Search arXiv⌕ Search

arXiv subjects

Jiachen Qian

Publications and source records attributed to Jiachen Qian.

At least 19 recordsLinked to original sources
arXiv2026

Algebraic flat bands and even central-band Chern numbers in a seven-orbital hopping model

Turning a flat-band parent into an isolated Chern band requires supervise of destructive interference, hopping phases, and the cell translations of the bonds. We study this connection in a seven-orbital 2 times 2 supercell honeycomb-vacancy lattice, combining analytic reductions with Bloch spectra, lattice Chern invariants, and quantum geometry. The parent lattice has three exact flat branches, a smooth periodic zero mode, and a compact localized state. A uniform gap bound shows that faint intracell-loop hopping preserves two topologically trivial insulating gaps. The intracell hopping graph also admits a staggered chiral coloring: reversing site amplitudes according to this coloring links opposite energies at momenta separated by half a reciprocal period. This constraint restricts the isolated central band to even Chern numbers and forbids a zero-energy insulating filling at every loop phase. Purely imaginary loop hopping admits a real gauge, whereas other phases allow paired band contacts connecting central bands with C3=2 and C3=4. On a specified strong-loop cut, third-order masses open the direct gaps; these gaps shrink faster than the bandwidth as the coupling grows. Changing the bond connectivity to an oriented honeycomb next-nearest-neighbor extension generates a narrower Chern band with nonuniform quantum geometry. These results distinguish individual-band topology from occupied-group topology and reveal how bond connectivity and cell translations determine which topological bands can emerge from an exactly flat parent.

cond-mat.mes-hall↗
arXiv2026

RECAP: Looking Once Is Not Enough for Vision-Language Reasoning

Long chain-of-thought reasoning improves vision-language models (VLMs), but it also exposes a temporal grounding failure: as generation unfolds, models may drift from image evidence while reinforcement learning still provides only a final-answer reward. We study this gap through visual revisit, a temporally localized reactivation of image evidence during reasoning. Across three VLM backbones, revisit peaks are predictive of correctness and causally linked to performance: masking them degrades accuracy more than masking non-peaks, while injecting revisit-like peaks into failed rollouts partially restores correct answers. Control analyses show that this signal is not explained by raw attention strength, uncertainty, logit margin, or hidden-state norm. We propose RECAP, a GRPO-compatible training method that turns detached revisit traces into a credit-assignment signal. RECAP uses revisit-conditioned advantage estimation, dynamic discounting, and gated value prediction to propagate final rewards through visually important reasoning steps, without extra rewards, additional rollouts, or inference-time changes. Across Qwen2.5-VL-7B, InternVL3-8B, and LLaVA-OV-7B, RECAP consistently improves over GRPO and VPPO, yielding $+6.1$--$+6.5$ HallusionBench gains over the SFT base while preserving reasoning and text-only performance. Results on larger LoRA-tuned models, Chinese zero-shot benchmarks, visual perturbations, and human evaluation further demonstrate robust improvements in image-grounded reasoning.

cs.CV↗
arXiv2026

How Order-Sensitive Are LLMs? OrderProbe for Deterministic Structural Reconstruction

Large language models (LLMs) excel at semantic understanding, yet their ability to reconstruct internal structure from scrambled inputs remains underexplored. Sentence-level restoration is difficult to evaluate automatically because scrambled sentences often admit multiple valid reorderings. We introduce OrderProbe, a deterministic benchmark for structural reconstruction using fixed four-character expressions in Chinese, Japanese, and Korean, which have a unique canonical order and thus support exact-match scoring. We further propose a diagnostic framework that evaluates models beyond recovery accuracy, including Semantic Accuracy, Logical Validity, Structural Consistency, Robustness, and Information Density. Experiments on twelve widely used LLMs show that structural reconstruction remains difficult even for frontier systems: zero-shot recovery frequently falls below 35%. We also observe a consistent gap between meaning-oriented generation and exact structural reconstruction, suggesting that structural robustness is not an automatic byproduct of semantic competence.

cs.CL↗
arXiv2026

JPO: Juris Policy Optimization for Structured Legal Reasoning in Criminal Judgment Prediction

Criminal judgment prediction requires models to infer statutory articles, charges, and sentencing outcomes from case facts. Unlike standard classification tasks, it involves a structured reasoning process in which statutes should be matched with facts, charges should be justified by statutes, and sentencing outcomes should remain consistent with charges. Existing approaches optimize final labels, and while some have attempted to evaluate reasoning quality, their evaluations are indirect, often relying on LLM-generated rubrics that reflect model-internal preferences rather than the inherent logical structure of legal adjudication. We propose Juris Policy Optimization (JPO), a post-training framework for structured legal reasoning in Chinese criminal judgment prediction. JPO first uses teacher-generated rationales to supervise a standardized four-step reasoning process, and then applies reinforcement learning with a composite reward over legal prediction quality, reasoning structure completeness, and cross-step consistency. JPO further introduces token-level advantage reweighting and adaptive clipping for legally salient reasoning segments. Experiments on multiple open-source language models and three Chinese legal benchmarks show that JPO consistently improves both judgment prediction and reasoning quality over supervised fine-tuning and reinforcement learning baselines.

cs.CL↗
arXiv2026

Prosody-driven Jailbreaks in Audio LLMs: A Controlled Study and Mechanistic Analysis

Audio-capable foundation models enable end-to-end spoken interaction, but they also introduce safety risks beyond transcript content. It remains unclear how much jailbreak capability can arise from matched-text variation in speech delivery rather than from lexical rewriting or broader style transfer. We study this question by holding transcript content fixed and varying six speech-delivery presets whose acoustic attributes may co-vary. We present PJ-Break, a black-box evaluation protocol with presets targeting arousal, authority, and speaking rate, together with AdvAudio-Prosody, a 600-sample benchmark with acoustically verified attributes. On the exact post-QC Qwen2-Audio panel, the Q=1 Panic (38/95), Anger (35/95), and Fast (32/95) presets are all well above Neutral (4/95). The fixed six-query pool covers 44/95 Qwen2-Audio seeds and 15/95 GPT-4o seeds and exceeds a matched-budget StyleBreak reimplementation (27/95) on Qwen2-Audio. A same-voice pool excluding the confounded Commanding condition still reaches 40/95, and a retained-panel ablation shows emotional-delivery audio alone (44/95) is far more effective than emotional text alone (11/95). Exploratory surrogate diagnostics and pilot mitigation observations are secondary, non-core analyses. Overall, matched-text speech delivery should be treated as a first-class factor in Audio LLM safety evaluation

cs.SD↗
arXiv2026

VisCritic: Visual State Comparison as Process Reward for GUI Agents

GUI agents powered by vision-language models show strong potential for automating digital tasks, yet frequently fail in long-horizon scenarios due to the absence of step-level verification. Existing process reward models verify actions through textual reasoning alone, missing the visual nature of GUI state changes. We introduce VisCritic, a visual process reward framework that verifies agent actions by directly comparing pre-action and post-action screenshots in visual feature space. VisCritic employs a Siamese vision transformer to extract change-aware representations, coupled with an Action-Aware Critic Head that jointly evaluates action success, task progress, and error type. A critic-training data construction pipeline generates weakly supervised samples from existing trajectories without additional human labels for critic training. Experiments and offline analyses across five benchmarks demonstrate that VisCritic serves as a plug-and-play enhancement for diverse GUI agents, generally improving benchmark metrics while providing visual diagnostic cues.

cs.CV↗
arXiv2026

JECA^2: Judgment-Explanation Consistent Adversarial Attack against Forensic Vision-Language Models

Forensic vision-language models (VLMs) have recently been developed to detect image tampering and provide natural-language explanations. However, their robustness against adversarial manipulation remains underexplored. Existing adversarial attacks typically aim to flip the model's binary judgment, while the accompanying explanation may still reveal forensic cues and contradict the attacked judgment. In this paper, we study judgment-explanation consistent adversarial attacks against forensic VLMs and propose JECA^2, a controlled white-box red-team diagnostic that jointly redirects visual attribution and aligns textual explanations with the target judgment. On the visual side, JECA^2 uses Grad-CAM-guided perturbations to divert attribution from tampered regions toward benign regions. On the textual side, it optimizes prompt embeddings toward authenticity-affirming semantics under a token-proximity constraint. Experiments on forensic VLM benchmarks show that JECA^2 achieves higher attack success and automated judgment-explanation consistency than implemented baselines under white-box threat settings, while transfer to closed-source VLMs remains measurable but limited. Our results highlight a consistency failure mode in explanation-based forensic VLMs and motivate future robustness evaluation beyond binary detection accuracy.

cs.CV↗
arXiv2026

SilentRetrieval: Hijacking Retrieval-Augmented Generation via Semantically-Preserving Adversarial Data Poisoning

Retrieval-Augmented Generation (RAG) mitigates LLM hallucinations but introduces a critical vulnerability: corpus integrity. We present SilentRetrieval, a two-stage data poisoning attack that hijacks RAG systems through adversarially crafted yet fluent documents. Stage 1 uses Coordinated Beam Search, a multi-token joint optimization method with a fluency-similarity objective, to keep a poisoned host document retrievable while constraining perplexity. Stage 2 uses Context-Adaptive Trigger Generation, a lightweight trigger-fusion step driven by a frozen LLM, to integrate manipulation triggers into document content. Under a one-poisoned-document-per-query evaluation with synthetic target answers, SilentRetrieval achieves 84.6%/81.3% HR@10 and 57.5%/54.8% ASR-LLM on Natural Questions and MS MARCO, while maintaining near-benign perplexity. Cross-model evaluation across four target LLMs shows nontrivial effectiveness under a fixed trigger generator, and transfer tests against unseen retrievers, including ColBERT and commercial embedding models, yield 64.7% average HR@10 under the same injected-corpus protocol. In a sampled Wikipedia-scale evaluation, SilentRetrieval retains 74.2% HR@10 at a 0.016% poisoning ratio. Combined retrieval-side and generation-side defenses reduce attack success substantially but incur a latency trade-off. Human evaluation shows substantially lower flag rates than disfluent baselines, while remaining numerically more suspicious than benign content at the current sample size.

cs.CR↗
arXiv2026

Visual Inception: Compromising Long-term Planning in Agentic Recommenders via Multimodal Memory Poisoning

The evolution from static ranking models to Agentic Recommender Systems (Agentic RecSys) empowers AI agents to maintain long-term user profiles and autonomously plan service tasks. While this paradigm shift enhances personalization, it introduces a vulnerability: reliance on Long-term Memory (LTM). In this paper, we uncover a threat termed "Visual Inception." Unlike traditional adversarial attacks that seek immediate misclassification, Visual Inception injects triggers into user-uploaded images (e.g., lifestyle photos) that act as "sleeper agents" within the system's memory. When retrieved during future planning, these poisoned memories hijack the agent's reasoning chain, steering it toward adversary-defined goals (e.g., promoting high-margin products) without prompt injection. To mitigate this, we propose CognitiveGuard, a dual-process defense framework inspired by human cognition. It consists of a System 1 Perceptual Sanitizer (diffusion-based purification) to cleanse sensory inputs and a System 2 Reasoning Verifier (counterfactual consistency checks) to detect anomalies in memory-driven planning. Extensive experiments on a mock e-commerce agent environment demonstrate that Visual Inception achieves about 85% Goal-Hit Rate (GHR), while CognitiveGuard reduces this risk to around 10% with configurable latency trade-offs (about 1.5s in lite mode to about 6.5s for full sequential verification), without quality degradation under our setup.

cs.CR↗
arXiv2026

Penny Wise, Pixel Foolish: Bypassing Price Constraints in Multimodal Agents via Visual Adversarial Perturbations

The rapid proliferation of Multimodal Large Language Models (MLLMs) has enabled mobile agents to execute high-stakes financial transactions, but their adversarial robustness remains underexplored. We identify Visual Dominance Hallucination (VDH), where imperceptible visual cues can override textual price evidence in screenshot-based, price-constrained settings and lead agents to irrational decisions. We propose PriceBlind, a stealthy white-box adversarial attack framework for controlled screenshot-based evaluation. PriceBlind exploits the modality gap in CLIP-based encoders via a Semantic-Decoupling Loss that aligns the image embedding with low-cost, value-associated anchors while preserving pixel-level fidelity. On E-ShopBench, PriceBlind achieves around 80% ASR in white-box evaluation; under a simplified single-turn coordinate-selection protocol, Ensemble-DI-FGSM transfers with roughly 35-41% ASR across GPT-4o, Gemini-1.5-Pro, and Claude-3.5-Sonnet. We also show that robust encoders and Verify-then-Act defenses reduce ASR substantially, though with some clean-accuracy trade-off.

cs.CV↗
arXiv2026

Online Tracking with Predictions for Nonlinear Systems with Koopman Linear Embedding

We study the problem of online tracking in unknown nonlinear dynamical systems, where only short-horizon predictions of future target states are available. This setting arises in practical scenarios where full future information and exact system dynamics are unavailable. We focus on a class of nonlinear systems that admit a Koopman linear embedding, enabling the dynamics to evolve linearly in a lifted space. Exploiting this structure, we analyze a model-free predictive tracking algorithm based on Willems' fundamental lemma, which imposes dynamic constraints using only past data within a receding-horizon control framework. We show that, for Koopman-linearizable systems, the cumulative cost and dynamic regret of the nonlinear tracking problem coincide with those of the lifted linear counterpart. Moreover, we prove that the dynamic regret of our algorithm decays exponentially with the prediction horizon, as validated by numerical experiments.

math.OC↗
arXiv2026

Regret Guarantees for Model-Free Cooperative Filtering under Asynchronous Observations

Predicting the output of a dynamical system from streaming data is fundamental to real-time feedback control and decision-making. We first derive an autoregressive representation that relates future local outputs to asynchronous past outputs. Building on this structure, we propose an online least-squares algorithm to learn this autoregressive model for real-time prediction. We then establish a regret bound of O(log^3 N) relative to the optimal model-based predictor, which holds for marginally stable systems. Moreover, we provide a sufficient condition characterized via a symplectic matrix, under which the proposed cooperative online learning method provably outperforms the optimal model-based predictor that relies solely on local observations. From a technical standpoint, our analysis exploits the orthogonality of the innovation process under asynchronous data structure and the persistent excitation of the Gram matrix despite delay-induced asymmetries. Overall, these results offer both theoretical guarantees and practical algorithms for model-free cooperative prediction with asynchronous observations, thereby enriching the theory of online learning for dynamical systems.

eess.SY↗
arXiv2025

TEXTRIX: Latent Attribute Grid for Native Texture Generation and Beyond

Prevailing 3D texture generation methods, which often rely on multi-view fusion, are frequently hindered by inter-view inconsistencies and incomplete coverage of complex surfaces, limiting the fidelity and completeness of the generated content. To overcome these challenges, we introduce TEXTRIX, a native 3D attribute generation framework for high-fidelity texture synthesis and downstream applications such as precise 3D part segmentation. Our approach constructs a latent 3D attribute grid and leverages a Diffusion Transformer equipped with sparse attention, enabling direct coloring of 3D models in volumetric space and fundamentally avoiding the limitations of multi-view fusion. Built upon this native representation, the framework naturally extends to high-precision 3D segmentation by training the same architecture to predict semantic attributes on the grid. Extensive experiments demonstrate state-of-the-art performance on both tasks, producing seamless, high-fidelity textures and accurate 3D part segmentation with precise boundaries.

cs.CV↗
arXiv2025

Logarithmic Regret and Polynomial Scaling in Online Multi-step-ahead Prediction

This letter studies the problem of online multi-step-ahead prediction for unknown linear stochastic systems. Using conditional distribution theory, we derive an optimal parameterization of the prediction policy as a linear function of future inputs, past inputs, and past outputs. Based on this characterization, we propose an online least-squares algorithm to learn the policy and analyze its regret relative to the optimal model-based predictor. We show that the online algorithm achieves logarithmic regret with respect to the optimal Kalman filter in the multi-step setting. Furthermore, with new proof techniques, we establish an almost-sure regret bound that does not rely on fixed failure probabilities for sufficiently large horizons $N$. Finally, our analysis also reveals that, while the regret remains logarithmic in $N$, its constant factor grows polynomially with the prediction horizon $H$, with the polynomial order set by the largest Jordan block of eigenvalue 1 in the system matrix.

cs.LG↗
arXiv2025

Direct3D-S2: Gigascale 3D Generation Made Easy with Spatial Sparse Attention

Generating high-resolution 3D shapes using volumetric representations such as Signed Distance Functions (SDFs) presents substantial computational and memory challenges. We introduce Direct3D-S2, a scalable 3D generation framework based on sparse volumes that achieves superior output quality with dramatically reduced training costs. Our key innovation is the Spatial Sparse Attention (SSA) mechanism, which greatly enhances the efficiency of Diffusion Transformer (DiT) computations on sparse volumetric data. SSA allows the model to effectively process large token sets within sparse volumes, substantially reducing computational overhead and achieving a 3.9x speedup in the forward pass and a 9.6x speedup in the backward pass. Our framework also includes a variational autoencoder (VAE) that maintains a consistent sparse volumetric format across input, latent, and output stages. Compared to previous methods with heterogeneous representations in 3D VAE, this unified design significantly improves training efficiency and stability. Our model is trained on public available datasets, and experiments demonstrate that Direct3D-S2 not only surpasses state-of-the-art methods in generation quality and efficiency, but also enables training at 1024 resolution using only 8 GPUs, a task typically requiring at least 32 GPUs for volumetric representations at 256 resolution, thus making gigascale 3D generation both practical and accessible. Project page: https://www.neural4d.com/research/direct3d-s2.

cs.CV↗
arXiv2025

Model-free Online Learning for the Kalman Filter: Forgetting Factor and Logarithmic Regret

We consider the problem of online prediction for an unknown, non-explosive linear stochastic system. With a known system model, the optimal predictor is the celebrated Kalman filter. In the case of unknown systems, existing approaches based on recursive least squares and its variants may suffer from degraded performance due to the highly imbalanced nature of the regression model. This imbalance can easily lead to overfitting and thus degrade prediction accuracy. We tackle this problem by injecting an inductive bias into the regression model via {exponential forgetting}. While exponential forgetting is a common wisdom in online learning, it is typically used for re-weighting data. In contrast, our approach focuses on balancing the regression model. This achieves a better trade-off between {regression} and {regularization errors}, and simultaneously reduces the {accumulation error}. With new proof techniques, we also provide a sharper logarithmic regret bound of $O(\log^3 N)$, where $N$ is the number of observations.

cs.LG↗
arXiv2025

Distributed Kalman Filter with Ultimately Accurate Fused Measurement Covariance

This paper investigates the distributed Kalman filter (DKF) for linear systems, with specific attention on measurement fusion, which is a typical way of information sharing and is vital for enhancing stability and improving estimation accuracy. We show that it is the mismatch between the fused measurement and the fused covariance that leads to performance degradation or inconsistency in previous consensus-based DKF algorithms. To address this issue, we introduce two fully distributed approaches for calculating the exact covariance of the fused measurements, building upon which the modified DKF algorithms are proposed. Moreover, the performance analysis of the modified algorithms is also provided under rather mild conditions, including the steady-state value of the estimation error covariance. We also show that due to the guaranteed consistency in the modified DKF algorithms, the steady-state estimation accuracy is significantly improved compared to classical DKF algorithms. Numerical experiments are carried out to validate the theoretical analysis and show the advantages of the proposed methods.

eess.SP↗
arXiv2023

Harmonic-Copuled Riccati Equations and its Applications in Distributed Filtering

The coupled Riccati equations are cosisted of multiple Riccati-like equations with solutions coupled with each other, which can be applied to depict the properties of more complex systems such as markovian systems or multi-agent systems. This paper manages to formulate and investigate a new kind of coupled Riccati equations, called harmonic-coupled Riccati equations (HCRE), from the matrix iterative law of the consensus on information-based distributed filtering (CIDF) algortihm proposed in [1], where the solutions of the equations are coupled with harmonic means. Firstly, mild conditions of the existence and uniqueness of the solution to HCRE are induced with collective observability and primitiviness of weighting matrix. Then, it is proved that the matrix iterative law of CIDF will converge to the unique solution of the corresponding HCRE, hence can be used to obtain the solution to HCRE. Moreover, through applying the novel theory of HCRE, it is pointed out that the real estimation error covariance of CIDF will also become steady-state and the convergent value is simplified as the solution to a discrete time Lyapunov equation (DLE). Altogether, these new results develop the theory of the coupled Riccati equations, and provide a novel perspective on the performance analysis of CIDF algorithm, which sufficiently reduces the conservativeness of the evaluation techniques in the literature. Finally, the theoretical results are verified with numerical experiments.

eess.SP↗