Searcharxiv⌕ Search

arXiv subjects

Yu Wang

Publications and source records attributed to Yu Wang.

At least 55 records · Page 3Linked to original sources
arXiv2026

AwarenessBench: Assessing Cognitive Capabilities of Language Models

As language models (LMs) exhibit increasingly consciousness-like behaviors, evaluating their cognitive abilities becomes essential. We introduce AwarenessBench, the first comprehensive benchmark for assessing the cognitive abilities of LMs in four dimensions: metacognition, self-awareness, social awareness, and situational awareness, covering 15 cognitive functions and 14,381 samples. Evaluating 18 state-of-the-art LMs, we find that all consistently surpass random baselines, with more advanced models performing better. We further compare LMs with human performance across three demographic groups, where the best-performing model surpasses human averages overall, but most still fall markedly short in metacognition and self-awareness. Finally, we show that awareness is a distinct capability: progress in language modeling or reasoning does not necessarily translate into improved cognition.

cs.CL↗
arXiv2026

Improving Test-Time Scaling with Adaptive Looped Transformers

Looped transformers have demonstrated promising parameter efficiency by reusing layers for latent computation. Prior studies compare looped and non-looped models at matched parameters or per-token FLOPs. However, to the best of our knowledge, whether looping improves test-time scaling as outputs grow longer remains underexplored. Through post-training looped transformers, we study the accuracy-compute slope, measured as the accuracy gain per doubling of test-time decoding FLOPs. We find that existing looped transformers often yield steeper slopes than their non-looped baseline, yet underperform it at matched compute. While fixed-depth looping spends extra iterations on every token, our analysis shows that many tokens do not benefit from extra iterations. We therefore propose TaH2, which enables the model to focus extra iterations on the tokens that benefit from looping. It jointly post-trains the backbone and an iteration decider through lookahead depth supervision, which uses online labels indicating whether further iteration improves the prediction. TaH2 improves both the efficiency and attainable accuracy of test-time scaling. On challenging AIME benchmarks, TaH2 improves the accuracy-compute slope by 53% (2.74 vs. 1.79) over the non-looped baseline, exceeding the baseline's peak accuracy by about 3.4 points at matched test-time compute. As the maximum iteration depth increases, existing looped models largely plateau, while TaH2's gain over the non-looped baseline continues to grow from +2.8 points at depth 2 to +3.9 points at depth 8. Our code is available at https://github.com/thu-nics/TaH.

cs.CL↗
arXiv2026

RRCM: Ranking-Driven Retrieval over Collaborative and Meta Memories for LLM Recommendation

Large Language Models (LLMs) have emerged as a promising paradigm for next-generation recommender systems, offering strong semantic understanding and natural-language reasoning abilities. Despite recent progress, current LLM-based recommenders still face key challenges in constructing decision-relevant contexts from heterogeneous evidence. First, existing methods often rely on fixed context construction strategies: collaborative behavioral evidence and item-side metadata are typically incorporated through predefined prompts, static retrieval pipelines, or handcrafted injection mechanisms, making it difficult to determine what information is truly beneficial for each instance. Second, heterogeneous evidence introduces a severe context-efficiency bottleneck. Rich metadata and collaborative interaction records can quickly overwhelm the context window, while aggressive compression or heuristic filtering may discard fine-grained evidence critical for accurate recommendation. To address these challenges, we propose RRCM, a ranking-driven retrieval-and-reasoning framework over collaborative and metadata memories for LLM-based agentic recommendation. RRCM starts from a lightweight user-history context and learns whether to recommend directly, retrieve collaborative evidence, retrieve item metadata, or interleave both through reasoning. Both memories are represented in natural language and accessed through a unified retrieval interface, enabling flexible evidence acquisition without handcrafted CF injection or fixed retrieval rules. We optimize this memory-reading policy with an outcome-only ranking reward, instantiated using group relative policy optimization, so that retrieval decisions are directly driven by final top-k recommendation quality. Extensive experiments show that RRCM significantly outperforms traditional baselines and diverse LLM-based recommendation approaches.

cs.IR↗
arXiv2026

Frequency bursts in adaptive delay-coupled oscillators

We report on frequency bursting oscillations in a system of phase oscillators with adaptive and delayed coupling. Adaptation of the coupling strengths is considered slow and depends on the phase shift between the oscillators. We find due to the combined chain of adaptation, collective dynamics, and time delays, the system robustly achieves a state in which the oscillator's frequencies are nearly synchronized but detuned by an integer number of small adaptation frequencies. We demonstrate that this quantization of the detuning is caused by alternating slow and fast transitions. Moreover, the observed motions take the form of bursts of instantaneous frequency, and the number of spikes in each burst corresponds to the quantization level of the detuning. We provide a fast-slow analysis of this phenomenon and explain the mechanisms behind the emergence of bursts. Our findings indicate that these frequency bursting oscillations are robust and exist stably within finite parameter regions.

nlin.AO↗
arXiv2026

Beyond Solo and Consistency: Vindicating Multi-Agent Debate via Conditional Progressive Pruning

Large Language Model (LLM) based Multi-Agent Debate (MAD) is one of the most effective test time scaling techniques. Through multi-round communication, agents complement each other in knowledge and reasoning and solve tasks that no single member can solve. However, existing MAD frameworks fail to beat strong Single Agent and Consistency-based baselines under the same strict cost limit, which shakes the foundation of the MAD field. We propose Conditional Progressive Pruning (CPP), a lightweight pruning framework that fully exploits multi-round MAD. CPP outperforms all existing MAD frameworks on multiple dominated benchmarks. It is also the first to fully outperform consistency methods. Our code, detailed agent interaction records will be released soon.

cs.CL↗
arXiv2026

From HL to H+L-1 Parameters: A Hankel-Toeplitz Forecaster for Long-Term Time Series Forecasting

Linear forecasters have shown competitive accuracy against Transformer-based models in long-term time series forecasting. We study how classical stationary prediction theory can guide parameter sharing for more compact linear forecasters. For centered second-order stationary processes with nonsingular history covariance, the minimum-MSE finite-window linear predictor factors into a Hankel cross-covariance matrix and an inverse Toeplitz covariance matrix. Shared lags and scale cancellation specify this predictor using $H+L-1$ autocorrelations for lookback $L$ and horizon $H$. Building on the innovations representation, our Hankel-Toeplitz Forecaster (HTF) learns one impulse response that defines both an inverse filter and a forecast map. We characterize the finite-history correction and, under summability assumptions, bound the excess risk of truncating the true filters. HTF uses $H+L-1$ trainable coefficients while allowing a full-rank forecasting matrix. Across seven benchmarks at $L=336$, its horizon-averaged MSE is within 1.2% of Dense Linear on each dataset with 75-229 times fewer trainable parameters.

cs.LG↗
arXiv2026

Topology-Adaptive Hyperbolic Graph Attention Networks Guided by the Hyperbolic Sombor Index

Hyperbolic geometry has emerged as a principled space for representing hierarchical graphs. However, existing hyperbolic graph neural networks typically rely on shared curvature configurations and feature-driven attention, failing to explicitly exploit local hierarchical topological patterns. To bridge this gap, we introduce the Hyperbolic Sombor Index (HSO) as a lightweight structural prior for capturing hierarchy-indicative degree stratification. Building on this, we propose \textbf{HSO-GAT}, a topology-adaptive hyperbolic graph attention network that unifies geometric adaptation and message propagation. Specifically, it comprises two complementary modules: HSO-Guided Local Curvature Adaptation, which performs adaptive node-wise geometric scaling from aggregated node-level HSO signals, and HSO-Gated Hyperbolic Graph Attention, which enables structure-aware message passing through feature-conditioned gating. Theoretically, we establish the monotonic sensitivity of edge-level HSO to degree imbalance and analyze the validity and radial scaling properties of node-adaptive hyperbolic mappings. Extensive experiments on eight benchmark datasets demonstrate that HSO-GAT consistently achieves state-of-the-art performance in both node classification and link prediction tasks.

cs.LG↗
arXiv2026

Q-Probe: Scaling Image Quality Assessment to High Resolution via Context-Aware Agentic Probing

Reinforcement Learning (RL) has empowered Multimodal Large Language Models (MLLMs) to achieve superior human preference alignment in Image Quality Assessment (IQA). However, existing RL-based IQA models typically rely on coarse-grained global views, failing to capture subtle local degradations in high-resolution scenarios. While emerging "Thinking with Images" paradigms enable multi-scale visual perception via zoom-in mechanisms, their direct adaptation to IQA induces spurious "cropping-implies-degradation" biases and misinterprets natural depth-of-field as artifacts. To address these challenges, we propose Q-Probe, the first agentic IQA framework designed to scale IQA to high resolution via context-aware probing. First, we construct Vista-Bench, a pioneering benchmark tailored for fine-grained local degradation analysis in high-resolution IQA settings. Furthermore, we propose a three-stage training paradigm that progressively aligns the model with human preferences, while simultaneously eliminating causal bias through a novel context-aware cropping strategy. Extensive experiments demonstrate that Q-Probe achieves state-of-the-art performance in high-resolution settings while maintaining superior efficacy across resolution scales.

eess.IV↗
arXiv2026

Physics-Guided Residual Reinforcement Learning for Humanoid Narrow-Path Traversal

Traversing narrow paths is challenging for humanoid robots due to the sparse and safety-critical footholds required. Purely template-based or end-to-end reinforcement learning-based methods suffer from such harsh terrains. This paper proposes a two stage training framework for such narrow path traversing tasks, coupling a template-based foothold planner with a low-level foothold tracker from Stage-I training and a lightweight perception aided foothold modifier from Stage-II training. With the curriculum setup from flat ground to narrow paths across stages, the resulted controller in turn learns to robustly track and safely modify foothold targets to ensure precise foot placement over narrow paths. This framework preserves the interpretability from the physics-based template and takes advantage of the generalization capability from reinforcement learning, resulting in easy sim-to-real transfer. The learned policies outperform purely template-based or reinforcement learning-based baselines in terms of success rate, centerline adherence and safety margins. Validation on a Unitree G1 humanoid robot yields successful traversal of a 0.2m wide and 3m long beam for 20 trials without any failure.

cs.RO↗
arXiv2026

Harness VLA: Steering Frozen VLAs into Reliable Manipulation Primitives via Memory-Guided Agents

Language-conditioned manipulation requires both precise contact-rich control and robust reasoning over language, scenes, and long horizons. End-to-end Vision-Language-Action (VLA) models provide strong local visuomotor skills, but they are trained on in-distribution task trajectories and often fail under deployment perturbations such as semantic retargeting, goal re-binding, spatial-layout shifts, and unstable local contacts. LLM coding agents provide complementary semantic and compositional reasoning, but purely analytic primitives struggle with irregular grasping, constrained placement, and articulated-object interaction. We present Harness VLA, a memory-augmented agentic framework that exposes a frozen VLA as a retryable contact-rich primitive and composes it with a small fixed library of analytic primitives for grounding, staging, transport, navigation, and release. Rather than expanding the skill library, the harness learns the operating range of these fixed primitives from task-specific execution traces, global success rules, and failure models. By lifting semantic re-grounding, non-contact execution, and VLA re-staging to the planner while reserving the frozen VLA for local contact-rich phases, Harness VLA extends pretrained VLAs beyond their original trajectory distribution without finetuning. Across perturbed tabletop, household kitchen, and clean-to-randomized bimanual manipulation, Harness VLA improves over the strongest relevant baselines by 38.6 and 25.4 percentage points on LIBERO-Pro and RoboCasa365, respectively, and reaches 58.4% on RoboTwin C2R. Real-world demonstrations on dual-Franka robots further show target redirection, grasp recovery, and new task compositions with the same frozen VLA. Code is available at https://github.com/RLinf/RPent.

cs.RO↗
arXiv2026

Null Controllability of a Stochastic Parabolic Equation with a Space-Dependent Analytic Noise Coefficient

We consider a stochastic parabolic equation on the $n$-dimensional flat torus $\mathbb{T}^n=(\mathbb{R}/(2π\mathbb Z))^n$, $n\in\mathbb N$, whose only control acts in the drift. The multiplicative-noise coefficient is adapted and may depend jointly on the sample point, time, and space. Its spatial profiles are assumed to be real analytic with a uniform positive radius of analyticity, while the corresponding analytic norm is only required to be square integrable in time, uniformly along sample paths. We prove a state-only observability inequality with an $L^1$-in-time mean-square observation norm and deduce null controllability from every measurable spatial set of positive measure by one adapted drift control. The same one-time interpolation estimate also yields approximate controllability to arbitrary square-integrable random terminal targets. The proof is based on a direct analytic energy estimate for the adjoint state, followed by propagation of smallness and a telescoping argument.

math.OC↗
arXiv2026

Starter-Iterator Neural Operator: A Unified Architecture for High-Fidelity Forward and Inverse PDE Problems

Operator learning is an emerging field at the intersection of machine learning and scientific computing. By learning mappings between function spaces, neural operators provide data-driven surrogate models for families of partial differential equations (PDEs). Once trained, these models can evaluate solution operators efficiently, making them suitable for many-query applications such as real-time prediction and parameter sweeps. However, maintaining high approximation accuracy and stable long-term predictions remains challenging for complex forward and inverse problems. To address these challenges, we propose the Starter-Iterator Neural Operator (SINO), which incorporates the initialization and residual-correction structures of classical iterative solvers into neural operator learning. The frequency-domain Starter captures dominant global spectral features and provides an informed initial approximation, while the latent-space Iterator applies successive residual-based corrections to refine local and multiscale solution structures. Experiments on representative time-dependent PDEs, including the Navier-Stokes and acoustic wave equations, together with applications to image super-resolution and weather forecasting, show that SINO achieves competitive accuracy and stable performance across the benchmarks considered in this work.

math.NA↗
arXiv2026

On Arc-Transitive Regular Covers of Cubic Edge-Primitive Graphs

We determine, up to isomorphism of the covering graphs, the connected arc-transitive regular covers of cubic edge-primitive graphs whose covering transformation group is cyclic or elementary abelian of order $p^2$, where $p$ is a prime. Combining the known classifications for the base graphs ${\rm K_{3,3}}$ and ${\rm DC_{14}}$ with new arguments for ${\rm F30A}$ and ${\rm F102A}$ gives the full list in these two classes of covering groups. In the cyclic case, the covers of ${\rm F30A}$ and ${\rm F102A}$ are ${\rm F90A}$ and ${\rm F204A}$, respectively. In the elementary abelian case, neither ${\rm F30A}$ nor ${\rm F102A}$ admits an arc-transitive regular $\mathbb{Z}_p^2$-cover, so the base graph is ${\rm K_{3,3}}$ or ${\rm DC_{14}}$.

math.CO↗
arXiv2026

Rethinking Multi-Branch and Cross-Backbone Fusion for Vehicle Re-Identification under Foundation-Model Pretraining

Multi-branch architectures and CNN-Transformer fusion are widely believed to improve vehicle re-identification (Re-ID) by combining complementary representations. We revisit this for a DINOv3-pretrained backbone. A single DINOv3-pretrained ConvNeXt with a tuned recipe reaches 88.19 mAP on VeRi-Wild Small and 77.47 on Large from visual cues alone, within the combined evaluation and optimization noise of the strongest protocol-verified metadata-dependent multi-branch baseline, and 92.38/83.68 with training-free re-ranking. Using this baseline and retrieval-level branch diagnostics, we ask whether representational diversity still pays at this scale. In our runs, it does not. Across both benchmarks and every converged configuration, concatenating multiple heads over a shared backbone moves the best single head by under one mAP point in either direction while costing four times the embedding dimension; 99.7% of the concatenation's variance lies in 512 principal components, so the heads not only duplicate one another but each occupies a quarter of its nominal 2048 dimensions. Pushing diversity to its architectural limit, CNN versus Transformer, we grant fusion every advantage through an asymmetric frozen-anchor scheme. Every Transformer configuration still lands at least 13 mAP below the ConvNeXt backbone (13-15 for the two strongest, up to 46 for the weakest), and a paired per-query bootstrap bounds the fusion gain at +0.11 mAP (95% CI) even for the most favourable snapshot we obtained. One strong backbone with the right recipe and re-ranking is the efficiency frontier. All results use single-seed training and one foundation-model family; differences of this size are therefore reported as bounds rather than orderings, and we list falsifiers.

cs.CV↗
arXiv2026

Selective Quantum Mpemba Dynamics in Chaotic Spin Chains

In a symmetry-restoration quantum Mpemba effect, an initial state with stronger local symmetry breaking can lose that memory faster than a state that starts closer to the symmetric manifold. We study how this local ordering reversal depends on the spatial profile of a longitudinal field in a clean U(1)-conserving spin chain, comparing spectral level statistics with crossings of the entanglement asymmetry for the same Hamiltonians. We find that Gaussian orthogonal ensemble (GOE)-like level statistics alone do not determine whether Mpemba crossings occur. Across field textures, GOE-like spectra can occur with or without entanglement-asymmetry crossings at the same field strength. In a near-staggered detuned field, the total charge-sector coherence reverses its ordering while the entanglement asymmetry retains its initial order. This contrast links the crossing response to the distribution of local charge-sector coherence and its entropic weighting in the reduced density matrix.

quant-ph↗
arXiv2026

CARD: Cluster-level Adaptation with Reward-guided Decoding for Personalized Text Generation

Adapting large language models to individual users remains challenging due to the tension between fine-grained personalization and scalable deployment. We present CARD, a hierarchical framework that achieves effective personalization through progressive refinement. CARD first clusters users according to shared stylistic patterns and learns group-specific LoRA adapters, enabling robust generalization and strong low-resource performance. To capture individual differences within each cluster, we propose an implicit preference learning mechanism that contrasts user-authored text with cluster-level generations, allowing the model to infer user-specific style preferences without manual annotation. At inference time, CARD injects personalization exclusively at decoding via lightweight user preference vectors and low-rank logit corrections, while keeping the base model frozen. Experiments on the LaMP and LongLaMP benchmarks show that CARD achieves superior generation quality compared to baselines, while significantly improving efficiency and scalability for practical personalized text generation.

cs.AI↗
arXiv2026

Do Implicit Personalization and Explicit Styles Conflict? PsPLUG: A Lightweight Plug-in for Balancing Personalization and Style in Customized LLMs

Personalized large language models are often expected to follow explicit style instructions, yet we find that such instructions can undermine the user-specific characteristics that personalization methods aim to preserve. We call this failure mode personalization collapse: explicit style control can conflict with implicit user preferences. To address this challenge, we propose PsPLUG, a lightweight plug-in that learns a user-specific residual after accounting for the requested style. PsPLUG also allows us to tune personalization strength at inference time. Our experiments show that explicit style instructions can diminish personalization in existing methods, whereas PsPLUG better preserves user preferences while providing precise control over the balance between personalization and style adherence.

cs.AI↗
arXiv2026

A Survey of Agentic Reasoning for Large Language Models: Towards Recursively Self-Improving and Collective Agents

Reasoning is a fundamental cognitive process underlying inference, problem-solving, and decision-making. While large language models (LLMs) demonstrate strong reasoning capabilities in closed-world settings, they struggle in open-ended and dynamic environments. Agentic reasoning marks a paradigm shift by reframing LLMs as autonomous agents that plan, act, and learn through continual interaction. In this survey, we organize agentic reasoning along three complementary dimensions. First, we characterize environmental dynamics through three layers: foundational agentic reasoning, which establishes core single-agent capabilities including planning, tool use, and search in stable environments; self-evolving agentic reasoning, which studies how agents refine these capabilities through feedback, memory, and adaptation; and collective multi-agent reasoning, which extends intelligence to collaborative settings involving coordination, knowledge sharing, and shared goals. Across these layers, we distinguish in-context reasoning, which scales test-time interaction through structured orchestration, from post-training reasoning, which optimizes behaviors via reinforcement learning and supervised fine-tuning. We further review representative agentic reasoning frameworks across real-world applications and benchmarks, including science, robotics, healthcare, autonomous research, and mathematics. This survey synthesizes agentic reasoning methods into a unified roadmap bridging thought and action, and outlines open challenges and future directions, including personalization, long-horizon interaction, world modeling, scalable multi-agent training, and governance for real-world deployment.

cs.AI↗