arXiv subjects

Yuting Wang

Publications and source records attributed to Yuting Wang.

At least 19 recordsLinked to original sources

arXiv2026

Periodicity-driven revision of the phase diagram of the generalized Baxter-Wu model with asymmetric complex couplings

The conventional self-dual lines of the generalized Baxter-Wu (GBW) model with asymmetric complex couplings are known to be $\sinh(2K)=\pm \cos(2\phi)$, where $K$ and $\phi$ are the real and imaginary parts of the coupling. We demonstrate that these lines are incomplete: the periodicity of the partition function, encoded in the cosine factor of the bundled Boltzmann weight, generates additional self-dual lines $\sinh(2K)=\pm \sin(2\phi)$. Guided by the complete set of self-dual candidates, we perform Monte Carlo simulations using brute-force reweighting (Metropolis) and the Wang-Landau methods. Simulations indicate that the self-dual lines at the partition-function minima $\phi_{\mathcal{Z}_{\min}}=(2n+1)\pi/8$ constitute a critical threshold. They are genuine critical boundaries for $|K| \ge \frac{1}{2}\operatorname{arsinh}(\cos(\pi/4)) \approx 0.32924$, while for smaller $|K|$ they are not. At $\phi_{\mathcal{Z}_{\min}}$, the sign problem is most severe and finite-size scaling corrections are largest; the local peak observed below the phase boundary in the temperature scan is thus a finite-size artifact, not a genuine new phase. We further clarify the capability and limitations of the average sign and its derivatives for detecting phase transitions. In particular, the negative peak of the average sign at $\phi_{\mathcal{Z}_{\min}}$ does not correspond to a genuine phase transition. We also evaluate the Wang-Landau method, which, despite formally circumventing the sign problem, still faces the exponential barrier.

cond-mat.str-el↗

arXiv2026

A Gaussian Covariance Matrix for Joint Pre- and Post-Reconstruction Full-Shape Power Spectrum Analysis

We apply the Gaussian covariance formalism to develop a semi-analytical covariance model for the joint analysis of pre-reconstruction, post-reconstruction, and cross full-shape galaxy power spectra. We model the reconstruction-reduced, scale-dependent cross shot noise using displacement-field statistics and introduce a new estimator that directly measures this term. Using the measured power spectra and the modeled shot-noise predictions as inputs, we construct the Gaussian covariance while accounting for correlations between the pre- and post-reconstruction density fields. We validate the resulting semi-analytical Gaussian covariance against mock catalogues. Using emulator-based parameter inference, we demonstrate that the semi-analytical Gaussian covariance adequately captures the dominant contribution to the covariance structure of the full data vector ($P_{\ell}^{\rm pre}, P_{\ell}^{\rm post}, P_{\ell}^{\rm cross}$). For the joint fit to these three power spectra, it yields cosmological constraints consistent with those obtained using the mock-based numerical covariance over the adopted fitting ranges: $k_{\rm max}=0.18\,h\,{\rm Mpc}^{-1}$ for $P_{\rm pre}$ and $P_{\rm post}$, and $k_{\rm max}=0.12\,h\,{\rm Mpc}^{-1}$ for $P_{\rm cross}$.

astro-ph.CO↗

arXiv2026

HERALD: Counterfactual Audits and Minimal Repairs for Proof-of-Retrieval Rewards

Search-agent rewards mix answer quality, citation grounding, tool cost, and anti-hacking terms; a high score therefore need not imply that cited evidence was retrieved, and added penalties can cancel. We introduce HERALD, an offline audit that applies exact same-question interventions, separates candidate-visible from oracle information, and enumerates detector contracts before policy optimization. On four Qwen3-8B pools from HotpotQA, 2WikiMultiHopQA, and MuSiQue, $R_0$ rejects search deletion and fake IDs, but a label-free citation-laundering attack succeeds. A complete $2^3$ ablation identifies targeted strengthening of $L$---citing a corpus passage absent from the retrieved evidence---as the observed inclusion-minimal repair: $R[L]$ has zero empirical ASR with a 0.50% one-sided cluster upper bound. The gap persists across pool rules, a visible BM25 attacker, and four models; broader hardening remains vulnerable when the attack removes an oracle support-ID penalty. Under strict 5M-token matched training evaluated on 256 paired questions per benchmark, $R[L]$ meets the EM non-inferiority gate on HotpotQA and 2Wiki but not MuSiQue. Equal-suite citation precision and support recall improve by 2.02 and 1.46 points, unsupported citations fall by 1.69, and laundering attackability falls on 2Wiki and MuSiQue. Natural $L$ is not reduced, and the detector appears in only 18 of 58,368 training trajectories. HERALD thus separates robust scoring, sparse learning signal, and policy transfer.

cs.AI↗

arXiv2026

Cosmological inference from the eBOSS QSO full-shape analysis with optimal redshift weights

We present a full-shape power-spectrum analysis of the eBOSS DR16 quasar sample with optimal redshift weights. The DR16 QSO catalog contains 343,708 quasars over $0.8<z<2.2$, a redshift interval broad enough to contain useful light-cone evolution but not naturally captured by a single effective-redshift measurement. We construct Karhunen--Lo\`eve weights for the parameters of interest and measure the resulting monopole and quadrupole with a cross-correlation estimator, which remains well defined for sign-changing weights. The theoretical spectra are convolved with the measured Fourier-space survey-window kernels for each Galactic cap and weighting scheme, and both the covariance matrix and the end-to-end validation are based on 1000 EZ light-cone mock catalogs. In $\Lambda$CDM, the redshift-weighted and standard analyses give consistent constraints, as expected from the near-standard effective redshifts of the weights targeting $h$, $\Omega_{\rm m}$, and $A_s$. In the Chevallier--Polarski--Linder (CPL) model, the redshift-weighted DR16 analysis reduces the marginalized uncertainties on $H_0$, $\sigma_8$, and $w_0$ by $43.3\%$, $19.7\%$, and $20.5\%$, respectively, and turns the standard one-sided constraint on $w_a$ into a bounded posterior, $w_a=-0.98^{+1.0}_{-1.3}$. The gain is therefore concentrated where the model contains genuine redshift evolution, demonstrating that optimal redshift weighting can recover tomographic information from a wide QSO light cone while keeping the full-shape data vector compact.

astro-ph.CO↗

arXiv2026

Towards Long-Horizon Vessel Trajectory and Destination Forecasting with Reasoning Large Language Models

Long-horizon maritime trajectory prediction is important for shipping management, logistics planning, and maritime risk analysis, yet month-level forecasting remains insufficiently studied. Existing deep learning methods mainly focus on short- and mid-term coordinate extrapolation and often struggle to preserve route feasibility and destination correctness over extended horizons. This paper investigates joint long-horizon vessel trajectory and destination forecasting with reasoning-capable large language models, and develops a Maritime LLM post-training framework based on Reinforcement Learning with Verifiable Reward (RLVR). An AIS-based benchmark is constructed with 60-day historical trajectories and 30-day forecasting horizons, where trajectories are converted into semantic textual representations for RL prompt construction. RLVR aligns LLMs with maritime forecasting objectives by enforcing physical validity, providing early-weighted trajectory supervision, and evaluating destination correctness through hierarchical matching and curriculum learning. Experimental results show that RLVR-trained LLMs substantially improve over zero-shot LLMs and representative deep learning baselines, especially on destination-related metrics. Among the evaluated RLVR-trained variants, 4B LLMs achieve the best overall performance, suggesting that reward-compatible optimization and task-specific capacity matching are more important than simply using larger 8B or 14B LLMs. The results also show that LSTM remains a strong deep learning baseline under limited fine-tuning data, while Transformer-style spatio-temporal models typically require larger datasets and richer structured inputs. Overall, this work advances semantic, verifier-aligned maritime forecasting for operational decision support.

cs.AI↗

arXiv2026

Tropical Cartan's second main theorem for hyperplanes in general position

We prove a tropical analogue of Cartan's second main theorem for holomorphic curves intersecting hyperplanes in general position--a setting that was not fully resolved by previous tropical Nevanlinna theory. Two versions are obtained. The first (Theorem 1.7) requires subnormal growth and involves the tropical Casorati determinant. The second and main version (Theorem 1.9) is completely free of growth conditions and exceptional sets; it replaces the Casorati term by the sum of the counting functions of the curve's components, yielding an inequality valid for every r. The proof uses a tropical Cramer theorem, bypassing the logarithmic derivative lemma. This improves upon previous results by Korhonen-Tohge and Cao-Zheng, where the coefficient could be suboptimal even under the general position hypothesis. We also clarify the relation between different notions of linear independence, and present the first counterexample to the truncated second main theorem in the tropical setting (Example 5.4).

math.AG↗

arXiv2026

Revisiting Uncertainty: On Evidential Learning for Partially Relevant Video Retrieval

Partially relevant video retrieval aims to retrieve untrimmed videos using text queries that describe only partial content. However, the inherent asymmetry between brief queries and rich video content inevitably introduces uncertainty into the retrieval process. In this setting, vague queries often induce semantic ambiguity across videos, a challenge that is further exacerbated by the sparse temporal supervision within videos, which fails to provide sufficient matching evidence. To address this, we propose Holmes, a hierarchical evidential learning framework that aggregates multi-granular cross-modal evidence to quantify and model uncertainty explicitly. At the inter-video level, similarity scores are interpreted as evidential support and modeled via a Dirichlet distribution. Based on the proposed three-fold principle, we perform fine-grained query identification, which then guides query-adaptive calibrated learning. At the intra-video level, to accumulate denser evidence, we formulate a soft query-clip alignment via flexible optimal transport with an adaptive dustbin, which alleviates sparse temporal supervision while suppressing spurious local responses. Extensive experiments demonstrate that Holmes outperforms state-of-the-art methods. Code is released at https://github.com/lijun2005/ICML26-Holmes.

cs.CV↗

arXiv2026

Efficient estimators for power spectrum and bispectrum multipole measurements

Large galaxy surveys demand fast and scalable estimators for anisotropic clustering statistics beyond the monopole. We present a suite of efficient FFT-based estimators for power-spectrum and bispectrum multipoles, built upon exact conjugation and parity symmetries of spherical-harmonic--weighted Fourier transforms of real fields. These symmetries eliminate redundant magnetic sub-configurations, thereby reducing the computational cost by a factor of 2. For the Yamamoto power-spectrum multipoles, we further decrease the cost of high-order even multipoles by algebraically expressing ${L}_{2n}$ in terms of lower-order Legendre polynomials, thereby measuring modified high-order multipoles using only low-$\ell$ fields with a small and controlled deviation from the traditional definition. We introduce a new TripoSH bispectrum estimator obtained by compressing the Scoccimarro bispectrum along an alternative triangle side, which substantially reduces the FFT scaling for commonly used quadrupole configurations in the large-$k$-bin limit. We also derive an analytic treatment of bispectrum shot noise by integrating spherical-harmonic kernels over the triangle-constrained $k$-space volumes, avoiding additional FFTs or costly spherical-Bessel evaluations and enabling fast and accurate shot-noise subtraction. Based on these optimizations, we also introduce CosmoNPC, an open-source Python package for large-scale-structure clustering measurements.

astro-ph.CO↗

arXiv2026

GLM-5V-Turbo: Toward a Native Foundation Model for Multimodal Agents

We present GLM-5V-Turbo, a step toward native foundation models for multimodal agents. As foundation models are increasingly deployed in real environments, agentic capability depends not only on language reasoning, but also on the ability to perceive, interpret, and act over heterogeneous contexts such as images, videos, webpages, documents, GUIs. GLM-5V-Turbo is built around this objective: multimodal perception is integrated as a core component of reasoning, planning, tool use, and execution, rather than as an auxiliary interface to a language model. This report summarizes the main improvements behind GLM-5V-Turbo across model design, multimodal training, reinforcement learning, toolchain expansion, and integration with agent frameworks. These developments lead to strong performance in multimodal coding, visual tool use, and framework-based agentic tasks, while preserving competitive text-only coding capability. More importantly, our development process offers practical insights for building multimodal agents, highlighting the central role of multimodal perception, hierarchical optimization, and reliable end-to-end verification.

cs.CV↗

arXiv2026

Imagine Before Concentration: Diffusion-Guided Registers Enhance Partially Relevant Video Retrieval

Partially Relevant Video Retrieval (PRVR) aims to retrieve untrimmed videos based on text queries that describe only partial events. Existing methods suffer from incomplete global contextual perception, struggling with query ambiguity and local noise induced by spurious responses. To address these issues, we propose DreamPRVR, which adopts a coarse-to-fine representation learning paradigm. The model first generates global contextual semantic registers as coarse-grained highlights spanning the entire video and then concentrates on fine-grained similarity optimization for precise cross-modal matching. Concretely, these registers are generated by initializing from the video-centric distribution produced by a probabilistic variational sampler and then iteratively refined via a text-supervised truncated diffusion model. During this process, textual semantic structure learning constructs a well-formed textual latent space, enhancing the reliability of global perception. The registers are then adaptively fused with video tokens through register-augmented Gaussian attention blocks, enabling context-aware feature learning. Extensive experiments show that DreamPRVR outperforms state-of-the-art methods. Code is released at https://github.com/lijun2005/CVPR26-DreamPRVR.

cs.CV↗

arXiv2026

TacSIm: A Dataset and Benchmark for Football Tactical Style Imitation

Current football imitation research primarily aims to opti mize reward-based objectives, such as goals scored or win rate proxies, paying less attention to accurately replicat ing real-world team tactical behaviors. We introduce Tac SIm, a large-scale dataset and benchmark for Tactical Style Imitation in football. TacSIm imitates the acitons of all 11 players in one team in the given broadcast footage of Pre mier League matches under a single broadcast view. Under a offensive or defensive broadcast footage, TacSIm projects the beginning positions and actions of all 22 players from both sides onto a standard pitch coordinate system. Tac SIm offers an explicit style imitation task and evaluation protocols. Tactics style imitation is measured by using spatial occupancy similarity and movement vector similarity in defined time, supporting the evaluation of spatial and tem poral similarities for one team. We run multiple baseline methods in a unified virtual environment to generate full team behaviors, enabling both quantitative and visual as sessment of tactical coordination. By using unified data and metrics from broadcast to simulation, TacSIm estab lishes a rigorous benchmark for measuring and modeling style-aligned tactical imitation task in football.

cs.CV↗

arXiv2026

Nonlinear Information from DESI Luminous Red Galaxies: An Emulator-Based Analysis of Pre- and Post-Reconstruction Power Spectra

We present joint measurements of the pre- and post-reconstruction power spectra, $P_{\rm pre}$ and $P_{\rm post}$, together with their cross-power spectrum, $P_{\rm cross}$, for the Luminous Red Galaxies (LRGs) in the DESI Data Release 1 (DR1). We jointly analyse these observables with an emulator-based full-shape modeling framework, thereby, for the first time, we extract complementary nonlinear information from the galaxy density field before and after reconstruction in real survey data. Specifically, including $P_{\rm post}$ and $P_{\rm cross}$ in addition to $P_{\rm pre}$ (hereafter $P_{\rm all}$) yields an improvement of approximately $18$-$27\%$ in the $\sigma_8$ constraint in both $\Lambda$CDM and $w$CDM, depending on the redshift bin, relative to the $P_{\rm pre}$-only analysis with the cosmic microwave background distance priors (hereafter CMB). In $w$CDM, the joint CMB+$P_{\rm all}$ analysis can tighten the constraints on $w$ by approximately $5$-$15\%$ across the two LRG redshift bins, compared to the CMB+$P_{\rm pre}$ combination. Further incorporating the Type Ia supernova dataset and comparing the cosmological constraints in $w$CDM from each individual power-spectrum component with those from the full combination, we find that $P_{\rm all}$ consistently provides the tightest constraints. From the joint CMB+$P_{\rm all}$+DES-Dovekie dataset, we obtain $\Omega_m = 0.314 \pm 0.0048$ and $w = -0.988 \pm 0.023$ for the \texttt{LRG1} sample, and $\Omega_m = 0.318 \pm 0.0046$ and $w = -0.988 \pm 0.025$ for \texttt{LRG2}. These results demonstrate that combining pre- and post-reconstruction power spectra with their cross-correlation enables DESI to harvest additional nonlinear information, leading to tighter constraints on cosmological parameters.

astro-ph.CO↗

arXiv2026

Flat Topological Nodal Lines in Heavy-Fermion Compound CeCoGe$_3$

The interplay between strong electronic correlations, unconventional superconductivity, and symmetry-protected topology provides a fertile ground for discovering exotic quantum states. In this work, we investigate the correlated electronic structure and topological properties of the heavy fermion material CeCoGe$_3$ using density functional theory combined with dynamical mean-field theory calculations. Our results reveal a crossover from high temperature incoherent states to low temperature coherent heavy quasiparticles, accompanied by a mass enhancement of $m^*/m_{\text{DFT}}\sim 52.6$ at $T=25$ K. The interplay between electronic correlation, spin-orbit coupling and the noncentrosymmetric $I4mm$ crystal symmetry stabilize flat topological nodal lines within 10 meV of the Fermi level, which could contribute a significant density of states. The proximity of topological nodal lines to the Fermi surface suggests a potential role in mediating pressure induced unconventional superconductivity. Our work establishes CeCoGe$_3$ as a prototype topological nodal line Kondo semimetal. The coexistence of strong correlation, non-trivial band topology and superconductivity indicate CeCoGe$_3$ as a potential candidate for realizing topological superconductivity.

cond-mat.str-el↗

arXiv2026

From Verbatim to Gist: Distilling Pyramidal Multimodal Memory via Semantic Information Bottleneck for Long-Horizon Video Agents

While multimodal large language models have demonstrated impressive short-term reasoning, they struggle with long-horizon video understanding due to limited context windows and static memory mechanisms that fail to mirror human cognitive efficiency. Existing paradigms typically fall into two extremes: vision-centric methods that incur high latency and redundancy through dense visual accumulation, or text-centric approaches that suffer from detail loss and hallucination via aggressive captioning. To bridge this gap, we propose MM-Mem, a pyramidal multimodal memory architecture grounded in Fuzzy-Trace Theory. MM-Mem structures memory hierarchically into a Sensory Buffer, Episodic Stream, and Symbolic Schema, enabling the progressive distillation of fine-grained perceptual traces (verbatim) into high-level semantic schemas (gist). Furthermore, to govern the dynamic construction of memory, we derive a Semantic Information Bottleneck objective and introduce SIB-GRPO to optimize the trade-off between memory compression and task-relevant information retention. In inference, we design an entropy-driven top-down memory retrieval strategy. Extensive experiments across 4 benchmarks confirm that MM-Mem achieves state-of-the-art performance on both offline and streaming tasks, demonstrating robust generalization and validating the effectiveness of cognition-inspired memory organization. Code and associated configurations are publicly available at https://github.com/EliSpectre/MM-Mem.

cs.CV↗

arXiv2026

RustCompCert: A Verified and Verifying Compiler for a Sequential Subset of Rust

We present our ongoing work on developing an end-to-end verified Rust compiler based on CompCert. It provides two guarantees: one is semantics preservation from Rust to assembly, i.e., the behaviors of source code includes the behaviors of target code, with which the properties verified at the source can be preserved down to the target; the other is memory safety ensured by the verifying compilation -- the borrow checking pass, which can simplify the verification of Rust programs, e.g., by allowing the verification tools focus on the functional correctness.

cs.PL↗

arXiv2026

Large-scale EM Benchmark for Multi-Organelle Instance Segmentation in the Wild

Accurate instance-level segmentation of organelles in electron microscopy (EM) is critical for quantitative analysis of subcellular morphology and inter-organelle interactions. However, current benchmarks, based on small, curated datasets, fail to capture the inherent heterogeneity and large spatial context of in-the-wild EM data, imposing fundamental limitations on current patch-based methods. To address these limitations, we developed a large-scale, multi-source benchmark for multi-organelle instance segmentation, comprising over 100,000 2D EM images across variety cell types and five organelle classes that capture real-world variability. Dataset annotations were generated by our designed connectivity-aware Label Propagation Algorithm (3D LPA) with expert refinement. We further benchmarked several state-of-the-art models, including U-Net, SAM variants, and Mask2Former. Our results show several limitations: current models struggle to generalize across heterogeneous EM data and perform poorly on organelles with global, distributed morphologies (e.g., Endoplasmic Reticulum). These findings underscore the fundamental mismatch between local-context models and the challenge of modeling long-range structural continuity in the presence of real-world variability. The benchmark dataset and labeling tool will be publicly released soon.

cs.CV↗

arXiv2026

AutoBinder Agent: An MCP-Based Agent for End-to-End Protein Binder Design

Modern AI technologies for drug discovery are distributed across heterogeneous platforms-including web applications, desktop environments, and code libraries-leading to fragmented workflows, inconsistent interfaces, and high integration overhead. We present an agentic end-to-end drug design framework that leverages a Large Language Model (LLM) in conjunction with the Model Context Protocol (MCP) to dynamically coordinate access to biochemical databases, modular toolchains, and task-specific AI models. The system integrates four state-of-the-art components: MaSIF (MaSIF-site and MaSIF-seed-search) for geometric deep learning-based identification of protein-protein interaction (PPI) sites, Rosetta for grafting protein fragments onto protein backbones to form mini proteins, ProteinMPNN for amino acid sequences redesign, and AlphaFold3 for near-experimental accuracy in complex structure prediction. Starting from a target structure, the framework supports de novo binder generation via surface analysis, scaffold grafting and pose construction, sequence optimization, and structure prediction. Additionally, by replacing rigid, script-based workflows with a protocol-driven, LLM-coordinated architecture, the framework improves reproducibility, reduces manual overhead, and ensures extensibility, portability, and auditability across the entire drug design process.

q-bio.BM↗

arXiv2026

Deep Learning Superresolution for 7T Knee MR Imaging: Impact on Image Quality and Diagnostic Performance

Background: Deep learning superresolution (SR) may enhance musculoskeletal MR image quality, but its diagnostic value in knee imaging at 7T is unclear. Objectives: To compare image quality and diagnostic performance of SR, low-resolution (LR), and high-resolution (HR) 7T knee MRI. Methods: In this prospective study, 42 participants underwent 7T knee MRI with LR (0.8*0.8*2 mm3) and HR (0.4*0.4*2 mm3) sequences. SR images were generated from LR data using a Hybrid Attention Transformer model. Three radiologists assessed image quality, anatomic conspicuity, and detection of knee pathologies. Arthroscopy served as reference in 10 cases. Results: SR images showed higher overall quality than LR (median score 5 vs 4, P<.001) and lower noise than HR (5 vs 4, P<.001). Visibility of cartilage, menisci, and ligaments was superior in SR and HR compared to LR (P<.001). Detection rates and diagnostic performance (sensitivity, specificity, AUC) for intra-articular pathology were similar across image types (P>=.095). Conclusions: Deep learning superresolution improved subjective image quality in 7T knee MRI but did not increase diagnostic accuracy compared with standard LR imaging.

eess.IV↗