comprises four key steps. First, we compute individual-level conditional average indirect/total treatment effect Second, we construct a distance matrix based on pairwise differences. Third, we apply tSNE to project this matrix into a low-dimensional Euclidean space, followed by K-means clustering to identify subgroup structures. Finally, we calibrate and refine the clusters using a threshold-based procedure to determine the optimal configuration. To the best of our knowledge, this is the first approach specifically designed to capture treatment effect heterogeneity in the presence of mediation. Experimental results validate the robustness and effectiveness of the proposed framework. Application to the real-world Jobs II dataset highlights the broad adaptability and potential applicability of our method.Code is available at https: //anonymous.4open.science/r/M-learner-C4BB.

stat.ML↗

arXiv2025

GPT-IMAGE-EDIT-1.5M: A Million-Scale, GPT-Generated Image Dataset

Recent advancements in large multimodal models like GPT-4o have set a new standard for high-fidelity, instruction-guided image editing. However, the proprietary nature of these models and their training data creates a significant barrier for open-source research. To bridge this gap, we introduce GPT-IMAGE-EDIT-1.5M, a publicly available, large-scale image-editing corpus containing more than 1.5 million high-quality triplets (instruction, source image, edited image). We systematically construct this dataset by leveraging the versatile capabilities of GPT-4o to unify and refine three popular image-editing datasets: OmniEdit, HQ-Edit, and UltraEdit. Specifically, our methodology involves 1) regenerating output images to enhance visual quality and instruction alignment, and 2) selectively rewriting prompts to improve semantic clarity. To validate the efficacy of our dataset, we fine-tune advanced open-source models on GPT-IMAGE-EDIT-1.5M. The empirical results are exciting, e.g., the fine-tuned FluxKontext achieves highly competitive performance across a comprehensive suite of benchmarks, including 7.24 on GEdit-EN, 3.80 on ImgEdit-Full, and 8.78 on Complex-Edit, showing stronger instruction following and higher perceptual quality while maintaining identity. These scores markedly exceed all previously published open-source methods and substantially narrow the gap to leading proprietary models. We hope the full release of GPT-IMAGE-EDIT-1.5M can help to catalyze further open research in instruction-guided image editing.

cs.CV↗

arXiv2025

Candidate Dark Galaxy-2: Validation and Analysis of an Almost Dark Galaxy in the Perseus Cluster

Candidate Dark Galaxy-2 (CDG-2) is a potential dark galaxy consisting of four globular clusters (GCs) in the Perseus cluster, first identified in Li et al. (2025) through a sophisticated statistical method. The method searched for over-densities of GCs from a \textit{Hubble Space Telescope} (\textit{HST}) survey targeting Perseus. Using the same \textit{HST} images and the new imaging data from the \textit{Euclid} survey, we report the detection of extremely faint but significant diffuse emission around the four GCs of CDG-2. We thus have exceptionally strong evidence that CDG-2 is a galaxy. This is the first galaxy detected purely through its GC population. Under the conservative assumption that the four GCs make up the entire GC population, preliminary analysis shows that CDG-2 has a total luminosity of $L_{V, \mathrm{gal}}= 6.2\pm{3.0} \times 10^6 L_{\odot}$ and a minimum GC luminosity of $L_{V, \mathrm{GC}}= 1.03\pm{0.2}\times 10^6 L_{\odot}$. Our results indicate that CDG-2 is one of the faintest galaxies having associated GCs, while at least $\sim 16.6\%$ of its light is contained in its GC population. This ratio is likely to be much higher ($\sim 33\%$) if CDG-2 has a canonical GC luminosity function (GCLF). In addition, if the previously observed GC-to-halo mass relations apply to CDG-2, it would have a minimum dark matter halo mass fraction of $99.94\%$ to $99.98\%$. If it has a canonical GCLF, then the dark matter halo mass fraction is $\gtrsim 99.99\%$. Therefore, CDG-2 may be the most GC dominated galaxy and potentially one of the most dark matter dominated galaxies ever discovered.

astro-ph.GA↗