Searcharxiv⌕ Search

arXiv subjects

Hao Zhang

Publications and source records attributed to Hao Zhang.

At least 91 records · Page 5Linked to original sources
arXiv2026

Twin-photon generation in a silicon nitride microresonator

Photonic chips with silicon nitride ($\mathrm{Si_3N_4}$) microring resonators are well established as heralded single-photon sources, but their operation as frequency-degenerate twin-photon sources has not previously been demonstrated. Here, we realise a twin-photon source at telecommunication wavelengths in a $\mathrm{Si_3N_4}$ ring microresonator via an inverse four-wave mixing (FWM) process, in which two photons from spectrally distinct pumps are converted into a pair of identical twin photons. The measurements show a maximum coincidence-to-accidental ratio (CAR) of $5.4\pm0.6$. In addition, the microresonator functions as a heralded single-photon source through pump-degenerate spontaneous four-wave mixing (SFWM), exhibiting a spectral purity of $P=0.67\pm0.05$ and a heralded anti-bunching of $g^{(2)}_h(0)=0.0042\pm0.0015$. Together, these results demonstrate both photon-generation schemes on a single integrated $\mathrm{Si_3N_4}$ platform, highlighting its potential for scalable, tailored quantum light generation.

physics.optics↗
arXiv2026

Towards AI-Driven Nanomedicine Discovery: A Benchmark and Multimodal Learning Framework for Nano Self-Assembly Prediction

Nano self-assembly organizes molecular components into bioactive nanoscale structures. Self-assembled nanoparticles (NAPs) derived from Chinese herbal formulas and applications such as anti-lung-cancer therapy demonstrate the substantial potential of self-assembly for nanomedicine discovery. Yet discovery still relies on costly wet-lab screening, while existing machine learning approaches lack standardized tasks, effective pairwise compatibility modeling, and public benchmarks with unified evaluation. To address these limitations, we formalize NSA prediction as a binary classification task for predicting self-assembly between molecular pairs and then establish NSA-Bench, the first public benchmark with curated molecular combinations, experimental conditions, self-assembly labels, and standardized evaluation protocols. We further develop NSA-Net, an interaction-aware multimodal framework that integrates complementary molecular evidence from graph topology, sequence semantics, and physicochemical descriptors to learn molecular-pair representations for self-assembly prediction. Extensive experiments on NSA-Bench show that NSA-Net achieves a ROC-AUC of $0.9470\pm0.0112$ (Small) and $0.9492\pm0.0062$ (Large). On the Small track, it surpasses the strongest machine-learning and graph-based baselines by 3.9 and 17.1 percentage points, respectively. Representation analyses reveal interpretable molecular characteristics associated with self-assembly prediction captured by the learned representations. Moreover, an NSA-Agent case study further demonstrates how NSA-Net predictions can support formulation refinement through experimental-condition-aware reasoning. Our code is available at https://github.com/developer-hq/NSA-Net.

q-bio.QM↗
arXiv2026

Error estimate of the nonuniform BDF3-L2 method for subdiffusion equations via multiscale solution decomposition

Numerical experiments reported by Quan and Wu [SIAM J Numer Anal 61 (2023) 2106-2132] show that the observed temporal convergence rates of nonuniform L2 methods for subdiffusion models are not consistent with the theoretically predicted order $3-α$. This discrepancy suggests that a more refined analysis is needed and motivates the development of a nonuniform BDF3-L2 method for the subdiffusion equation. To account for the initial solution singularity, we employ the multiscale solution decomposition to decompose the original solution and approximate a smoother unknown variable that satisfies the subdiffusion model with a smoother source term. The resulting formulation, however, involves restrictive high-order boundary conditions on the source term and initial data. To overcome this difficulty, we introduce a spectral truncation technique that requires only slightly stronger regularity of the data and a controllable truncation error. We establish high-order regularity estimates of the solution to the truncated problem and develop a nonuniform BDF3-L2 method for its numerical approximation, based on which we derive a rigorous error estimate of temporal convergence order $2+α$. Numerical experiments are carried out to substantiate the theoretical findings.

math.NA↗
arXiv2026

ProbeMatchDTI: Probe-Driven Multi-Scale Biochemical Pattern Matching for Drug-Target Interaction Prediction

Drug-target interaction (DTI) prediction is an important task in AI-driven drug discovery. Although recent biochemical representation learning methods have improved DTI prediction, their passive feature aggregation tends to favor dominant molecular patterns while suppressing weak yet binding-relevant signals, such as functional groups and residue-context patterns, limiting the modeling of multi-scale biochemical correspondences. To address this issue, we propose ProbeMatchDTI, a pattern-probe-driven framework comprising IterProbe and BindingProbe. IterProbe explicitly retains contextual states across refinement depths and uses learnable probes to select them at each position before cross-entity matching, thereby preserving weak biochemical patterns and strengthening associations among functional groups, local motifs, and molecular scaffolds. BindingProbe then characterizes cross-entity drug-protein complementarity at local biochemical-unit and whole-pair levels, jointly modeling fine-grained interactions and multi-scale correspondences while preserving weaker binding-relevant associations. Extensive experiments demonstrate the superiority of ProbeMatchDTI, achieving 2.0% and 0.5% higher AUC-ROC on BindingDB and DrugBank, respectively. Feature-level pattern analyses further characterize its probe-driven behavior in cross-scale biochemical pattern matching. We further connect ProbeMatchDTI predictions with an evidence-guided downstream drug-discovery workflow, demonstrating their utility for candidate refinement and validation planning. Our code is available at https://github.com/developer-hq/ProbeMatchDTI

cs.LG↗
arXiv2026

ByteX: A Unified AI Search Engine at ByteDance

Since 2016, ByteX has been the foundation of ByteDance's search infrastructure, scaling to more than 7,000 clusters and 300 PB of indexed data. Driven by the demands of AI workloads, ByteX has evolved from a text search engine into a unified AI search system supporting vector retrieval, lexical matching, and predicate filtering. Its largest deployment indexes nearly one trillion high-dimensional vectors. This scale exposes two central bottlenecks in AI-era retrieval: memory-intensive graph-index construction under sustained ingestion, and the prohibitive cost of keeping vector indexes entirely in memory. ByteX addresses these bottlenecks with two techniques. First, it introduces a quantization-aware vector kernel based on SymRaBitQ, a new symmetric quantization scheme with tight theoretical guarantees that allows index construction to run directly in the quantized space accurately and efficiently without retaining a copy of full-precision vectors. Second, it provides a hybrid storage engine that supports memory-resident, hybrid, and SSD-resident deployments, with fine-grained record-level caching to trade memory for latency under operational control. On large-scale benchmarks, ByteX improves throughput by up to 3x, reduces indexing memory by 80%, and lowers operating cost by 86% compared with prior systems, while supporting trillion-vector scale, write-heavy or latency-sensitive workloads in production.

cs.DB↗
arXiv2026

CAT-Flow: Curvature-Adaptive sTeps for Flow Matching

Flow Matching has emerged as a leading framework for generative modeling, powering state-of-the-art systems such as FLUX and Stable Diffusion 3.5. However, the iterative nature of its ODE-based sampling process creates a fundamental efficiency bottleneck: the quality of generated samples is highly sensitive to the choice of step-sizes, and current models typically require 20 to 30 steps for good quality. In this work, we propose two lightweight, training-free algorithms, CAT-OV and CAT-OT that adapt step-sizes at inference time based on a novel connection between Flow Matching sampling and gradient flow. Our algorithms are computed efficiently by not requiring additional neural function evaluations. Specifically, CAT-OT estimates curvature over time via a finite-difference approximation of the time-derivative of the vector field, while CAT-OV approximates curvature over the state space via a gradient of the vector field. Under suitable conditions, both methods have truncation error bounds of constant order. Empirically, CAT-OV and CAT-OT outperform existing step-size heuristics in image quality metrics across four text- to-image Flow Matching models, reducing the number of generation steps required to reach comparable quality by up to 40%.

cs.LG↗
arXiv2026

CM2: Multimodal Cultural Reasoning via an Integrated Multi-Agent Framework

Multimodal Large Language Models (MLLMs) have shown remarkable success in STEM domains, where progress is often driven by vertical, step-by-step deduction under relatively stable symbol systems. Their horizontal, interdisciplinary cultural reasoning, however, remains underexplored.We propose CM2, a multi-agent framework grounded in the cognitive pathway of human cultural interpretation. CM2 integrates multimodal perception, retrieval-augmented generation, networked reasoning, gated fusion, and reward-driven feedback.Experiments on CM2D across multiple MLLM backbones show consistent gains over CoT and typical reasoning paradigms; ablations validate each module's contribution, and conflict analyses confirm genuine cross-modal arbitration.

cs.AI↗
arXiv2026

Exclusive Leptonium Electroproduction

Purely leptonic bound states provide precision probes of QED. Positronium $(e^+e^-)$ and muonium $(μ^+e^-)$ have long been observed, whereas dimuonium $(μ^+μ^-)$ and tauonium $(τ^+τ^-)$ remain undiscovered. We study exclusive vector-leptonium electroproduction in $ep$ collisions within nonrelativistic QED. We include the Bethe--Heitler and double deeply virtual Compton scattering contributions and their interference, and calculate the NLO QCD hard-scattering kernels entering the dominant Compton form factor $\Hcal$ within collinear GPD factorization. The NLO QCD correction to the DDVCS contribution changes from a strong suppression at low photon virtuality to a sizable enhancement as the lower virtuality cut is raised, with the gluon channel providing the dominant contribution. Bethe--Heitler production dominates the exclusive rate, supporting dedicated dimuonium searches at the EIC and JLab, with larger samples expected at higher-energy electron--proton colliders. The much larger positronium samples provide a high-statistics environment for precision QED studies, whereas tauonium production remains strongly suppressed.

hep-ph↗
arXiv2026

Large zeta sums and zeros of the Riemann zeta function

For real $t$ and $x\ge 1$, set \[ S(x,t)=\sum_{n\le x} n^{\ii t}. \] We prove an unconditional inverse theorem relating large values of $S(x,t)$, with $|t|$ large, to zeros of the Riemann zeta function near height $t$. More precisely, if $T\le |t|\le 2T$, $\exp(\sqrt{\log T})\le x\le \sqrt T$, and $|S(x,t)|=x/N$ with $N\le (\log x)^{1/100}$, then for every $cN^6\le L\le (\log x)/2$ a disk centered at $1+\iiϕ$, where $|ϕ-t|\ll N$, contains at least $L/360$ zeros of $ζ(s)$. As a consequence, a quantitative restriction on zeros in a short family of arbitrarily thin fixed windows to the left of the line $\Ree s=1$ yields $S(x,t)\ll x/(\log x)^{1/100}$ in polynomial ranges of $x$. The proof adapts the zero-forcing mechanism of Granville and Soundararajan for large character sums. In the zeta setting the spectral height is shifted by $t$, and an additional residue from the pole of $ζ(s)$ appears in the Gaussian transform; in the range considered here that residue is exponentially small.

math.NT↗
arXiv2026

SORT: A Systematically Optimized Ranking Transformer for Industrial-scale Recommenders

While Transformers have achieved remarkable success in LLMs through superior scalability, their application in industrial-scale ranking models remains nascent, hindered by the challenges of high feature sparsity and low label density. In this paper, we propose SORT (Systematically Optimized Ranking Transformer), a scalable model designed to bridge the gap between Transformers and industrial-scale ranking models. We address the high feature sparsity and low label density challenges through a series of optimizations, including request-centric sample organization, local attention, query pruning, and generative pre-training. Furthermore, we introduce a suite of refinements to the tokenization, multi-head attention (MHA), and feed-forward network (FFN) modules, which collectively stabilize the training process and enlarge the model capacity. To maximize hardware efficiency, we optimize our training system to elevate the model FLOPs utilization (MFU) to 45%. Extensive experiments demonstrate that SORT outperforms strong baselines and exhibits excellent scalability across data size, model size, and sequence length, while remaining flexible at integrating diverse features. Finally, online A/B testing in large-scale e-commerce scenarios confirms that SORT achieves significant gains in key business metrics, including orders (+7.47%), buyers (+6.67%), and GMV (+8.65%), while simultaneously cutting latency by 62% and boosting throughput nearly sevenfold (+589%). SORT has been fully deployed in production, serving all users on AliExpress.

cs.IR↗
arXiv2026

Large-Scale Bayesian Tensor Reconstruction via Approximate Message Passing

While CANDECOMP/PARAFAC (CP) decomposition (CPD) is fundamental for tensor reconstruction, Bayesian CPD often scales poorly because variational updates require repeated matrix inversions. We develop CP generalized approximate message passing (CP-GAMP) for incomplete noisy Bayesian CPD. The algorithm uses Gaussian message approximations to avoid high-dimensional inversions, and it combines a Bernoulli-Gaussian prior with expectation-maximization updates to estimate effective CP rank and noise variance. We also give a formal state evolution (SE) recursion and relate its fixed points to replica-symmetric saddle points, so CP-GAMP's SE-predicted error can be compared with the formal replica-symmetric minimum mean-squared error (MMSE) benchmark in the matched limit. Synthetic and image-inpainting experiments show that CP-GAMP substantially reduces runtime relative to variational Bayesian CPD while maintaining competitive reconstruction accuracy.

cs.LG↗
arXiv2026

Learning to Allocate Incentives for Incentivized Advertising via Offline Model-Based Reinforcement Learning

Complete your ad view and grab a 5-cent bonus! In incentivized advertising, a platform promises users a bonus before observing downstream ad revenue, encouraging them to click and complete ads. It must balance the incentive promised in advance against the revenue realized afterward: insufficient incentives forfeit monetization opportunities, whereas excessive incentives reduce net profit. Because current incentives may also shape user expectations and future engagement, incentive allocation is a sequential decision problem with delayed revenue, cost sensitivity, and carryover effects. Existing work has not studied decision-making algorithms for this setting. Auto-bidding assumes available ad opportunities, while targeted promotion optimizes incentives outside the ad monetization pipeline. We formulate the problem as an MDP and develop an offline model-based RL framework for cost-controllable sequential incentive allocation. It learns a world model of user feedback and ad revenue, then performs conservative policy optimization. An independent counterfactual scorer evaluates each learned policy on held-out logs, enabling pre-launch selection without costly online exposure. Experiments on large-scale industrial data and online A/B tests show that the scorer provides a stable offline signal. The deployment path from causal inference to offline RL and then Offline-MBRL further validates the framework: MB-IQL improves per-user net profit by 7.96\% over TD3+BC, whereas reverting to plain IQL reduces it by 6.56\% (both \(p<0.0001\)).

cs.AI↗
arXiv2026

An Attention-Assisted AI Model for Real-Time Underwater Sound Speed Estimation Leveraging Remote Sensing Sea Surface Temperature Data

The estimation of underwater sound velocity distribution serves as a critical basis for facilitating effective underwater communication and precise positioning, given that variations in sound velocity influence the path of signal transmission. Conventional techniques for the direct measurement of sound velocity, as well as methods that involve the inversion of sound velocity utilizing acoustic field data, necessitate on--site data collection. This requirement not only places high demands on device deployment, but also presents challenges in achieving real-time estimation of sound velocity distribution. In order to construct a real-time sound velocity field and eliminate the need for underwater onsite data measurement operations, we propose a self-attention embedded multimodal data fusion convolutional neural network (SA-MDF-CNN) for real-time underwater sound speed profile (SSP) estimation. The proposed model seeks to elucidate the inherent relationship between remote sensing sea surface temperature (SST) data, the primary component characteristics of historical SSPs, and their spatial coordinates. This is achieved by employing CNNs and attention mechanisms to extract local and global correlations from the input data, respectively. The ultimate objective is to facilitate a rapid and precise estimation of sound velocity distribution within a specified task area. The comparative analysis demonstrates that the proposed approach achieves superior performance in terms of both accuracy and stability, exhibiting reduced error rates and enhanced resistance to disturbances when benchmarked against existing advanced techniques.

eess.SP↗
arXiv2026

CoGeo-GS: Concept-Driven and Geometry-Aware Multi-Object Removal in 3D Scenes

Multi-object removal in 3D scenes is challenging due to severe occlusions, semantic entanglement, and the difficulty of maintaining geometric and multi-view consistency. Existing 3D Gaussian Splatting (3DGS) methods perform well for single-object editing but scale poorly to multi-object scenarios, often requiring repetitive optimization and yielding unstable geometry in removed regions. We propose CoGeo-GS, a concept-driven framework for controllable multi-object removal in 3D scenes. CoGeo-GS assigns concept-aware semantic tags to Gaussians, enabling flexible object selection and reducing interference between foreground objects and background structures within a single optimization stage. To recover plausible geometry, we introduce a geometry-aware completion pipeline that combines monocular depth priors with diffusion-based refinement and boundary-aligned blending. A geometry-regularized refinement strategy further stabilizes reconstruction and preserves multi-view consistency. Experiments demonstrate that CoGeo-GS outperforms existing methods in visual quality and reconstruction fidelity.

cs.CV↗
arXiv2026

Astar: Learning to Propose Evolution Directions for Self-Evolving Industrial AI Systems

Modern AI systems advance through continuous iteration: a loop of proposing evolution directions, implementing code, training, and evaluation. While the latter three stages are increasingly automated, the starting point --- proposing effective evolution directions --- remains a critical bottleneck that still relies heavily on senior experts. In this work, we explore whether AI can take over this role. We find that general-purpose LLMs, even the advanced GPT-5.5, offer only generic and misaligned suggestions: the required expertise is accumulated through experience rather than explicitly codified, and thus hard to inject directly. To this end, we propose Astar, a training-based approach that learns a specialized evolution-guiding model from the abundant iteration histories of industrial systems. Realizing this idea, however, raises four challenges: sparse supervision, noisy data, a vast direction space, and prohibitively expensive verification. We address them along two fronts. On the data side, we design a pipeline that turns noisy historical commits into a large, clean evolutionary corpus via pairwise sample expansion and noise filtering. On the model side, we train the model through mid-training, SFT, and RL, guiding evolution direction generation with hierarchical hints and using the reward model in RL as a fast surrogate evaluator. Astar has been deployed in Alibaba's Lazada advertising system for evolution direction proposal. Astar-8B achieves a single-proposal success rate of 0.6786 in real-execution evaluation, far exceeding human experts (0.3229) and the strongest general-purpose LLM (0.3071). More importantly, Astar closes the loop and enables fully automatic iteration: it guided 20 consecutive iterations over two weeks, improving offline Hitrate@200 by 23.6%, while an online A/B test yielded relative lifts of 4.86% in GMV and 1.82% in advertising revenue.

cs.IR↗
arXiv2026

Stochastic transport of a Goldstone mode in a self-organized atomic crystal

Spontaneous breaking of a continuous symmetry produces a massless Goldstone mode that can evolve across a degenerate manifold at zero energy cost. Goldstone modes have been identified primarily through excitation spectra, mode softening or collective oscillations. However, their time-domain transport under intrinsic fluctuations and dissipation has remained largely unexplored. Here we directly track the stochastic transport of a Goldstone mode in a self-organized atomic crystal inside an optical ring cavity. The ring cavity maps the order-parameter phase onto the real-space position of the emergent crystal. Without any external perturbation, fundamental photon-scattering recoil drives the collective transport, while cavity dissipation generates friction. We monitor individual trajectories of the atoms and their self-generated optical lattice by measuring the cavity output phase. We find that the diffusion constant decreases as $1/N$, indicating that all atoms move collectively as a rigid object rather than independently. By tuning the Langevin driving force and cavity-mediated damping, we show that the normalized diffusion constant collapses onto a single universal curve. This work extends the study of continuous symmetry breaking from excitation-frequency measurements to real-time tracking of transport, and opens routes for studying non-equilibrium collective transport, phonon dynamics, and defect formation in driven-dissipative quantum matter.

quant-ph↗
arXiv2026

Realization of Air-Stable Two-Dimensional Superconductor Nb2Pd3Te5 With Quasi-One-Dimensional Pair Density Modulation

Two-dimensional (2D) superconductors provide a fertile platform for exploring reduced-dimensional superconductivity and emergent quantum phenomena. Incorporating quasi-one-dimensional (quasi-1D) structural motifs into 2D superconductors offers a powerful route to engineer strong electronic anisotropy, enabling unconventional superconducting states and anisotropic superconducting transport functionalities. However, such systems remain rarely realized. Here we report the realization of a 2D superconductor Nb2Pd3Te5, exhibiting an intrinsic quasi-1D pair density modulation. Monolayer and bilayer Nb2Pd3Te5 is synthesized via van-der-Waals epitaxy. Using ultralow-temperature scanning tunneling microscopy/spectroscopy, we observe the quasi-1D crystal structure and superconductivity below ~0.6 K with a pronounced quasi-1D pair density modulation. Remarkably, both monolayer and bilayer Nb2Pd3Te5 show strong air stability. Our findings establish atomically 2D Nb2Pd3Te5 as a robust and promising platform for exploring novel low-dimensional quantum phenomena and anisotropy-enabled superconducting devices.

cond-mat.mtrl-sci↗