arXiv subjects

Hongyang Wang

Publications and source records attributed to Hongyang Wang.

At least 19 recordsLinked to original sources

arXiv2026

A/B Agent: A Self-Evolving Agent for Strategy Iteration in Industrial A/B Testing

Industrial recommendation strategy iteration heavily relies on large-scale A/B experimentation. Traditional tuning requires experts to repeatedly design strategies, configure experiments, analyze results, and adjust parameters, making the process labor-intensive and time-consuming. Meanwhile, valuable knowledge from historical experiments is often fragmented, making systematic reuse difficult through manual expert effort alone. Existing RAG agents partially alleviate this burden by retrieving prior strategies, but typically organize experience in a flat manner, overlooking the hierarchical relationships among business scenarios, recommendation stages, optimization objectives, and experimental contexts. This often results in mismatched retrieval and limited cross-scenario transfer, while preventing agents from continuously refining strategies and parameters through sequential A/B feedback. % To address these limitations, we propose A/B Agent, a closed-loop A/B agent for industrial recommendation strategy optimization. The framework comprises three tightly coupled core components: Historical Strategy Knowledge Organization, Autonomous Target-Aware Strategy Generation, and Experiment-Guided Strategy Self-Evolution. It organizes historical strategies into a hierarchical experience tree, retrieves transferable evidence through multi-path Tree-RAG to generate executable strategies, and continuously analyzes online A/B feedback to guide autonomous tuning and update the experience tree for self-evolution. Extensive offline and online evaluations demonstrate its effectiveness, including a 4.829% improvement in GMV in a real-world short-video e-commerce recommendation system while maintaining positive gains across all guardrail metrics.

cs.AI↗

arXiv2026

FAS-R1: A Unified Multi-Task MLLM for Reasoning Face Anti-Spoofing

Face anti-spoofing (FAS) is increasingly expected to provide not only bona fide/spoof decisions, but also attack semantics and image-grounded evidence for human inspection. Existing discriminative FAS models remain largely label-centric, while recent MLLM-based methods offer structured outputs but still rely mainly on supervised fine-tuning, often producing template-like rationales and weak optimization for difficult attacks. We propose FAS-R1, a two-stage reasoning-oriented MLLM framework for unified FAS prediction, covering authenticity classification, attack-type recognition and spoof-region localization. FAS-R1 first uses FAS-R1-23K, a high-quality long-CoT dataset, for cold-start supervised fine-tuning, and then performs FAS-specific GRPO post-training. Degradation-Simulated Augmentation (DSA) encourages stable spoof-cue reasoning across visual-quality shifts, while Difficulty-Aware GRPO (DA-GRPO) mitigates easy-sample dominance that may leave difficult task--attack groups under-optimized, especially for subtle or ambiguous attacks such as makeup and mask attacks. The main 3B FAS-R1 model achieves 98.75\% authenticity accuracy, 93.33\% attack-type accuracy, and 96.30/94.73\% AP@40/AP@50 in-domain. It also outperforms the compared systems in cross-domain authenticity generalization and answer-and-rationale quality. Experiments with different base models further show favorable scaling behavior. The code will be released soon.

cs.CV↗

arXiv2026

WuYuEval: A Multi-Level Benchmark for Large Language Models in Solid Waste Management

Large language models (LLMs) are increasingly used as technical assistants, but their competence in solid waste management (SWM) remains difficult to assess because existing benchmarks emphasize general knowledge rather than professional decisions under engineering, environmental, and policy constraints. We introduce WuYuEval, a multi-level benchmark for evaluating LLMs in SWM across foundational knowledge, domain reasoning, and expert decision-making. After quality auditing, WuYuEval contains a Foundation Module with 4,590 closed-ended multiple-choice questions across six task types and eight domain categories, together with an Expert Module with 247 scenario-based open-ended questions involving multi-objective optimization, constraint trade-offs, and system design. For expert tasks, we combine anchor-calibrated LLM-as-a-Judge scoring with Elo-based pairwise comparison. Across 33 LLMs, performance varied widely. The leading model reached 94.64\% accuracy on the Foundation Module, but average accuracy still fell from 84.14\% on easy questions to 42.50\% on hard questions, with lower performance concentrated in calculation, experimental design, urban planning, and open-ended expert tasks. Reasoning-oriented Thinking modes improve most matched model pairs after auditing, but the gains depend on baseline capability and are not uniformly positive. These results suggest that visible deliberation helps only when it remains anchored to units, assumptions, and engineering constraints; otherwise, it may drift from decisive answer boundaries. WuYuEval therefore provides both an evaluation resource and an empirical basis for developing SWM-oriented foundation models with professional reasoning chains and explicit constraint control.

cs.CL↗

arXiv2026

Partition-selected flow polynomials and associated arrangements

We introduce a partition-selection method to generalize the flow, chromatic, and Tutte polynomials of a graph by restricting the standard edge subgraph expansions to subgraphs given by prescribed connected vertex partitions. We establish similar deletion-contraction formulas and specialization relations for these polynomials, recovering all classical polynomial invariants when the selection is the set of all partitions. Next we study a relation between Jaeger et al.'s nonhomogeneous flows and a special class of partition-selected flow polynomials (called affine flow polynomials). Specifically, we give a geometric realization of nowhere-zero nonhomogeneous flows by restricting the edge-coordinate arrangement to affine flow spaces. The resulting characteristic polynomials coincide with Kochol's admissible assigning polynomials and with affine flow polynomials, which enumerate nowhere-zero nonhomogeneous flows over finite fields. To see the key role of the partition-selection framework, we further introduce boundary arrangements determined by the bond structure of a graph. Using the intersection posets of boundary arrangements, we obtain the classification of all restricted arrangements mentioned above, the comparison of unsigned coefficients of affine flow polynomials, and the decomposition formulas for affine flow polynomials.

math.CO↗

arXiv2026

UniShield: Unified Face Attack Detection via KG-Informed Multimodal Reasoning

Unified face attack detection (UAD) requires recognizing physical spoofing and digital forgery within a shared decision space, yet existing discriminative or prompt-based methods largely rely on appearance correlations and provide limited evidence-grounded reasoning. We propose UniShield, a knowledge-grounded multimodal reasoning framework for unified face attack defense. UniShield constructs a Face Attack Knowledge Graph (FAKG) that links attack categories to diagnostic visual cues and attack-conditioned relations, and uses it to synthesize 52,025 FAKG-QA examples for Attack-Graph Instruction Tuning (AGIT). To improve rationale consistency, we further introduce Graph-Consistent Reasoning Optimization (GCRO), a GRPO-based objective with a KG-consistency reward that encourages generated rationales to match graph-supported cues while penalizing incompatible claims. Experiments on our multimodal UAD benchmark show that UniShield achieves strong performance across binary, coarse-grained, and fine-grained protocols, with consistently high ACC and low HTER. These results suggest that structured attack knowledge can improve both detection accuracy and reasoning reliability over discriminative baselines and general-purpose MLLMs. Our code will be released at https://anonymous.4open.science/r/Unishield-A6A3/.

cs.CV↗

arXiv2026

Beyond Static Collision Handling: Adaptive Semantic ID Learning for Multimodal Recommendation at Industrial Scale

Modern recommendation systems involve massive catalogs of multimodal items, where scalable item identification must balance compactness, semantic fidelity, and downstream effectiveness. Semantic IDs (SIDs) address this need by representing items as short discrete token sequences derived from multimodal signals, providing a compact interface for retrieval, ranking, and generative recommendation. However, effective SID learning is hindered by collisions, where different items are assigned identical or highly confusable codes. Existing methods mainly rely on improved quantization or fixed overlap regularization, but they do not adaptively distinguish whether an overlap should be suppressed or preserved. We propose AdaSID, an adaptive semantic ID learning framework for recommendation. AdaSID regulates SID overlaps through a two-stage process. First, it relaxes repulsion for observed overlaps when the involved items are semantically compatible, preserving admissible sharing rather than uniformly separating all collisions. Second, it allocates the remaining regulation pressure according to local collision load and training progress, strengthening control in congested regions while gradually rebalancing optimization toward recommendation alignment. This design adaptively decides which overlaps to penalize, how strongly to regulate them, and when to shift the learning focus. Extensive offline and online experiments validate AdaSID. On two public benchmarks, AdaSID improves Recall and NDCG by about 4.5% on average over strong baselines, while improving codebook utilization and SID diversity. In Kuaishou e-commerce, an online A/B test on short-video retrieval covering tens of millions of users achieves statistically significant gains, including a 0.98% GMV improvement, and industrial ranking evaluation shows consistent AUC improvements.

cs.IR↗

arXiv2026

Stop Treating Collisions Equally: Qualification-Aware Semantic ID Learning for Recommendation at Industrial Scale

Semantic IDs (SIDs) are compact discrete representations derived from multimodal item features, serving as a unified abstraction for ID-based and generative recommendation. However, learning high-quality SIDs remains challenging due to two issues. (1) Collision problem: the quantized token space is prone to collisions, in which semantically distinct items are assigned identical or overly similar SID compositions, resulting in semantic entanglement. (2) Collision-signal heterogeneity: collisions are not uniformly harmful. Some reflect genuine conflicts between semantically unrelated items, while others stem from benign redundancy or systematic data effects. To address these challenges, we propose Qualification-Aware Semantic ID Learning (QuaSID), an end-to-end framework that learns collision-qualified SIDs by selectively repelling qualified conflict pairs and scaling the repulsion strength by collision severity. QuaSID consists of two mechanisms: Hamming-guided Margin Repulsion, which translates low-Hamming SID overlaps into explicit, severity-scaled geometric constraints on the encoder space; and Conflict-Aware Valid Pair Masking, which masks protocol-induced benign overlaps to denoise repulsion supervision. In addition, QuaSID incorporates a dual-tower contrastive objective to inject collaborative signals into tokenization. Experiments on public benchmarks and industrial data validate QuaSID. On public datasets, QuaSID consistently outperforms strong baselines, improving top-K ranking quality by 5.9% over the best baseline while increasing SID composition diversity. In an online A/B test on Kuaishou e-commerce with a 5% traffic split, QuaSID increases ranking GMV-S2 by 2.38% and improves completed orders on cold-start retrieval by up to 6.42%. Finally, we show that the proposed repulsion loss is plug-and-play and enhances a range of SID learning frameworks across datasets.

cs.IR↗

arXiv2025

Optimal chromatic bound for ($P_2\cup P_4$, HVN)-free graphs

The HVN is a graph formed by removing two edges incident to the same vertex from the complete graph $K_5$. In this paper, we prove that every ($P_2\cup P_4$, HVN)-free graph $G$ satisfies $\chi(G)\leq\lceil\frac{4}{3}\omega(G)\rceil$ when $\omega(G)\ge4$, where $\chi(G)$ and $\omega(G)$ denote the chromatic number and clique number of $G$, respectively. Furthermore, this bound is optimal for every $\omega(G)\ge4$. Constructions demonstrating the optimality of the bound are provided. Our work unifies several previously known results on $\chi$-binding functions for several graph classes.

math.CO↗

arXiv2025

Colouring ($P_2\cup P_4$, diamond)-free graphs with $\omega$ colours

In this paper, we establish an optimal $\chi$-binding function for $(P_2\cup P_4,\text{ diamond})$-free graphs. We prove that for any graph $G$ in this class, $\chi(G)\le 4$ when $\omega(G)=2$, $\chi(G)\le 6$ when $\omega(G)=3$, and $\chi(G)=\omega(G)$ when $\omega(G)\ge 4$, where $\chi(G)$ and $\omega(G)$ denote the chromatic number and clique number of $G$, respectively. This result extends the known chromatic bounds for $(P_2\cup P_3,\text{ diamond})$-free graphs by showing that $(P_2\cup P_4,\text{ diamond})$-free graphs admit the same $\chi$-binding function. It also refines the chromatic bound obtained by Angeliya, Karthick and Huang [arXiv:2501.02543v3 [math.CO], 2025] for $(P_2\cup P_4,\text{ diamond})$-free graphs.

math.CO↗

arXiv2025

Perfect divisions in ($P_2 \cup P_4$, bull)-free graphs

A graph $G$ has a perfect division if its vertex set can be partitioned into two sets $A$, $B$ such that $G[A]$ is perfect and $\omega(G[B]) < \omega(G)$. We call $G$ perfectly divisible if every induced subgraph of $G$ admits a perfect division. We prove that every ($P_2 \cup P_4$, bull)-free graph $G$ with $\omega(G) \geq 3$ has a perfect division if $G$ contains no homogeneous set. The clique-number condition is tight: a counterexample exists for $\omega(G) = 2$. Additionally, we present a short proof of the perfect divisibility of ($P_5$, bull)-free graphs, originally established by Chudnovsky and Sivaraman [J. Graph Theory 90 (2019), 54-60.].

math.CO↗

arXiv2025

AMSbench: A Comprehensive Benchmark for Evaluating MLLM Capabilities in AMS Circuits

Analog/Mixed-Signal (AMS) circuits play a critical role in the integrated circuit (IC) industry. However, automating Analog/Mixed-Signal (AMS) circuit design has remained a longstanding challenge due to its difficulty and complexity. Although recent advances in Multi-modal Large Language Models (MLLMs) offer promising potential for supporting AMS circuit analysis and design, current research typically evaluates MLLMs on isolated tasks within the domain, lacking a comprehensive benchmark that systematically assesses model capabilities across diverse AMS-related challenges. To address this gap, we introduce AMSbench, a benchmark suite designed to evaluate MLLM performance across critical tasks including circuit schematic perception, circuit analysis, and circuit design. AMSbench comprises approximately 8000 test questions spanning multiple difficulty levels and assesses eight prominent models, encompassing both open-source and proprietary solutions such as Qwen 2.5-VL and Gemini 2.5 Pro. Our evaluation highlights significant limitations in current MLLMs, particularly in complex multi-modal reasoning and sophisticated circuit design tasks. These results underscore the necessity of advancing MLLMs' understanding and effective application of circuit-specific knowledge, thereby narrowing the existing performance gap relative to human expertise and moving toward fully automated AMS circuit design workflows. Our data is released at this URL.

cs.LG↗

arXiv2025

AMSnet 2.0: A Large AMS Database with AI Segmentation for Net Detection

Current multimodal large language models (MLLMs) struggle to understand circuit schematics due to their limited recognition capabilities. This could be attributed to the lack of high-quality schematic-netlist training data. Existing work such as AMSnet applies schematic parsing to generate netlists. However, these methods rely on hard-coded heuristics and are difficult to apply to complex or noisy schematics in this paper. We therefore propose a novel net detection mechanism based on segmentation with high robustness. The proposed method also recovers positional information, allowing digital reconstruction of schematics. We then expand AMSnet dataset with schematic images from various sources and create AMSnet 2.0. AMSnet 2.0 contains 2,686 circuits with schematic images, Spectre-formatted netlists, OpenAccess digital schematics, and positional information for circuit components and nets, whereas AMSnet only includes 792 circuits with SPICE netlists but no digital schematics.

cs.CV↗

arXiv2025

FaceShield: Explainable Face Anti-Spoofing with Multimodal Large Language Models

Face anti-spoofing (FAS) is crucial for protecting facial recognition systems from presentation attacks. Previous methods approached this task as a classification problem, lacking interpretability and reasoning behind the predicted results. Recently, multimodal large language models (MLLMs) have shown strong capabilities in perception, reasoning, and decision-making in visual tasks. However, there is currently no universal and comprehensive MLLM and dataset specifically designed for FAS task. To address this gap, we propose FaceShield, a MLLM for FAS, along with the corresponding pre-training and supervised fine-tuning (SFT) datasets, FaceShield-pre10K and FaceShield-sft45K. FaceShield is capable of determining the authenticity of faces, identifying types of spoofing attacks, providing reasoning for its judgments, and detecting attack areas. Specifically, we employ spoof-aware vision perception (SAVP) that incorporates both the original image and auxiliary information based on prior knowledge. We then use an prompt-guided vision token masking (PVTM) strategy to random mask vision tokens, thereby improving the model's generalization ability. We conducted extensive experiments on three benchmark datasets, demonstrating that FaceShield significantly outperforms previous deep learning models and general MLLMs on four FAS tasks, i.e., coarse-grained classification, fine-grained classification, reasoning, and attack localization. Our instruction datasets, protocols, and codes will be released at https://github.com/Why0912/FaceShield.

cs.CV↗

arXiv2024

Decentralised Governance for Autonomous Cyber-Physical Systems

This paper examines the potential for Cyber-Physical Systems (CPS) to be governed in a decentralised manner, whereby blockchain-based infrastructure facilitates the communication between digital and physical domains through self-governing and self-organising principles. Decentralised governance paradigms that integrate computation in physical domains (such as 'Decentralised Autonomous Organisations' (DAOs)) represent a novel approach to autono-mous governance and operations. These have been described as akin to cybernetic systems. Through the lens of a case study of an autonomous cabin called "no1s1" which demonstrates self-ownership via blockchain-based control and feedback loops, this research explores the potential for blockchain infrastructure to be utilised in the management of physical systems. By highlighting the considerations and challenges of decentralised governance in managing autonomous physical spaces, the study reveals that autonomy in the governance of autonomous CPS is not merely a technological feat but also involves a complex mesh of functional and social dynamics. These findings underscore the importance of developing continuous feedback loops and adaptive governance frameworks within decentralised CPS to address both expected and emergent challenges. This investigation contributes to the fields of infra-structure studies and Cyber-Physical Systems engineering. It also contributes to the discourse on decentralised governance and autonomous management of physical spaces by offering both practical insights and providing a framework for future research.

cs.CY↗

arXiv2024

Indications of superconductivities in blend of variant apatite and covellite

Through heavily doping sulfur into an apatite framework, we synthesize a new blend mainly comprising variant apatite and covellite (copper sulfide). Magnetic measurement exhibits that significant diamagnetism appears at around 260 K and drops dramatically below 30 K implying coexistence of two superconducting phases. The upper critical magnetic field is larger than 1000 Oe at 250 K. Electric measurement manifests that the current-voltage curves deviate from the normal linear lineshape suggesting the presence of zero-resistance effect, and the critical current is around 50 $\mu$A at 140 K. These exotic magnetic and electric features strongly indicate these two components, variant apatite and covellite, individually trigger two superconducting phases at near-room and low temperatures.

cond-mat.supr-con↗

arXiv2024

Observation of diamagnetic strange-metal phase in sulfur-copper codoped lead apatite

By codoping sulfur and copper into lead apatite, the crystal grains are directionally stacked and the room-temperature resistivity is reduced from insulating to $2\times10^{-5}~\Omega\cdot$m. The resistance-temperature curve exhibits a nearly linear relationship at low temperature suggesting the presence of strange-metal phase, and a second-order phase transition is then observed at around 230~K during cooling the samples. A possible Meissner effect is present in dc magnetic measurements. Further hydrothermal lead-free synthesis results in smaller resistance and stronger diamagnetism, demonstrating the essential component might be sulfur-substituted copper apatite and the alkalis matter as well. A clear pathway towards superconductivity in this material is subsequently benchmarked.

cond-mat.supr-con↗

arXiv2024

Mode-resolved thermometry of trapped ion with Deep Learning

In trapped ion system, accurate thermometry of ion is crucial for evaluating the system state and precisely performing quantum operations. However, when the motional state of a single ion is far away from the ground state, the spatial dimension of the phonon state sharply increases, making it difficult to realize accurate and mode-resolved thermometry with existing methods. In this work, we apply deep learning for the first time to the thermometry of trapped ion, providing an efficient and mode-resolved method for accurately estimating large mean phonon numbers. Our trained neural network model can be directly applied to other experimental setups without retraining or post-processing, as long as the related parameters are covered by the model's effective range, and it can also be conveniently extended to other parameter ranges. We have conducted experimental verification based on our surface trap, of which the result has shown the accuracy and efficiency of the method for thermometry of single ion under large mean phonon number, and its mode resolution characteristic can make it better applied to the characterization of system parameters, such as evaluating cooling effectiveness, analyzing surface trap noise.

quant-ph↗

arXiv2024

SHIELD : An Evaluation Benchmark for Face Spoofing and Forgery Detection with Multimodal Large Language Models

Multimodal large language models (MLLMs) have demonstrated strong capabilities in vision-related tasks, capitalizing on their visual semantic comprehension and reasoning capabilities. However, their ability to detect subtle visual spoofing and forgery clues in face attack detection tasks remains underexplored. In this paper, we introduce a benchmark, SHIELD, to evaluate MLLMs for face spoofing and forgery detection. Specifically, we design true/false and multiple-choice questions to assess MLLM performance on multimodal face data across two tasks. For the face anti-spoofing task, we evaluate three modalities (i.e., RGB, infrared, and depth) under six attack types. For the face forgery detection task, we evaluate GAN-based and diffusion-based data, incorporating visual and acoustic modalities. We conduct zero-shot and few-shot evaluations in standard and chain of thought (COT) settings. Additionally, we propose a novel multi-attribute chain of thought (MA-COT) paradigm for describing and judging various task-specific and task-irrelevant attributes of face images. The findings of this study demonstrate that MLLMs exhibit strong potential for addressing the challenges associated with the security of facial recognition technology applications.

cs.CV↗