Searcharxiv⌕ Search

arXiv subjects

Yu Gao

Publications and source records attributed to Yu Gao.

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

GRAB: An LLM-Inspired Sequence-First Click-Through Rate Prediction Modeling Paradigm

Traditional Deep Learning Recommendation Models (DLRMs) face increasing bottlenecks in performance and efficiency, often struggling with generalization and long-sequence modeling. Inspired by the scaling success of Large Language Models (LLMs), we propose Generative Ranking for Ads at Baidu (GRAB), an end-to-end generative framework for Click-Through Rate (CTR) prediction. GRAB integrates a novel Causal Action-aware Multi-channel Attention (CamA) mechanism to effectively capture temporal dynamics and specific action signals within user behavior sequences. Full-scale online deployment demonstrates that GRAB significantly outperforms established DLRMs, delivering a 3.05% increase in revenue and a 3.49% rise in CTR. Furthermore, the model demonstrates desirable scaling behavior: its expressive power shows a monotonic and approximately linear improvement as longer interaction sequences are utilized.

cs.IR↗
arXiv2026

An Autonomous Agent Framework for Feature-Label Extraction from Device Dialogues and Automatic Multi-Dimensional Device Hosting Planning Based on Large Language Models

With the deep integration of artificial intelligence and smart home technologies, the intelligent transformation of traditional household appliances has become an inevitable trend. This paper presents AirAgent--an LLM-driven autonomous agent framework designed for home air systems. Leveraging a voice-based dialogue interface, AirAgent autonomously and personally manages indoor air quality through comprehensive perception, reasoning, and control. The framework innovatively adopts a two-layer cooperative architecture: Memory-Based Tag Extraction and Reasoning-Driven Planning. First, a dynamic memory tag extraction module continuously updates personalized user profiles. Second, a reasoning-planning model integrates real-time environmental sensor data, user states, and domain-specific prior knowledge (e.g., public health guidelines) to generate context-aware decisions. To support both interpretability and execution, we design a semi-streaming output mechanism that uses special tokens to segment the model's output stream in real time, simultaneously producing human-readable Chain-of-Thought explanations and structured, device-executable control commands. The system handles planning across 25 distinct complex dimensions while satisfying more than 20 customized constraints. As a result, AirAgent endows home air systems with proactive perception, service, and orchestration capabilities, enabling seamless, precise, and personalized air management responsive to dynamic indoor and outdoor conditions. Experimental results demonstrate up to 94.9 percent accuracy and more than 20 percent improvement in user experience metrics compared to competing commercial solutions.

cs.HC↗
arXiv2026

Graviton energy spectra arising from the KSVZ axion model

Axion, the goldstone boson arising from the spontaneous breaking of a global $U(1)$ Peccei-Quinn symmetry, provides a dynamical solution to the strong CP problem and is an excellent dark matter candidate. Various experiments are designed to search for the axion, however no confirmative signal has been observed. On the other hand, there are also hypothetical heavy particles in axion models, such as the heavy scalar $s$, which is the CP-even component of the complex scalar that carries $U(1)_{PQ}$ charge, and the vector-like heavy quark (VLQ) in the Kim-Shifman-Vainshtein-Zakharov~(KSVZ) axion model. Studying signals induced by them are helpful for axion searches. In this paper, we calculate the graviton bremsstrahlung energy spectrum arising from the decay of the heavy scalar or VLQ in the KSVZ model. The result shows that these heavy particles can emit ultrahigh-frequency gravitational waves (GWs), with the peak frequency depending on the model's parameter inputs. In addition, the graviton spectrum is distinguished from the thermal GW background at high frequencies if there is an early matter-dominated era induced by these heavy particles. Future measurements of ultrahigh-frequency GWs may provide indirect evidence for the KSVZ axion.

hep-ph↗
arXiv2026

Retrieval-Augmented Self-Taught Reasoning Model with Adaptive Chain-of-Thought for ASR Named Entity Correction

End-to-end automatic speech recognition (ASR) systems frequently misrecognize domain-specific phrases like named entities, which can cause catastrophic failures in downstream tasks. A new family of named entity correction methods based on large language models (LLMs) has recently emerged. However, these approaches have yet to fully exploit the sophisticated reasoning capabilities inherent to LLMs. To bridge this gap, we propose a novel retrieval-augmented generation framework for correcting named entity errors in ASR. Our approach consists of two key components: (1) a rephrasing language model (RLM) for named entity recognition, followed by candidate retrieval using a phonetic-level edit distance; and (2) a novel self-taught reasoning model with adaptive chain-of-thought (A-STAR) that dynamically adjusts the depth of its reasoning based on task difficulty. Experiments on the AISHELL-1 and Homophone datasets demonstrate the effectiveness of our method, which achieves relative reductions in the named entity character error rate of 17.96\% and 34.42\%, respectively, compared to a strong baseline.

cs.CL↗
arXiv2026

The Structure of an 80 pc Long Massive Filament

Using new Institut de Radioastronomie Millimétrique (IRAM) 30m telescope $\rm N_2H^+$, $\rm C^{18}O$ $J$=1-0 and Atacama Pathfinder Experiment (APEX) telescope $\rm ^{13}CO$ and $\rm C^{18}O$ $J$=2-1 maps together with archival far-infrared continuum data, and $\rm ^{12}CO$, and $\rm ^{13}CO$ $J$=1-0 data, we present a comprehensive analysis of the massive filament CFG024.00$+$0.48 (G24) across clump-to-cloud scales. Our results show that G24 is an $\sim$80 pc giant filament with a total mass of $\sim$$10^5$ M$_{\odot}$. In the different tracers the filament width is measured to be about $\sim$2 times the beam size of the observations, as expected for power-law density distributions, giving beam-deconvolved widths in the range from 0.8 to 2.8 pc. We determine a line-of-sight thickness of $\sim$2.2 pc demonstrating that G24 is not an edge-on, flatten structure. The virial parameter obtained from line mass ($α_{\rm line,vir}=M_{\rm line,vir}/M_{\rm line}$) from the $\rm C^{18}O$ (1-0) data is 0.85, and that obtained from $Herschel$-based H$_2$ column density is 0.52, suggesting G24 is globally close to virial equilibrium. The distribution of the 40 dust clumps appears to have a ''two-tier'' fragmentation pattern. For the clump groups, the separation, with a mean/median of 3.68/3.46 pc, is very close to expected length associated with the maximum fragmentation growth rate of $λ_{\rm max}=3.55 \pm0.32$ pc estimated for the dust. However, the longitudinal centroid velocity profiles of $\rm C^{18}O$ and $\rm N_2H^+$ show oscillation patterns with wavelengths of 9.8$\pm$0.1 pc and 9.9$\pm$0.1 pc, respectively. This is $\sim$2 times larger than the corresponding values of $λ_{\rm max}$ of 4.96$\pm$0.63 pc and 4.65$\pm$1.34 pc, respectively. This suggests that the velocity structure is not dominated by flows directly associated with the fragmentation seen in the dust emission.

astro-ph.GA↗
arXiv2026

Adaptive Speaker Embedding Self-Augmentation for Personal Voice Activity Detection with Short Enrollment Speech

Personal Voice Activity Detection (PVAD) is crucial for identifying target speaker segments in the mixture, yet its performance heavily depends on the quality of speaker embeddings. A key practical limitation is the short enrollment speech--such as a wake-up word--which provides limited cues. This paper proposes a novel adaptive speaker embedding self-augmentation strategy that enhances PVAD performance by augmenting the original enrollment embeddings through additive fusion of keyframe embeddings extracted from mixed speech. Furthermore, we introduce a long-term adaptation strategy to iteratively refine embeddings during detection, mitigating speaker temporal variability. Experiments show significant gains in recall, precision, and F1-score under short enrollment conditions, matching full-length enrollment performance after five iterative updates. The source code is available at https://anonymous.4open.science/r/ASE-PVAD-E5D6 .

eess.AS↗
arXiv2026

RISE: Rule-Driven SQL Dialect Translation via Query Reduction

Translating SQL dialects across different relational database management systems (RDBMSs) is crucial for migrating RDBMS-based applications to the cloud. Traditional SQL dialect translation tools rely on manually-crafted rules, necessitating significant manual effort to support new RDBMSs and dialects. Although large language models (LLMs) can assist in translating SQL dialects, they often struggle with lengthy and complex SQL queries. In this paper, we propose RISE, a novel LLM-based SQL dialect translation approach that can accurately handle lengthy and complex SQL queries. Given a complex source query $Q_c$ that contains a SQL dialect $d$, we first employ a dialect-aware query reduction technique to derive a simplified query $Q_{s}$ by removing $d$-irrelevant SQL elements from $Q_c$. Subsequently, we utilize LLMs to translate $Q_{s}$ into $Q_{s^{'}}$, and automatically extract the translation rule $r_d$ for dialect $d$ based on the relationship between $Q_{s}$ and $Q_{s^{'}}$. By applying $r_d$ to $Q_c$, we can effectively translate the dialect $d$ within $Q_c$, thereby bypassing the complexity of the source query $Q_c$. We evaluate RISE on two real-world benchmarks, i.e., TPC-DS and SQLProcBench, comparing its performance against both the traditional rule-based tools and the LLM-based approaches with respect to translation accuracy. RISE achieves accuracies of 97.98% on TPC-DS and 100% on SQLProcBench, outperforming the baselines by an average improvement of 24.62% and 238.41%, respectively.

cs.DB↗
arXiv2025

Seedance 1.5 pro: A Native Audio-Visual Joint Generation Foundation Model

Recent strides in video generation have paved the way for unified audio-visual generation. In this work, we present Seedance 1.5 pro, a foundational model engineered specifically for native, joint audio-video generation. Leveraging a dual-branch Diffusion Transformer architecture, the model integrates a cross-modal joint module with a specialized multi-stage data pipeline, achieving exceptional audio-visual synchronization and superior generation quality. To ensure practical utility, we implement meticulous post-training optimizations, including Supervised Fine-Tuning (SFT) on high-quality datasets and Reinforcement Learning from Human Feedback (RLHF) with multi-dimensional reward models. Furthermore, we introduce an acceleration framework that boosts inference speed by over 10X. Seedance 1.5 pro distinguishes itself through precise multilingual and dialect lip-syncing, dynamic cinematic camera control, and enhanced narrative coherence, positioning it as a robust engine for professional-grade content creation. Seedance 1.5 pro is now accessible on Volcano Engine at https://console.volcengine.com/ark/region:ark+cn-beijing/experience/vision?type=GenVideo.

cs.CV↗
arXiv2025

Signatures of inhomogeneous dark matter annihilation on 21-cm

The energy released from dark matter (DM) annihilation leads to additional ionization and heating of the intergalactic gas, impacting the hydrogen 21-cm signal during the cosmic dawn. The dark matter annihilation rate scales with its density squared and becomes inhomogeneously boosted with structure formation. This paper examines the inhomogeneity in DM annihilation rate induced by the growth of DM halo structures, and we show that this effect can significantly amplify the spatial fluctuations in temperature and ionization fraction of the gas. Consequently, the fluctuations in the 21-cm brightness temperature may also be enhanced. We showcase these effects for a DM mass of 100 MeV annihilating into $\rm{e}^-\rm{e}^+$ at a rate of $\left<σv\right>/m_χ\sim 10^{-27} {\rm cm^3 s^{-1} GeV^{-1}}$, which is consistent with current constraints set by the cosmic microwave background. We find that, compared to the homogeneous calculations, inhomogeneous annihilation can enhance the 21-cm power spectrum by up to a factor of 130 over the scales of $k \in [0.05, 3]\ {\rm{Mpc^{-1}}}$ at redshifts $11-16$. Such signatures could potentially be detected by upcoming radio observatories such as the Square Kilometer Array telescope.

astro-ph.CO↗
arXiv2025

Reject or Not?: A Benchmark for Voice Assistant Query Rejection in Smart Home Scenario and an Improved Method Based on LLMs

In smart-home voice assistant scenario, deciding whether to accept or reject a user query is the first step before any downstream processing. To address the limited query-rejection capability of current voice assistants, this paper presents the first Chinese-oriented open-source benchmark and evaluation suite for smart homes, together with a personalized query-rejection method based on large language models. On the data side, we construct the first multimodal query-rejection dataset tailored for domestic scenarios, containing 11,913 manually labeled text-speech pairs that systematically cover twelve typical dialogue types (e.g., chit-chat, non-human sounds, valid commands, ambiguous references, device-irrelevant requests). Fine-grained labels, conversational context and multi-turn information are provided to support both zero-shot and fine-tuning evaluations across language and multimodal large models. On the method side, we propose a three-tier collaborative architecture: first, a Qwen-2.5-3B adapter fine-tuned to model family-agnostic semantic boundaries; second, a dynamic household-level historical dialogue module to capture personalized habits; third, a household-specific RAG knowledge base that explicitly memorizes and revises past false-rejection cases. Experiments show that the proposed approach significantly outperforms zero-shot and fine-tuned general LLMs on the constructed dataset, with pronounced gains in rejection accuracy for family-specific expressions and complex multi-turn scenarios. This work provides a reproducible data foundation, evaluation standard and extensible technical framework for reliability research in smart-home voice interaction.

cs.HC↗
arXiv2025

Seedream 4.0: Toward Next-generation Multimodal Image Generation

We introduce Seedream 4.0, an efficient and high-performance multimodal image generation system that unifies text-to-image (T2I) synthesis, image editing, and multi-image composition within a single framework. We develop a highly efficient diffusion transformer with a powerful VAE which also can reduce the number of image tokens considerably. This allows for efficient training of our model, and enables it to fast generate native high-resolution images (e.g., 1K-4K). Seedream 4.0 is pretrained on billions of text-image pairs spanning diverse taxonomies and knowledge-centric concepts. Comprehensive data collection across hundreds of vertical scenarios, coupled with optimized strategies, ensures stable and large-scale training, with strong generalization. By incorporating a carefully fine-tuned VLM model, we perform multi-modal post-training for training both T2I and image editing tasks jointly. For inference acceleration, we integrate adversarial distillation, distribution matching, and quantization, as well as speculative decoding. It achieves an inference time of up to 1.8 seconds for generating a 2K image (without a LLM/VLM as PE model). Comprehensive evaluations reveal that Seedream 4.0 can achieve state-of-the-art results on both T2I and multimodal image editing. In particular, it demonstrates exceptional multimodal capabilities in complex tasks, including precise image editing and in-context reasoning, and also allows for multi-image reference, and can generate multiple output images. This extends traditional T2I systems into an more interactive and multidimensional creative tool, pushing the boundary of generative AI for both creativity and professional applications. We further scale our model and data as Seedream 4.5. Seedream 4.0 and Seedream 4.5 are accessible on Volcano Engine https://www.volcengine.com/experience/ark?launch=seedream.

cs.CV↗
arXiv2025

The MALATANG survey: star formation, dense gas, and AGN feedback in NGC 1068

We aim to investigate the interplay between dense molecular gas, star formation, and active galactic nucleus (AGN) feedback in the luminous infrared galaxy (LIRG) NGC 1068 at sub-kiloparsec scales. We present the HCN (4-3) and HCO$^+$ (4-3) maps of NGC 1068, obtained with JCMT as part of the Mapping the dense molecular gas in the strongest star-forming galaxies (MALATANG) project, and perform spatially resolved analyses of their correlations with infrared luminosity and soft X-ray emission. Spatially resolved relations between the luminosities of infrared dust emission and dense molecular gas tracers ($L_{\rm IR}-L'_{\rm dense}$) are found to be nearly linear, without clear evidence of excess contributions from AGN activity. The spatially resolved X-ray emission ($L^{\rm gas}_{0.5-2\,\mathrm{keV}}$) displays a radially-dependent twofold correlation with the star formation rate (SFR), suggesting distinct gas-heating mechanisms between the galaxy center and the outer regions. A super-linear scaling is obtained in galactic center regions with SFR surface density ($Σ_{\rm SFR}$) $>$ 8.2 $\times$ 10$^{-6}$ $M_\odot$ yr$^{-1}$ kpc$^{-2}$: log($L^{\rm gas}_{0.5-2\,\mathrm{keV}}$/erg s$^{-1}$) = 2.2 log(SFR/$M_\odot$ yr$^{-1}$) + 39.1. We further found a statistically significant super-linear correlation ($β= 1.34$ $\pm$ 0.86) between $L^{\rm gas}_{0.5-2\,\mathrm{keV}}$/SFR and HCN(4-3)/CO(1-0) intensity ratio, whereas no such trend is seen for HCO$^+$(4-3)/CO(1-0) or CO(3-2)/CO(1-0). These findings indicate that AGN feedback does not dominate star formation regulation on sub-kiloparsec scales, and that the excitation of dense gas traced by HCN (4-3) may be more directly influenced by high-energy feedback processes compared to HCO$^+$ (4-3) and CO (3-2).

astro-ph.GA↗
arXiv2025

Probing Sub-eV Dark Photon, Scalar and Axion-like Particle Dark Matters with Transmon Qubits

In this paper, we investigate constraints of the transmon qubit, an improved version of the charge qubit, on bosonic light dark matters. Phonon excitations induced by the scattering or absorption of dark matter on a superconductor may destroy the Cooper pair, leading to the production of quasiparticles made by the electron. By measuring the production rate of the quasiparticle density, one may read out the coupling between dark matter and ordinary matter, assuming that these quasiparticles are solely induced by dark matter interactions. For the first time, we show constraints on the parameter space of the dark photon, light scalar dark matter, and axion-like particles from the measurement of quasiparticles in transmon qubit experiments. This study offers insights for the development of quantum qubit experiments aimed at the direct detection of dark matter in underground laboratories.

hep-ph↗
arXiv2025

DiffVLA++: Bridging Cognitive Reasoning and End-to-End Driving through Metric-Guided Alignment

Conventional end-to-end (E2E) driving models are effective at generating physically plausible trajectories, but often fail to generalize to long-tail scenarios due to the lack of essential world knowledge to understand and reason about surrounding environments. In contrast, Vision-Language-Action (VLA) models leverage world knowledge to handle challenging cases, but their limited 3D reasoning capability can lead to physically infeasible actions. In this work we introduce DiffVLA++, an enhanced autonomous driving framework that explicitly bridges cognitive reasoning and E2E planning through metric-guided alignment. First, we build a VLA module directly generating semantically grounded driving trajectories. Second, we design an E2E module with a dense trajectory vocabulary that ensures physical feasibility. Third, and most critically, we introduce a metric-guided trajectory scorer that guides and aligns the outputs of the VLA and E2E modules, thereby integrating their complementary strengths. The experiment on the ICCV 2025 Autonomous Grand Challenge leaderboard shows that DiffVLA++ achieves EPDMS of 49.12.

cs.RO↗
arXiv2025

The spin-switch scanning tunneling microscopy: an architecture to probe electron-phonon interactions in the atomic scale

On the spin-valve-like ferromagnet/spin glass/ferromagnet (FM/SG/FM) structure, the tunneling current is dominated by resistance switch (RS) instead of the local density of states according to the conventional tunneling theory. Here we show lattice-site dependent RS behaviors in one-quintuple-layer Bi2Te3 deposited on single MnBi2Te4 septuple layer, which comes from the difference in the efficiency of tunneling electrons to induce focused current or phonons at different sites, switching remotely the spin valve by spin-transfer torque or spin-phonon interactions. These lead to the observation of the dynamic 2-state lattice when the tip scans across the surface as well as the ability of scanning tunneling microscope (STM) to reveal atomic-scale features of electron-phonon (EP) interactions. Our work demonstrates the possibility of the spin-switch STM to image lattice-site dependent EP interactions of any materials deposited on the FM/SG/FM structure.

cond-mat.mtrl-sci↗
arXiv2025

Prox-PINNs: A Deep Learning Algorithmic Framework for Elliptic Variational Inequalities

Elliptic variational inequalities (EVIs) present significant challenges in numerical computation due to their inherent non-smoothness, nonlinearity, and inequality formulations. Traditional mesh-based methods often struggle with complex geometries and high computational costs, while existing deep learning approaches lack generality for diverse EVIs. To alleviate these issues, this paper introduces Prox-PINNs, a novel deep learning algorithmic framework that integrates proximal operators with physics-informed neural networks (PINNs) to solve a broad class of EVIs. The Prox-PINNs reformulate EVIs as nonlinear equations using proximal operators and then approximate the solutions via neural networks that enforce boundary conditions as hard constraints. Then the neural networks are trained by minimizing physics-informed residuals. The Prox-PINNs framework advances the state-of-the-art by unifying the treatment of diverse EVIs within a mesh-free and scalable computational architecture. The framework is demonstrated on several prototypical applications, including obstacle problems, elasto-plastic torsion, Bingham visco-plastic flows, and simplified friction problems. Numerical experiments validate the method's accuracy, efficiency, robustness, and flexibility across benchmark examples.

math.OC↗
arXiv2025

Emerging axion detection in artificial magnetoelectric materials

Axions are considered a key component of dark matter, characterized by very weak couplings to fermions and Chern-Simons couplings to gauge fields. We propose a novel detection mechanism based on symmetry-breaking magnetoelectric materials with a linear axionic coupling between magnetization and ferroelectric polarization. The focus is on a strain gradient Sr2IrO4 film, where the breaking of space-inversion symmetry results in an emergent polar phase and an out-of-plane magnetic moment, exhibiting a flexomagnetoelectric effect. In this material, the linear P||M enables a direct coupling between the external axion field and the intrinsic axion-like field, which amplifies the weak electromagnetic signals induced by axions, paving the way for pioneering axion detection. In contrast to conventional detection techniques, this mechanism is expected to enhance the sensitivity of the axion-electron and axion-photon coupling, providing a novel platform for axion detection and advancing the study of dark matter through the magnetoelectric effect.

hep-ph↗
arXiv2025

Detecting gravitational waves with spin systems

The observation of gravitational waves has opened a new window into the Universe through gravitational-wave astronomy. However, high-frequency gravitational waves remain undetected. In this work, we propose that spin systems can be employed to detect gravitational waves in this unexplored frequency regime. We derive the spin's response to gravitational waves and identify three distinct effects: the well-known Gertsenshtein effect, a metric-induced interaction, and the gravitational spin Hall effect. We focus on nuclear spins and utilize nuclear magnetic resonance to enhance the gravitational response, leveraging the advantages of long coherence time, high polarization, and a small gyromagnetic ratio. The proposed experimental scheme is capable of probing gravitational waves in the kilohertz to gigahertz range, with projected sensitivities reaching $\sqrt{S_h}\approx10^{-20}~\mathrm{Hz}^{-1/2}$.

gr-qc↗