Searcharxiv⌕ Search

arXiv subjects

Jiale Huang

Publications and source records attributed to Jiale Huang.

At least 19 recordsLinked to original sources

From Trajectories to Prefixes: Reusing Teacher Trajectories via Replayed Prefixes and Online Continuation

Small language models are attractive backbones for interactive agents, but direct distillation from strong teacher trajectories often turns rich multi-turn behavior into one-shot imitation targets. This is inefficient in long-horizon environments, where early decisions shape later states and rewards. We propose Prefix-GRPO, a reinforcement learning framework that decomposes teacher trajectories into replay-aligned prefix queries and online continuations. Each prefix is replayed in the environment to recover a valid intermediate state, after which the student continues online interaction and receives task reward. Unlike response-only GRPO, Prefix-GRPO also applies clipped policy updates to historical assistant tokens inside the replayed prefix, using a policy-distilled SFT checkpoint to estimate their old log-probabilities. This unifies prefix learning and continuation learning within the same policy-optimization form. Experiments on TextCraft, BabyAI, and ALFWorld show that Prefix-GRPO improves small-model agents over distillation and standard RL baselines, while ablations show that replay alone is insufficient without explicit prefix-token optimization. The implementation and reproduction scripts are available at https://github.com/HappynessI/Prefix_GRPO.

cs.LG↗

SetGRPO: Calibrating Grader-Induced Policy Optimization

Rubric-based reinforcement learning uses fine-grained grader feedback to train large language models on open-ended tasks, but grader randomness can alter GRPO updates in ways that score discrepancies alone cannot characterize. Because GRPO jointly normalizes rewards within each response group and combines the resulting training coefficients with response-level gradients, a scoring discrepancy can shift the coefficients of the entire group and induce policy-dependent update changes. Selective execution further affects the direction-reversal risk of retained updates. To address this, we propose Set-Valued Group Relative Policy Optimization (SetGRPO). It comprises three modules: (a) Joint Reward Set Calibration constructs a joint uncertainty set for the training coefficients; (b) Geometry-Guided Safe Recovery maps candidate coefficients into the current policy's local update geometry and determines the largest safe recovery ratio for each response group; and (c) Selective Risk-Gated Optimization selects a global threshold through independent risk calibration to control the conditional direction-reversal risk of accepted updates while preserving the original GRPO objective. Under exchangeability, SetGRPO provides finite-sample marginal coverage for the joint coefficient set and simultaneous risk control across candidate thresholds. On ten open-ended benchmarks spanning three domains, SetGRPO achieves the highest macro average among all compared training methods on both Qwen3.5 and Llama3.2 model families.

cs.LG↗

TopoField: Technical Report for ECCV 2026 SceneFun3D Challenge

Functional segmentation requires locating small interactive parts, such as knobs, buttons, and sockets, in indoor point clouds and returning confidence-ranked instance masks. A point-wise semantic predictor produces a continuous probability field, whereas the benchmark evaluates a discrete set of scored instances. This creates a substantial field-to-instance gap. We present TopoField, a deterministic decoder built on top of VOLT. TopoField first converts 13 test-time forward passes into a semantic consensus field, then decodes instance candidates from the consensus geometry, and finally ranks the candidates in a class-aware manner. The augmentation stage retains the processed-point identity of every fragment prediction, accumulates probabilities in a common point array, normalizes by coverage count, and fuses the resulting caches. The decoding stage separates high-confidence cores from broader support regions, constructs a symmetric radius-pruned point graph, and uses a multi-source graph-geodesic partition to split support regions that contain multiple seeds. On 561 ground-truth instances across 30 validation scenes, an IoU-0.5-valid candidate exists for 67.2% of instances, and top-1 coverage is 20.0%. These complementary measurements motivate joint attention to candidate construction and ranking. TopoField achieves 28.286% AP, 41.383% AP50, and 47.558% AP25. TopoField achieved second place on the ECCV 2026 SceneFun3D Challenge.

cs.CV↗

GradCO: Behavior-Aware Gradient-Conflict Offloading for Text-Based Person Anomaly Search

Text-based Person Anomaly Search (TPAS) retrieves a target pedestrian using natural-language descriptions of appearance and normal or anomalous behavior. Unlike conventional text-based person retrieval, TPAS must preserve identity cues while distinguishing fine-grained behaviors. It presents two challenges: same-identity, different-behavior hard negatives cause gradient conflict during granular alignment, while strong appearance cues can obscure weak behavioral evidence during aggregation. We propose GradCO (behavior-aware Gradient-Conflict Offloading), comprising three complementary components. Behavior-Aware Mixture-of-Experts (B-MoE) forms appearance, normal-behavior, and anomalous-behavior streams whose roles are learned from retrieval objectives without expert-specific semantic labels. Conflict Offload Learning (COL) uses pair-conditioned offload anchors in similarity computation to reduce the conflicting-gradient burden on the main text representations. Hierarchical Intent Alignment (HIA) aggregates expert similarities and features using query-conditioned weights. Experiments on five TPAS/TIPR benchmarks show GradCO significantly surpasses SOTA methods and validate our design.

cs.CV↗

SWE-at-Scale: Scalable Issue Synthesis from Historical Bug Fixes

As coding agents become increasingly capable, robust evaluation requires a broader and more diverse supply of verifiable software engineering (SWE) tasks. However, real-world tasks are constrained by the natural frequency of issue reporting and the substantial cost of manual curation. Furthermore, existing synthetic approaches rely primarily on the pretrained knowledge of large models coupled with heuristic prompting, often failing to capture the authentic, complex mechanics of software failures. To exponentially scale SWE task production, we introduce SWE-at-Scale, a framework that automatically synthesizes realistic issues by reverse-engineering real-world software fixes. Our key insight is that historical fixes encapsulate deep mechanistic knowledge of bug formation: inverting a fix transforms repair supervision into bug-induction supervision grounded in actual development scenarios. SWE-at-Scale transfers these learned bug patterns to novel code contexts, repeatedly sampling diverse bug typologies to construct executable SWE tasks. Extensive experiments show that SWE-at-Scale generates more diverse and realistic SWE tasks than current synthesis baselines while posing greater challenges to state-of-the-art coding agents. Notably, SWE-at-Scale can generate distinct bug types from a single code context, with bug diversity scaling robustly even up to 200 generations per file. Overall, SWE-at-Scale turns finite real-world fixes into reusable supervision for scalable synthesis of realistic, diverse, and challenging SWE tasks.

cs.SE↗

Scale-QLoRA: Code-Invariant Adapter Merging for Native 4-bit Microscaling LLMs

Merging a LoRA adapter into its base model is standard deployment practice: it removes the runtime adapter's per-forward overhead and leaves a single standalone checkpoint any serving stack can load. On a native 4-bit microscaling checkpoint (NVFP4, MXFP4) that step stops being free. The merged weights must be written back through a quantizer, which re-derives the checkpoint's discrete E2M1 code plane (roughly 90% of the artifact's bytes), so the deployed artifact becomes coupled to one quantization convention, and every later code-touching event in its lifecycle can move it. Done naively the step is worse than fragile: it deletes the adaptation, by up to 39 pp, because against an already-on-grid base the reconstruction optimum is that base. Scale-QLoRA instead adapts only the native per-block scale field, trains those scales on the deployment grid, and freezes every E2M1 code. Within a fixed native format, scale grid, block layout and code plane, merging is then a bit-exact identity and the merged artifact is code-invariant. Across four models and four tasks, Scale-QLoRA and merge-aware QAT-LoRA are both accuracy-lossless, so we claim no accuracy ordering between them; they differ structurally, in that QAT-LoRA re-derives the code plane through a quantizer while Scale-QLoRA preserves it exactly. That difference is what the lifecycle prices: nearest-rounding implementations disagree by about a point on the measured task, and more extreme rule mismatches can drive the weight-space artifact to ~0%, which we report as a sensitivity bound rather than a deployment frequency. Preserving the code plane also drops the weight-space straight-through estimator from training (3.9x per step on the dense 8B model) and enables exact rollback, code-plane deduplication, and a ~125x faster scale-only task swap.

cs.CL↗

Repo0: Design-Driven Zero-to-All Code Generation

Large language model agents have made substantial progress in code generation, yet most existing systems assume a predefined repository architecture. This assumption does not hold in zero-to-all code generation, where an agent must construct an entire software project directly from natural-language requirements while maintaining a modular repository architecture throughout development. We present Repo0, a continuous structural evolution framework for zero-to-all code generation. Repo0 maintains an explicit architectural state instantiated as a Dual-Directed-Acyclic-Graph (Dual-DAG), consisting of a requirement-level DAG, a component-level DAG, and their alignment relation. Starting from natural-language requirements, it iteratively evolves component boundaries through structural actions guided by modularity metrics until structural convergence, after which the converged architecture guides test-driven development code generation. We evaluate Repo0 on six real-world repositories from RepoCraft using GPT-5 mini and DeepSeek V3.2. Repo0 achieves the highest Functionality Coverage and Pass Rate across all settings. Compared with RPG, the strongest repository-planning baseline, Repo0 improves Functionality Coverage by up to 20.08 percentage points and Pass Rate by up to 29.74 percentage points. Ablation and structural-evolution analyses further demonstrate the importance of the Dual-DAG architectural state, modularity-guided structural evolution, and explicit structural convergence.

cs.SE↗

RankVR: Low-Rank Structure Perception and Value Recalibration for Robust Composed Image Retrieval

Composed Image Retrieval (CIR) constitutes a pivotal paradigm requiring models to perform joint reasoning on reference images and modification texts. However, the prevalence of Noisy Triplet Correspondence (NTC) in large-scale datasets severely constrains model performance. Existing denoising methods either target binary mismatches or rely on scalar-based point-wise estimation, neglecting rich global structural correlations among sample populations and dynamic value variations during training, thereby yielding suboptimal results. This paper identifies two critical unresolved challenges: Global Structural Inconsistency of Semantic Correlations and Hard Sample Discrimination Uncertainty. To address these, we propose RankVR, a framework designed to construct a robust CIR model via global structure consistency and dynamic value perception. Specifically, we introduce the Global Structure Consistency Perception (GSCP) module, which utilizes the Effective Rank of the Correlation Matrix to decouple clean samples from structural noise. By measuring rank difference, GSCP identifies samples disrupting macroscopic semantic symmetry. Furthermore, we develop the Adaptive Semantic Value Calibration (ASVC) module to distinguish high-value hard clean samples. By integrating training potential and reliability, it dynamically quantifies the semantic value of each triplet, ensuring effective utilization of hard samples while suppressing noise characterized by logical conflicts. Extensive experiments on the FashionIQ and CIRR benchmark datasets demonstrate that RankVR significantly outperforms existing state-of-the-art methods, validating its superior robustness in noisy environments.

cs.CV↗

IMAGINE: Adaptive Schema-Imagery Enhanced Composition for Composed Video Retrieval

Composed Video Retrieval (CVR) is designed to retrieve a target video that matches a reference video modified by a modification text. While existing methods explore cross-modal correspondences, they often assume modified objects appear directly in videos. However, modification texts frequently describe concepts not explicitly presented but implicitly expressed through semantically related visual cues (e.g., "cake" implying "birthday party"). Current approaches typically rely on aligning explicit feature representations within the concrete space, neglecting critical latent associations. To address this, we propose an adaptIve scheMa-ImAGery enhanced composItional NEtwork (IMAGINE). Unlike standard explicit matching, IMAGINE materializes implicit semantics (termed schema imagery) via dynamic multimodal prototypes. These prototypes capture shared latent concepts to adaptively modulate visual features, effectively injecting implicit guidance into the retrieval process. By bridging the gap between explicit visual contents and implicit retrieval intentions, IMAGINE achieves state-of-the-art performance in both CVR and Composed Image Retrieval (CIR) across three widely used benchmarks.

cs.CV↗

INTENT: Invariance and Discrimination-aware Noise Mitigation for Robust Composed Image Retrieval

Composed Image Retrieval (CIR) is a challenging image retrieval paradigm that enables to retrieve target images based on multimodal queries consisting of reference images and modification texts. Although substantial progress has been made in recent years, existing methods assume that all samples are correctly matched. However, in real-world scenarios, due to high triplet annotation costs, CIR datasets inevitably contain annotation errors, resulting in incorrectly matched triplets. To address this issue, the problem of Noisy Triplet Correspondence (NTC) has attracted growing attention. We argue that noise in CIR can be categorized into two types: cross-modal correspondence noise and modality-inherent noise. The former arises from mismatches across modalities, whereas the latter originates from intra-modal background interference or visual factors irrelevant to the coarse-grained modification annotations. However, modality-inherent noise is often overlooked, and research on cross-modal correspondence noise remains nascent. To tackle above issues, we propose the Invariance and discrimiNaTion-awarE Noise neTwork (INTENT), comprising two components: Visual Invariant Composition and Bi-Objective Discriminative Learning, specifically designed to handle the two-aspect noise. The former applies causal intervention on the visual side via Fast Fourier Transform (FFT) to generate intervened composed features, enforcing visual invariance and enabling the model to ignore modality-inherent noise during composition. The latter adopts collaborative optimization with both positive and negative samples, and constructs a scalable decision boundary that dynamically adjusts decisions based on the loyalty degree, enabling robust correspondence discrimination. Extensive experiments on two widely used benchmark datasets demonstrate the superiority and robustness of INTENT.

cs.CV↗

Dismagicker: Unitary Gate for Non-Stabilizerness Reduction

We introduce the notion of dismagicker: non-Clifford unitary gate designed to reduce the non-stabilizerness (also called magic) of quantum many-body states. Although both entanglement and non-stabilizerness are fundamental quantum resources, they require distinct control strategies. While disentanglers (unitary operations that lower entanglement) are well-established in tensor network methods, analogous concept for non-stabilizerness suppression has been largely missing. In this work, we define dismagicker as non-Clifford unitary operation that actively suppresses non-stabilizerness, steering states toward classically simulatable stabilizer states. We develop optimization method for constructing dismagickers within the Matrix Product States framework. Our numerical results show that the non-stabilizerness reduction procedure, when combined with entanglement reduction steps with Clifford circuits, significantly improves the accuracy for both classical simulation of many-body systems and quantum state preparation on quantum devices. Dismagicker enriches our toolkit for the manipulation of many-body states by unifying non-stabilizerness and entanglement reduction.

quant-ph↗

Floquet-induced bosonic pair condensate with unconventional symmetry

In this study, we propose a dynamical pairing mechanism other than the pair-wise interactions. Starting from a two-dimensional hard-core boson model with periodically modulated hopping amplitude, we derive an effective Floquet Hamiltonian with three-site interactions that are responsible for unconventional pairing between adjacent bosons. By performing a density matrix renormalization group study on this three-site interacting Hamiltonian, we reveal a bosonic pair condensate with $s+id$ wave symmetry, while the single-particle Bose-Einstein condensate is completely depleted. The experimental implementations of the proposed model on superconducting quantum circuit have also been discussed.

cond-mat.quant-gas↗

Augmenting Density Matrix Renormalization Group with Matchgates and Clifford circuits

Matchgates and Clifford circuits are two types of quantum circuits which can be efficiently simulated classically, though the underlying reasons are quite different. Matchgates are essentially the single particle basis transformations in the Majorana fermion representation which can be easily handled classically, while the Clifford circuits can be efficiently simulated using the tableau method according to the Gottesman-Knill theorem. In this work, we propose a new wave-function ansatz in which matrix product states are augmented with the combination of Matchgates and Clifford circuits (dubbed MCA-MPS) to take advantage of the representing power of all of them. Moreover, the optimization of MCA-MPS can be efficiently implemented within the Density Matrix Renormalization Group method. Our benchmark results on one-dimensional hydrogen chain show that MCA-MPS can improve the accuracy of the ground-state calculation by several orders of magnitude over MPS with the same bond dimension. This new method provides us a useful approach to study quantum many-body systems. The MCA-MPS ansatz also expands our understanding of classically simulatable quantum many-body states.

quant-ph↗

LongCat-Image Technical Report

We introduce LongCat-Image, a pioneering open-source and bilingual (Chinese-English) foundation model for image generation, designed to address core challenges in multilingual text rendering, photorealism, deployment efficiency, and developer accessibility prevalent in current leading models. 1) We achieve this through rigorous data curation strategies across the pre-training, mid-training, and SFT stages, complemented by the coordinated use of curated reward models during the RL phase. This strategy establishes the model as a new state-of-the-art (SOTA), delivering superior text-rendering capabilities and remarkable photorealism, and significantly enhancing aesthetic quality. 2) Notably, it sets a new industry standard for Chinese character rendering. By supporting even complex and rare characters, it outperforms both major open-source and commercial solutions in coverage, while also achieving superior accuracy. 3) The model achieves remarkable efficiency through its compact design. With a core diffusion model of only 6B parameters, it is significantly smaller than the nearly 20B or larger Mixture-of-Experts (MoE) architectures common in the field. This ensures minimal VRAM usage and rapid inference, significantly reducing deployment costs. Beyond generation, LongCat-Image also excels in image editing, achieving SOTA results on standard benchmarks with superior editing consistency compared to other open-source works. 4) To fully empower the community, we have established the most comprehensive open-source ecosystem to date. We are releasing not only multiple model versions for text-to-image and image editing, including checkpoints after mid-training and post-training stages, but also the entire toolchain of training procedure. We believe that the openness of LongCat-Image will provide robust support for developers and researchers, pushing the frontiers of visual content creation.

cs.CV↗

Refractive Index Engineering: Insights from Biological Systems for Advanced Optical Design

This review explores the innovative design to achieve advanced optical functions in natural materials and intricate optical systems inspired by the unique refractive index profiles found in nature. By understanding the physical principles behind biological structures, we can develop materials with tailored optical properties that mimic these natural systems. One key area discussed is biomimetic materials design, where biological systems such as apple skin and the vision system inspire new materials. Another focus is on intricate optical systems based on refractive index contrast. These principles can be extended to design devices like waveguides, photonic crystals, and metamaterials, which manipulate light in novel ways. Additionally, the review covers optical scattering engineering, which is crucial for biomedical imaging. By adjusting the real and imaginary parts of the refractive index, we can control how much light is scattered and absorbed by tissues. This is particularly important for techniques like optical coherence tomography and multiphoton microscopy, where tailored scattering properties can improve imaging depth and resolution. The review also discusses various techniques for measuring the refractive index of biological tissues which provide comprehensive insights into the optical properties of biological materials, facilitating the development of advanced biomimetic designs. In conclusion, the manipulation of refractive index profiles in biological systems offers exciting opportunities for technological advancements. By drawing inspiration from nature and understanding the underlying physical principles, we can create materials and devices with enhanced performance and new functionalities. Future research should focus on further exploring these principles and translating them into practical applications to address real-world challenges.

physics.bio-ph↗

MADiff: Text-Guided Fashion Image Editing with Mask Prediction and Attention-Enhanced Diffusion

Text-guided image editing model has achieved great success in general domain. However, directly applying these models to the fashion domain may encounter two issues: (1) Inaccurate localization of editing region; (2) Weak editing magnitude. To address these issues, the MADiff model is proposed. Specifically, to more accurately identify editing region, the MaskNet is proposed, in which the foreground region, densepose and mask prompts from large language model are fed into a lightweight UNet to predict the mask for editing region. To strengthen the editing magnitude, the Attention-Enhanced Diffusion Model is proposed, where the noise map, attention map, and the mask from MaskNet are fed into the proposed Attention Processor to produce a refined noise map. By integrating the refined noise map into the diffusion model, the edited image can better align with the target prompt. Given the absence of benchmarks in fashion image editing, we constructed a dataset named Fashion-E, comprising 28390 image-text pairs in the training set, and 2639 image-text pairs for four types of fashion tasks in the evaluation set. Extensive experiments on Fashion-E demonstrate that our proposed method can accurately predict the mask of editing region and significantly enhance editing magnitude in fashion image editing compared to the state-of-the-art methods.

cs.CV↗

FashionFAE: Fine-grained Attributes Enhanced Fashion Vision-Language Pre-training

Large-scale Vision-Language Pre-training (VLP) has demonstrated remarkable success in the general domain. However, in the fashion domain, items are distinguished by fine-grained attributes like texture and material, which are crucial for tasks such as retrieval. Existing models often fail to leverage these fine-grained attributes from both text and image modalities. To address the above issues, we propose a novel approach for the fashion domain, Fine-grained Attributes Enhanced VLP (FashionFAE), which focuses on the detailed characteristics of fashion data. An attribute-emphasized text prediction task is proposed to predict fine-grained attributes of the items. This forces the model to focus on the salient attributes from the text modality. Additionally, a novel attribute-promoted image reconstruction task is proposed, which further enhances the fine-grained ability of the model by leveraging the representative attributes from the image modality. Extensive experiments show that FashionFAE significantly outperforms State-Of-The-Art (SOTA) methods, achieving 2.9% and 5.2% improvements in retrieval on sub-test and full test sets, respectively, and a 1.6% average improvement in recognition tasks.

cs.CV↗

Clifford circuits Augmented Matrix Product States for fermion systems

Clifford circuits Augmented Matrix Product States (CAMPS) was recently proposed to leverage the advantages of both Clifford circuits and Matrix Product States (MPS). Clifford circuits can support large entanglement and can be efficiently simulated classically according to the Gottesman-Knill theorem. So in CAMPS, MPS needs only to handle the so-called Non-stabilizerness Entanglement Entropy which significantly improves the simulation accuracy for a given bond dimension. In this work, we generalize CAMPS to study the Fermion system by taking advantage of the Jordan-Wigner transformation which can map the studied Fermion system to a spin system. We benchmark the method on both the spinless $t-V$ model and the spinful Hubbard model. Our test results show significant improvement of the accuracy of CAMPS over MPS, especially when the interactions are strong. Fermionic CAMPS provides a useful tool for the accurate study of many-body fermion systems in the future and has the potential to help resolve long-standing issues.

cond-mat.str-el↗