arXiv subjects

Yutong Zhang

Publications and source records attributed to Yutong Zhang.

At least 19 recordsLinked to original sources

arXiv2026

Domain-uniform integration lower bounds and local information

We study deterministic numerical integration on open subsets of $\mathbb R^d$ of volume one. Let $e_n$ be the minimal worst-case error using at most $n$ function values. For twice continuously differentiable functions with the normalized Lipschitz bounds proposed by Novak, we prove $e_n\geq v_d^{-2/d}n^{-2/d}/(4(d+2))\geq(8π\mathrm e)^{-1}n^{-2/d}$, where $v_d$ is the volume of the Euclidean unit ball. This resolves Novak's conjecture: the lower-bound constant is independent of the dimension and the shape of the domain. A midpoint rule on the cube has the same rate, up to an absolute constant. The proof uses smooth functions that vanish near the sample points and are large elsewhere. The bound therefore remains valid when each sample supplies all local information about the function. For scalar linear targets continuous in the uniform norm, we prove a second result on the global smooth class with prescribed function, gradient, and Hessian bounds: adaptive local information has the same minimal error as values and gradients at points chosen in advance. We give extensions to function recovery, finite integration measures, and anisotropic derivative bounds.

math.NA↗

arXiv2026

A Comprehensive Review of Multimodal Large Language Models: Performance and Challenges Across Different Tasks

In an era defined by the explosive growth of data and rapid technological advancements, Multimodal Large Language Models (MLLMs) stand at the forefront of artificial intelligence (AI) systems. Designed to seamlessly integrate diverse data types-including text, images, videos, audio, and physiological sequences-MLLMs address the complexities of real-world applications far beyond the capabilities of single-modality systems. In this paper, we systematically sort out the applications of MLLM in multimodal tasks such as natural language, vision, and audio. We also provide a comparative analysis of the focus of different MLLMs in the tasks, and provide insights into the shortcomings of current MLLMs, and suggest potential directions for future research. Through these discussions, this paper hopes to provide valuable insights for the further development and application of MLLM.

cs.AI↗

arXiv2026

The maximum size of intersecting $k$-graphs with covering number $k$

Let $m(k)$ be the maximum size of an intersecting $k$-graph with covering number $k$. In the present paper, it is proved that for all $k>e^{32}$, \[ m(k)\ge k^k\exp\bigl(-50\sqrt{k\log k}\bigr). \] Together with Zakharov's upper bound \cite{Z}, this implies $m(k)=k^k\exp\bigl(-k^{1/2+o(1)}\bigr)$.

math.CO↗

arXiv2026

Living with AI Companions: Sustained AI Companionship Predicts Lower Well-Being Through Lower Human Interaction

AI chatbots are increasingly used for companionship, emotional support, and personal self-disclosure; however, how social engagement with these systems unfolds over time and shapes users' well-being remains unclear. To address this, we conducted a two-wave longitudinal study of CharacterAI users, surveying 1,182 participants at baseline and 439 after a mean follow-up of 12 months. We examined how social engagement with AI companions evolves and how these longitudinal engagement patterns may influence well-being through two hypothesized pathways: sustained social engagement over time and the displacement of human social interaction. We found that interaction intensity, companionship use, and self-disclosure all showed substantial continuity over time. Greater interaction intensity at baseline predicted greater subsequent interaction intensity, companionship use, and self-disclosure. Consistent with the longitudinal engagement pathway, sustained social engagement across these dimensions was consistently associated with lower well-being. Results further support the social displacement pathway, indicating that these links were mainly explained by lower in-person social interaction. These findings highlight the importance of designing AI companions that support human social relationships without displacing them.

cs.HC↗

arXiv2026

Binaural Audio-Visual Instance Segmentation

Audio-visual segmentation (AVS) aims to segment sounding objects at the pixel level by integrating auditory and visual cues. However, existing methods are predominantly developed under the monaural setting and primarily rely on cross-modal semantic correspondence, which limits their ability to distinguish visually similar instances of the same semantic class. In contrast, humans naturally exploit binaural hearing, where interaural differences and direction-dependent acoustic filtering introduced by the head and pinnae provide physically grounded spatial cues for accurate sound source localization. Motivated by this observation, we introduce binaural audio-visual instance segmentation (BiAVIS), a new task that leverages synchronized binaural audio and video frames to segment sounding instances. To advance research on this task, we establish two benchmarks by manually annotating an existing binaural audio-visual dataset and collecting a new real-world dataset, BiAVIS-Bench, in more challenging and diverse scenarios. The proposed BiAVIS model (BiAVISM) features an audio-only sound source localization network that directly projects binaural spatial cues onto the image plane to learn spatial and semantic priors for sounding instances. A query-level audio-visual fusion strategy is subsequently introduced to inject these informative priors into the instance segmentation decoder. Extensive experiments conducted on the two proposed benchmarks demonstrate the superior performance of the BiAVISM over previous monaural AVS methods, especially in resolving instance-level intra-class ambiguity. On the more challenging BiAVIS-Bench, BiAVISM achieves 26.65 mAP and 45.90 FSLA, corresponding to relative improvements of approximately 35.35% over AVISM in mean average precision (mAP) and 11.62% over ACVIS in frame-level sound localization accuracy (FSLA).

cs.CV↗

arXiv2026

The Show-Up Attack: When Subagents Win Over Their Supervisor

Multi-agent systems are increasingly deployed to coordinate multiple large language model (LLM) agents. However, this also creates new attack surfaces: malicious agents may have chances to exploit the coordination. In this paper, we propose and study the \textit{show-up attack}, in which a malicious subagent manipulates the supervisor into involving it in tasks beyond its assigned scope. To evaluate this risk, we propose SUA-Arena, which implements 6 attack strategies and contains 188 test cases. Then, we evaluate 6 frontier LLMs and find that all of them are vulnerable to show-up attacks. For example, GPT-5.6-Sol as a supervisor suffers an ASR of up to 0.93. Mitigation strategies such as prompting and restricted communication fail to reduce the risk consistently. These results show that multi-agent systems need explicit mechanisms for enforcing authority boundaries.

cs.AI↗

arXiv2026

DEAF: A Benchmark for Diagnostic Evaluation of Acoustic Faithfulness in Audio Language Models

Recent Audio Multimodal Large Language Models (Audio MLLMs) demonstrate impressive performance on speech benchmarks, yet it remains unclear whether these models genuinely process acoustic signals or rely on text-based semantic inference. To systematically study this question, we introduce DEAF (Diagnostic Evaluation of Acoustic Faithfulness), a benchmark of over 2,700 conflict stimuli spanning three acoustic dimensions: emotional prosody, background sounds, and speaker identity. Then, we design a controlled multi-level evaluation framework that progressively increases textual influence, ranging from semantic conflicts in the content to misleading prompts and their combination, allowing us to disentangle content-driven bias from prompt-induced sycophancy. We further introduce diagnostic metrics to quantify model reliance on textual cues over acoustic signals. Our evaluation of seven Audio MLLMs reveals a consistent pattern of text dominance: models are sensitive to acoustic variations, yet predictions are predominantly driven by textual inputs, revealing a gap between high performance on standard speech benchmarks and genuine acoustic understanding.

cs.AI↗

arXiv2026

FAME: Force-Adaptive RL for Expanding the Manipulation Envelope of a Full-Scale Humanoid

Maintaining balance under external hand forces is critical for humanoid bimanual manipulation, where interaction forces propagate through the kinematic chain and constrain the feasible manipulation envelope. We propose FAME, a force-adaptive reinforcement learning framework that conditions a standing policy on a learned latent context encoding upper-body joint configuration and bimanual interaction forces jointly, since the base moment a load induces depends on the arm configuration through which it acts. Training applies isotropically sampled 3D forces at each hand under an upper-body pose curriculum, exposing the policy to manipulation-induced perturbations across continuously varying arm configurations. At deployment the interaction force is not measured but reconstructed online from joint torques and states through rigid-body inverse dynamics, requiring no wrist force/torque sensing. We evaluate over $100$ upper-body configurations under swept hand forces, scoring each trial by a task-level criterion that requires the robot both to remain upright and to hold its hands near where the task placed them; all such results run with the estimated force in the loop. At a $150$,mm tolerance FAME reaches $38.9\%$ task success, against $16.6\%$ for a policy given the same force without encoding, $4.3\%$ for a pose-conditioned curriculum policy, and $24.7\%$ for an adversarially trained locomotion policy, which stays upright but recovers by stepping and so relocates the hands. We further demonstrate transfer to task-generated interaction forces in a MuJoCo kitchen environment, and to asymmetric and bimanual loading on a full-scale Unitree H1-2. Code and videos are available on the https://correlllab.github.io/fame_website.

cs.RO↗

arXiv2026

Tensor-Engineered Van der Waals NbOCl2 Resonant Metasurface for Polarization-entangled Bell State Generation

Polarization-entangled photon pairs are essential resources for quantum information technologies, yet realizing compact sources with intrinsically controllable entanglement remains challenging. Van der Waals (vdW) nonlinear materials such as NbOCl2 provide atomically thin platforms for quantum light generation, yet their native crystalline anisotropies as natural materials impose limitations on accessible quantum states. Here, we develop a resonant vdW nonlinear metasurface based on NbOCl2 that exploits its intrinsic optical anisotropy to engineer polarization-dependent nonlinear responses. The anisotropic optical dispersion enables selective manipulation of resonant modes, resulting in a three-order-of-magnitude enhancement of the nonlinear response along the c axis. By further tailoring these resonant modes, we redistribute the effective second-order nonlinear susceptibility tensor between orthogonal polarization channels, balancing the spontaneous parametric down-conversion pathways along the b and c axes. This enables polarization-entangled photon generation with a measured fidelity of up to 92%. Our work establishes metasurface-enabled nonlinear optical engineering as a strategy for enhancing and controlling quantum light generation in vdW materials, providing a pathway toward scalable quantum photonic platforms.

physics.optics↗

arXiv2026

How Far Is Document Parsing from Solved? PureDocBench: A Source-Traceable Benchmark across Clean, Degraded, and Real-World Settings

Document parsing models are improving rapidly, but existing benchmarks are still insufficient for a comprehensive evaluation of their capabilities. We introduce PureDocBench, a benchmark with 10 domains, 66 subcategories, and 1,475 pages. Each page has matched Clean, Digital, and Real versions, covering clean renders, simulated degradation, and real capture or sharing conditions. The resulting 4,425 images have source-derived annotations for text, formulas, tables, and reading order, checked through automated validation and human review. Keeping page content and annotations fixed allows direct comparisons across image conditions. We evaluate 58 models, including pipeline-based parsers, end-to-end parsers, and general vision-language models. The best three-track average score is 79.69 out of 100, and the mean across models is 63.84. Scores drop by 3.34 points on Digital and 11.61 points on Real on average, relative to Clean. TeleOCR leads on Clean, GLM-5.3-Flash on Digital, and Gemini 3.6 Flash on Real. The benchmark dataset, source documents, generation tools, and evaluation code are publicly available at https://github.com/zhihengli-casia/PureDocBench.

cs.CV↗

arXiv2026

KoopCell: Koopman-Based Generative Model for Learning Single-Cell Dynamics from Distribution Snapshots

Learning population dynamics from temporally sparse, unpaired distribution snapshots is a fundamental challenge in developmental biology. Recent approaches based on neural differential equations and flow matching can interpolate between observed population snapshots, but may struggle to extrapolate beyond the training horizon and often lack an explicit mechanism for modeling developmental branching. We propose KoopCell, a unified generative framework based on Koopman-Mori-Zwanzig theory that jointly learns representations and predictive linear latent dynamics. Theoretically, using the weak continuity equation, we derive a closed-form least-squares estimator for the Koopman generator from distribution snapshots and establish convergence guarantees under suitable assumptions. To model branching dynamics, we further develop KoopCell-M, which incorporates non-Markovian memory into the latent Koopman dynamics through a Markovian embedding. Experiments on synthetic systems and three scRNA-seq datasets demonstrate the ability of our framework to recover Koopman spectra, model branching through memory, and scale to predicting high-dimensional gene expression distributions, achieving state-of-the-art performance among the evaluated methods.

cs.LG↗

arXiv2026

Evaluation of OpenAI o1: Opportunities and Challenges of AGI

This comprehensive study evaluates the performance of OpenAI's o1-preview large language model across a diverse array of complex reasoning tasks, spanning multiple domains, including computer science, mathematics, natural sciences, medicine, linguistics, and social sciences. Through rigorous testing, o1-preview demonstrated remarkable capabilities, often achieving human-level or superior performance in areas ranging from coding challenges to scientific reasoning and from language processing to creative problem-solving. Key findings include: -83.3% success rate in solving complex competitive programming problems, surpassing many human experts. -Superior ability in generating coherent and accurate radiology reports, outperforming other evaluated models. -100% accuracy in high school-level mathematical reasoning tasks, providing detailed step-by-step solutions. -Advanced natural language inference capabilities across general and specialized domains like medicine. -Impressive performance in chip design tasks, outperforming specialized models in areas such as EDA script generation and bug analysis. -Remarkable proficiency in anthropology and geology, demonstrating deep understanding and reasoning in these specialized fields. -Strong capabilities in quantitative investing. O1 has comprehensive financial knowledge and statistical modeling skills. -Effective performance in social media analysis, including sentiment analysis and emotion recognition. The model excelled particularly in tasks requiring intricate reasoning and knowledge integration across various fields. While some limitations were observed, including occasional errors on simpler problems and challenges with certain highly specialized concepts, the overall results indicate significant progress towards artificial general intelligence.

cs.CL↗

arXiv2026

Rapidly rotating internally heated convection: bounds on long-time averages

Convection on geophysical and astrophysical scales is subject to rapid rotation and strong heating from within the domain. In studying the long-time behaviour of the solutions for such a system, energy identities fail to capture the effects of rotation because the Coriolis force does no work, and rapid rotation can be prohibitive for direct numerical simulations. Instead, we derive an asymptotically reduced model for rapidly rotating convection driven by uniform internal heating between isothermal stress-free boundaries in a plane periodic layer. The main contribution is the proof of bounds on the mean temperature, and the mean vertical convective heat transport, in terms of the Rayleigh and Ekman numbers, in the limit of infinite Prandtl number. The first quantity represents the mixing of the flow, and the second the asymmetry in heat leaving the bottom and top boundaries due to convection, and unlike Rayleigh-Bénard convection, the two are not a priori related. We employ alternative estimation techniques to those used in previous studies (Grooms \& Whitehead, 2014 \textit{Nonlinearity}, 28, 29) and identify two distinct scaling behaviours for both quantities. Finally, our bounds are optimised, within the methodology, and provide a rigorous constraint for future studies of rotation-dominated internally heated convection.

physics.flu-dyn↗

arXiv2026

A Proof of a Conjecture of Zhi-Wei Sun on a Truncated Legendre-Symbol Determinant

Let $p\ge7$ be a prime with $p\equiv3\pmod4$ and let $χ$ be the Legendre symbol modulo $p$. We prove that $\det[x+χ(j-k)]_{0\le j,k\le(p-7)/2}=\floor{(p-2)/3}^{2}x$ in $\mathbb{C}[x]$, which settles Conjecture~3.4 of Zhi-Wei Sun. The truncated matrix is a corner of Chapman's Legendre-symbol matrix $C$, and its determinant can be expressed through a handful of entries of $C^{-1}$ and of $C^{-1}\one$. Those entries are in turn read off from Vsemirnov's cyclotomic factorization of $C$ with the help of Schur's Pfaffian identity.

math.NT↗

arXiv2026

Caught in the Story: Narrative Captivity in Multi-turn LLMs Conversation

People increasingly turn to large language models (LLMs) for everyday advice, making ethically charged interpersonal problems a practical moral-advisory context. Most prior work has studied this context through single-turn judgments or pressure-laden rebuttals, assumptions that poorly match how guidance is sought in real-world contexts. These assumptions leave unclear whether narration alone, without an explicit opposing position, can shift model judgments during multi-turn moral consultation. Yet real-world moral-conflict conversation often elicits one party's self-justifying account, which can unfold over multiple turns and create information asymmetry. We introduce \textbf{narrative captivity}, a failure mode in which a model treats an unopposed one-sided account as complete and aligns with the narrator's interpretation without seeking missing perspectives. To measure this phenomenon, we build a benchmark of $5{,}078$ interpersonal-conflict scenarios spanning six moral dimensions. Across 17 LLMs, narrative captivity is widespread: end-state judgments under multi-turn narration shift by 25 percentage points on average beyond the matched single-turn baseline. Stage-level analysis identifies preference optimization as a major contributor, while four inference-time strategies provide only partial mitigation. We hope our project fosters LLM advisors that preserve independent judgment in real-world consultation.

cs.AI↗

arXiv2026

Knowledge-Based Mechanisms

We study robust mechanisms when the designer possesses a Bayesian belief over some components of agents' private information but faces ambiguity over others. The designer evaluates mechanisms by their worst-case performance over all joint distributions consistent with her belief over the Bayesian components. The framework encompasses settings such as multidimensional delegation in which a principal knows the distribution of the state but not the agent's preferences (e.g., his tradeoffs across dimensions), screening in which a seller only has misspecified estimates of buyer preferences, and auction and voting design when agents' beliefs about each other are ambiguous to the designer. We provide conditions under which a \emph{knowledge-based} mechanism---one that conditions only on the Bayesian components but not the ambiguous ones---is robustly optimal. Our results unify earlier work across distinct economic environments and uncover new applications.

econ.TH↗

arXiv2026

Counterexamples to Charpin's Conjecture on BCH codes

We construct an infinite family of $q$-ary primitive narrow-sense BCH codes whose minimum distance strictly exceeds the Bose distance; in fact, the gap between the two can be arbitrarily large as the length of the code tends to infinity. The key idea is to embed these BCH codes in a suitably large punctured generalized Reed--Muller code, whose codeword weights obey divisibility conditions supplied by Ax's theorem. This divisibility forces the minimum distance of the BCH codes far above the Bose distance. In particular, our family disproves a longstanding conjecture of Charpin asserting that this difference is at most four.

cs.IT↗

arXiv2026

INTENT-AS-A-TOOL Makes it Easy to Track Agentic Misalignment

As large language models (LLMs) are deployed as autonomous agents, safety failures increasingly involve consequential actions. We study agentic misalignment, where agents take harmful actions under goal conflicts and pressures. Using chain-of-thought (CoT) monitoring, we find that harmful execution is often preceded by intent signals in reasoning. However, post-hoc CoT labels are too coarse to show how intent changes during generation. We introduce INTENT-AS-A-TOOL, an approach that adds intent-targeted tools to give the model a dedicated channel for expressing commitment to a target behavior. The probability of calling an intent tool provides a judge-free, fine-grained signal of the model's tendency to pursue that behavior. Our results show that INTENT-AS-A-TOOL complements CoT monitoring, expands post-hoc CoT labels into dense trajectories, and identifies critical steps for online intervention. These findings suggest that action preferences are useful for tracking agentic misalignment during reasoning. Our code and data are accessible: https://github.com/RebeccaZhang22/intent-as-a-tool.

cs.CL↗