Search arXiv⌕ Search

arXiv subjects

Yang Hu

Publications and source records attributed to Yang Hu.

At least 19 recordsLinked to original sources
arXiv2026

Engineering Canalized Near-Field Radiative Heat Transfer between Fixed Nanoparticles with Twisted $α$-MoO$_3$ Trilayers

Canalized phonon polaritons in twisted van der Waals crystals offer a route to control nanoscale thermal energy flow. Building on the established twisted $α$-MoO$_3$ trilayer platform, we examine how internal twist angles modify near-field radiative heat transfer (NFRHT) between a prescribed nanoparticle pair while both the nanoparticle axis and the bottom-layer crystallographic orientation remain fixed. Under this constraint and at equal total $α$-MoO$_3$ thickness, the optimized trilayer reaches 2.84 times the bilayer maximum radiative power, while its mean power over the sampled internal twist configurations is 1.51 times the bilayer value. The second internal twist coordinate also enlarges the set of configurations supporting strong transfer, with local angular tolerance when either internal angle is varied separately. Momentum-space Green-tensor diagnostics and real-space energy-density maps associate the enhancement with stronger receiver-directed scattering and improved channel alignment. Selected trilayer configurations retain enhanced coupling across the sampled family of nanoparticle resonances and at long distances. These results quantify the additional internal configurational freedom available for engineering radiative heat transfer under a fixed laboratory-frame geometry.

physics.optics↗
arXiv2026

From Generation to Scoring: Scaling LLM-Based Semantic Labeling at LinkedIn

For efficiency, industrial large language model (LLM) labelers often pack many candidate labels into one prompt and generate a YES/NO answer for each, even when each decision depends only on the entity and one candidate. Generating these independent decisions as one answer list couples them, spends decoding on the list, and relies on the list being well-formed. In this paper, we present an end-to-end recipe that reformulates the labeler as semantic scoring, fine-tuning it on single-candidate prompts to return each decision as one answer-token score from a prefill-only pass. We also adapt Multiple-Item Scoring (MIS) to score hundreds of candidates in one sequence with attention masked between them. Deployed for profile qualification labeling at LinkedIn, semantic scoring changes at most 0.01% of decisions when candidates are reordered and matches the agreement with reference labels that packed generation reaches only with about 4.7x as many training pairs. The MIS-based stack raises throughput over the deployed stack by 1.24x offline and 1.9x in model serving under a one-second p99 latency target. For offline batch inference, we implement MIS-based scoring as a reusable, model-agnostic Ray Data API proposed for upstream adoption. Model serving uses MIS support contributed to SGLang.

cs.LG↗
arXiv2026

Iterative Randomized Rounding for Better-Than-2 k-Median Approximation

We study the classic $k$-median problem and refine the iterative rounding framework recently introduced by Byrka, Guo, Hu, Li, Wan, and Wang [FOCS'26], which yields a $(2+\varepsilon)$-approximation. Our algorithm opens $k+O(1)$ facilities and achieves an expected connection cost of at most $1.9999$ times the value of the standard LP relaxation. Although the numerical improvement over $2$ is small, it establishes a conceptual result: the integrality gap of the standard relaxation is strictly below $2$ when a constant number of extra facilities is allowed. This partially answers an open question of Li and Svensson [STOC'13]. Combined with their reduction, our result yields a $1.9999$-approximation algorithm for $k$-median. Our approach builds on several technical ingredients. First, we develop a preprocessing routine that makes the LP solution $\mathrm{poly}(\varepsilon)$-integral, at the cost of a $(1+\varepsilon)$ factor in the connection cost and $O(1)$ extra facilities. Unlike the preprocessing of Byrka et al. [FOCS'26], our conversion is unconditional: it requires no special treatment of clients whose fractional facilities are concentrated. Next, we show that the main rounding algorithm can open facilities one at a time, eliminating the batch processing used by Byrka et al. [FOCS'26]. Finally, the improvement over 2 comes from a tighter analysis for clients whose distances from to its facilities are not all equal, and a complementary algorithm for the case in which these distances are nearly uniform.

cs.DS↗
arXiv2026

Data-driven accelerated multigrid method for Riesz space fractional diffusion equations

The multigrid method is one of the most efficient iterative methods for solving large structured linear systems, and its efficiency stems from the complementarity between the smoother and the coarse grid correction. In this paper, we consider the block Toeplitz linear system obtained by discretizing time dependent Riesz space fractional diffusion equations, and employ the weighted Jacobi iterative method as a smoother. The weighted Jacobi smoother features a simple structure and low computational cost, yet its smoothing property is highly dependent on the relaxation parameter. We adopt Gaussian process regression to select the relaxation parameter to accelerate the convergence of the multigrid method. Numerical experiments confirm the theoretical analysis, which simultaneously verifies the effectiveness and superiority of the data-driven parameter selection strategy.

math.NA↗
arXiv2026

Experimental Asynchronous Measurement-Device-Independent Quantum Cryptographic Conferencing

The quantum cryptographic conferencing (QCC) protocol, which distributes identical secure keys to user groups, is a crucial component of the quantum network. Previous experimental works have implemented the measurement-device-independent (MDI) QCC, of which the key rate in an N-user network scales as $R\sim O(η^N)$. Building on the MDI QCC protocol, the asynchronous MDI QCC protocol theoretically integrates the mode-pairing scheme into QCC, significantly boosting the key rate to $R\sim O(η)$ in the ideal case, which is independent of the number of users, thus demonstrating greater application potential. Experimentally, in this Letter, we implement the three-user asynchronous MDI QCC network without phase locking by adopting the fast Fourier transform-based frequency difference estimation and the phase drift compensation technique. Finally, we achieve a key rate of about $4.470\times 10^{-9}$ bits per pulse under a maximum overall loss of about 59.6 dB. This Letter provides a scalable solution for the development of large-scale quantum communication networks in the future.

quant-ph↗
arXiv2026

Detecting Inconsistencies in Model Specifications with LLM-as-Verifier Reasoning

Model specifications define how large language models (LLMs) should behave, guiding alignment training, inference-time behavior, and evaluation. Yet these specifications may themselves contain defects: two individually reasonable principles may prescribe incompatible behavior when applied to the same situation, leaving no response that satisfies both. Detecting such inconsistencies is challenging. Formalizing natural-language specifications risks losing subtle distinctions, while behavior-based testing cannot reliably distinguish specification defects from differences in model behavior. We introduce VeriSpec, the first approach to directly detect inconsistencies in model specifications by auditing the specification text itself. Our key insight is to preserve the specification in natural language while using an LLM as a verifier. VeriSpec extracts structured, context-aware rules, constructs a topic-guided graph to cluster behaviorally related rules at the same authority level, and applies LLM-as-verifier reasoning to detect inconsistencies. Applying VeriSpec to the OpenAI Model Spec, we extract 405 rules and manually validate five inconsistencies, all reported to its developers, who responded positively and have initiated internal discussions. Compared with five baselines, VeriSpec identifies the most validated inconsistencies, achieves the highest precision (38.5%), and incurs the lowest cost per validated inconsistency ($11.12). These results establish direct specification auditing as a practical complement to behavioral alignment evaluation, catching defects at the source before they shape any model. The code is available at https://github.com/HIPREL-Group/VeriSpec.

cs.SE↗
arXiv2026

It's the Geometry, Not the Model: Effective Rank and Subspace Alignment in Functional Connectivity Classification

Resting-state functional connectivity (FC) is widely used to classify brain phenotypes and disorders. Most pipelines use the full connectome and seek gains through model design. We instead examine how FC geometry constrains classification and cross-site transfer. Across-subject FC variation concentrates in a small effective subspace, suggesting substantial redundancy in nominal dimensions. Across cohorts, these subspaces may differ in orientation even when their effective ranks are comparable, potentially limiting transfer. Across 2,330 subjects from HCP, ABIDE, and ADHD-200, effective-rank analysis reveals strong spectral concentration. Projection onto leading components at the effective-rank scale recovers most of the full-FC classification performance. In ABIDE, site-specific effective subspaces are weakly aligned, and their principal-angle overlap predicts pairwise transfer after covariate adjustment despite comparable per-site effective ranks. Controlled rotations that alter subspace orientation while preserving the mean and covariance spectrum drive transfer toward chance, whereas displacement-matched label-orthogonal rotations do not. These results identify subspace orientation as a key factor in transfer degradation under controlled perturbations. This study offers a geometric diagnostic of FC generalization and suggests evaluating cross-site harmonization by its ability to align effective subspaces alongside classification accuracy.

cs.CG↗
arXiv2026

A Fast and Scalable Transformer Pipeline for Binary Black Hole Detection

With the projected increase in the detection rate of compact-binary coalescences in the coming decade, there is critical need to develop fast, robust, and scalable alternatives to matched filtering for gravitational-wave searches. Transformer models have revolutionized natural language and audio processing but their application to gravitational-wave astronomy is still largely unexplored. In this work, we introduce \castor, a transformer-based coincident search pipeline for detecting binary black hole gravitational-wave signals from Advanced LIGO detectors. One of the major features of our model is that it allows the false-alarm rate to be estimated via time slides cheaply without requiring repeated evaluations of the neural network. We evaluate \castor\ on datasets from the Machine-Learning Gravitational-Wave Search Challenge (MLGWSC-1) and on approximately five months of real O3b observing strain. When tested on benchmark datasets, \castor\ ranks among the most sensitive machine-learning pipelines and successfully recovers the majority of confident events from the GWTC-3 catalog that lie within its training range. We also benchmark \castor\ against another transformer architecture, GW-Whisper, a domain-adaptation of OpenAI's audio foundation model. We find that \castor\ substantially outperforms the repurposed audio model in sensitivity and also reduces the computational cost of background estimation by a factor of 20. Our results demonstrate a highly practical, scalable approach for deep-learning gravitational wave searches and empirical background estimation for future observing runs.

gr-qc↗
arXiv2026

D-IMPL: A Diffusion-based Solver for Parameterized BBOs

Diffusion models have demonstrated strong power in generative modeling tasks across multiple domains, exhibiting a remarkable capability of learning complex distributions from samples. In this paper, we leverage such capability to design an efficient universal diffusion-based solver for parameterized black-box optimizations (BBO), where the optimizer has only black-box access to queries of the objective function at the learning stage, yet is able to reduce the additional computational cost at the inference stage for each BBO instance while also capturing the potential multi-modal landscape of non-convex objectives. To cast our formulation as a compatible generative modeling task, we introduce the notion of minimization policy as a new solution concept, which defines a sampling distribution over the solutions that should concentrate around the minimizer set for each BBO instance. We then propose Diffusion-based Iterative Minimization Policy Learning (D-IMPL), a practical generative-model-based solver for solving parameterized BBOs that employs diffusion models to learn a minimization policy, whose density is proportional to the exponential of the negated objective values, thereby amortizing the computational costs across different BBO instances. Furthermore, we demonstrate the performance of our D-IMPL algorithm by establishing a sample complexity guarantee showing that a $δ$-approximate minimization policy can be effectively learned within $O(\log(1/δ))$ iterations, and by extensive empirical evaluations over a range of constrained and unconstrained BBO tasks.

cs.LG↗
arXiv2026

A Thresholding Based Operator-Splitting Method for Curvature-Regularized Surface Reconstruction

Surface reconstruction from point clouds is a fundamental problem in computational geometry with broad applications in computer graphics, medical imaging, and manufacturing. In this paper, we propose an efficient method for solving a curvature-regularized surface reconstruction model. To avoid the computational cost associated with reinitialization in traditional level set methods, we represent the reconstructed surface by an indicator function. By introducing an auxiliary variable, we reformulate the original optimization problem as the computation of the steady-state solution of an initial value problem. We then develop an operator-splitting method to decompose the resulting problem into two tractable subproblems, one of which can be efficiently solved by iterative thresholding. The proposed approach combines the advantages of curvature regularization, operator splitting, and threshold dynamics. Numerical experiments show that the proposed method is computationally efficient and can accurately reconstruct sharp corners and concave features.

math.NA↗
arXiv2026

Quadratic Probing Insertions Are $ε^{-(1+o(1))}$ Time

First proposed in 1968, quadratic probing has stood for more than half a century as one of the simplest and most widely used hash-table designs in computer science. It is conjectured that, at load factor $1 - ε$, the hash table achieves $O(ε^{-1})$ expected insertion time. But even proving a bound of the form $f(ε^{-1})$ for any function $f$ has remained open. In this paper, we prove that the expected insertion time is $ε^{-(1 + o(1))}$. This settles the complexity of the data structure up to sub-polynomial factors in $ε^{-1}$.

cs.DS↗
arXiv2026

WaferTrans: Enabling IOMMU-free Distributed Virtual Address Translation for Wafer-scale GPUs

Wafer-scale GPUs (WSGs) provide sufficient on-wafer bandwidth to make near-lossless Unified Memory feasible. However, existing designs still rely on a CPU-IOMMU to translate remote virtual-address accesses. This centralized mechanism scales poorly to tens of GPU dies: translation requests must traverse costly off-wafer hierarchies and contend for limited CPU-side resources, making address translation a critical bottleneck. We propose WaferTrans, an IOMMU-free distributed virtual-address translation design for WSGs. WaferTrans introduces PTE Presence Consistency (PTE-PC), a lightweight consistency model that tracks PTE insertions and removals, and equips each GPU with a PTE Presence Directory (PPD) that locates the GPU holding a requested PTE. It further employs a distributed PTE-PC mapping and a cooperative query mechanism to localize PTE-PC maintenance while preserving complete lookup coverage. Together, these mechanisms enable the GPU array to resolve remote translations within the wafer, eliminating its dependence on the CPU-IOMMU. Compared with the SOTA Trans-FW design, WaferTrans improves performance by 2.5x on average.

cs.AR↗
arXiv2026

Realization of some Stanley-Reisner algebras and graph colorings

It is a classical problem in algebraic topology to decide whether a given graded $\mathbb{Z}$-algebra can be realized as the cohomology ring of a space. In this paper, we introduce families of Stanley-Reisner algebras depending on graphs, and relate their realizability to the span coloring of the graph.

math.AT↗
arXiv2026

High-corank torsion in homotopy of unitary groups via topological modular forms and higher real $K$-theories

Work of Toda identifies groups of metastable vector bundles on even-dimensional spheres with stable homotopy groups of certain stunted projective spectra. Using Weiss' unitary calculus, the second-named author generalized this identification to show that metastable, stably trivial vector bundles on even cell complexes can naturally be identified with stable homotopy classes of maps into a shifted stunted projective spectrum. Thus, certain classical questions about vector bundles (or homotopy of unitary groups) can be rephrased as stable computations. In this note, we show that certain generalized cohomology theories arising in chromatic and equivariant homotopy theory can be used to deduce the existence of non-trivial, stably trivial vector bundles on spheres and complex projective spaces.

math.AT↗
arXiv2026

Enumerating stably trivial vector bundles with higher real $K$-theory

This paper explores periodic phenomena in the group $\operatorname{Vect}_r^0(\mathbb{CP}^{r+c})$ of stably trivial, complex rank $r$ topological vector bundles on $\mathbb{CP}^{r+c}$. For $1 \leq c < r$ and $c\leq 2p-3$, we give a complete computation of the $p$-torsion in $\operatorname{Vect}_r^0(\mathbb{CP}^{r+c})$, and we relate these $p$-torsion bundles to vector bundles on spheres. We also compute $\operatorname{Vect}_r^0(\mathbb{CP}^{r+c})$ in full when $c=3$, extending computations of the second-named author when $c=1$ and $c=2$. Finally, for a fixed corank $c$ which is larger relative to the prime $p$, we show there are families of $p$-torsion in $\operatorname{Vect}_r^0(\mathbb{CP}^{r+c})$ that are periodic in $r$ with period $p$. We detect these using Weiss calculus and certain higher real $K$-theory homology groups.

math.AT↗
arXiv2026

Once Generated, Ranked: End-to-End Generative Slate Recommendation with Unified Semantic-Collaborative IDs

Slate recommendation treats a slate rather than an individual item as the recommendation unit, requiring joint optimization of item interactions and slate utility. Existing approaches typically separate candidate generation from ranking and restrict optimization to retrieved candidates. Generative recommendation with Semantic IDs (SIDs) offers a path to end-to-end recommendation, but existing SID construction often lacks recommendation-aware semantics and effective local collaborative signals, while next-token prediction is misaligned with slate-level objectives. We propose OGR, an end-to-end framework that directly generates ordered slates-"Once Generated, Ranked." OGR first introduces TUSID, which adaptively fuses item-specific semantic and local collaborative information into hierarchical SIDs. It then uses list-wise preference planning and pipelined position-wise SID decoding to model global preferences and inter-item dependencies while generating ordered slates. We further propose SPA, a reward-guided conservative policy optimization method that aligns generated slates with user preferences beyond likelihood imitation. Offline experiments show that OGR outperforms representative baselines, with 48.2% and 27.2% relative NDCG@5 gains on industrial and public datasets, respectively. Online A/B testing on Kuaishou further yields a 1.120% improvement in Effective Views.

cs.IR↗
arXiv2026

Spatial Message Passing in Language Space for Pathology Image Interpretation

Multimodal Large Language Models (MLLMs) can generate pathological descriptions from histological images, but gigapixel Whole Slide Images (WSIs) exceed their visual context limits. The standard tiling workaround makes WSIs tractable yet severs the tissue neighborhoods that define tumor-stroma interfaces and morphology. We introduce Spatial Language Message Passing (SLMP), a framework that performs spatial reasoning entirely in language space, human-readable by construction. SLMP represents a WSI region as a spatial text graph: tiles are nodes initialized with MLLM descriptions, and edges encode spatial adjacency. For each tile, an LLM refines its description by integrating language messages from adjacent tiles under a shared aggregation policy that, on the tile grid, acts as an adaptive local kernel operating on text rather than learned embeddings. This policy is an inspectable prompt that can be refined from model-observed tissue phenotypes via textual gradients, enabling automatic semantic optimization from local cellular context to broader tissue morphology without fine-tuning MLLM weights. On representative HER2 and CAMELYON16 regions, SLMP improves tile-level tumor description accuracy in settings spanning general-purpose and pathology-specialized backbones, with gains of +3.3 to +19.6 percentage points. Random-neighbor ablations confirm that these gains stem from spatial context rather than additional text alone, and inspecting the optimized policies reveals interpretable, tissue-specific decision rules. Besides, without any weight updates or fine-tuning the backbone MLLM, SLMP substantially improves general-purpose MLLMs and narrows its gap to pathology-specialized counterparts, offering a transparent and flexible mechanism for incorporating spatial reasoning into MLLM-based pathology analysis.

cs.CV↗