arXiv subjects

Yujie Guo

Publications and source records attributed to Yujie Guo.

13 recordsLinked to original sources

arXiv2026

Post-Training Zero-Shot TTS for Fine-Grained Emotion and Duration Control via Natural Language

Audiobook narration, conversational agents, and audiovisual dubbing require speech that conveys changing emotions and adapts its pacing within a single utterance. But most existing TTS systems typically rely on utterance-level style conditioning, making such fine-grained control difficult to achieve. In light of this, and inspired by the success of post-training in large language models, we propose a unified post-training framework that equips pretrained text-to-speech models with natural-language control over segment-level emotion and duration. Supervised fine-tuning establishes instruction-conditioned speech generation, while reinforcement learning with group relative policy optimization refines control accuracy using emotion and duration rewards alongside content and speaker preservation objectives. By reusing the pretrained architecture, our approach avoids additional inference-time control modules. Experiments demonstrate significantly improved fine-grained controllability while maintaining speech intelligibility and speaker identity, highlighting post-training as a practical approach to extending existing speech synthesis models.

cs.SD↗

arXiv2026

Self-Improving Large Language Models via Progressive Experience Evolution

Large language models (LLMs) capable of self-improvement require not only effective policy optimization, but also a principled mechanism for transforming transient interaction experience into persistent model capabilities. Existing self-improvement paradigms remain fragmented: test-time methods can explicitly extract experience but cannot internalize it into model parameters, whereas training-time optimization methods can update model parameters but lack an explicit mechanism for accumulating transferable experience. Bridging these two paradigms requires a critical intermediate stage that remains underexplored, namely \emph{experience distillation}. To address this gap, we propose \textbf{SPEE} (\textbf{S}elf-\textbf{P}rogressive \textbf{E}xperience \textbf{E}volution), a unified post-training framework that sequentially performs explicit experience evolution followed by implicit policy optimization. During explicit experience evolution, SPEE reflects on trajectories collected from multiple interactions to extract, verify, and progressively evolve transferable experience, which is subsequently internalized into the policy through privilege-guided On-Policy Self-Distillation (OPSD). During implicit policy optimization, reward-driven reinforcement learning leverages these internalized priors to explore novel solution strategies. In the experience evolution stage, a continuously evolving global experience pool consolidates knowledge from both successful and failed trajectories, filters out low-utility experience, and mitigates post-hoc rationalization induced by individual trajectories. Experiments on five mathematical reasoning benchmarks demonstrate that SPEE consistently outperforms both test-time and training-time self-evolution baselines across three model scales. The source code is available at https://github.com/rrrsj/SPEE.

cs.CL↗

arXiv2026

H-SAGE: Holistic Speaker-Aware Guided Experts for MoE-based Multi-Talker ASR

Multi-talker Automatic Speech Recognition (MTASR) faces significant challenges in accurately transcribing overlapping speech, particularly under complex high-overlap conditions. While recent Mixture-of-Experts (MoE) approaches have shown promise, they typically rely on frame-independent routing that leads to temporal myopia, and depend solely on the downstream ASR objective, which results in implicit and ungrounded representation learning. To address these limitations, we propose Holistic Speaker-Aware Guided Experts (H-SAGE) for MoE-based MTASR. Specifically, we introduce a Speaker-Aware Global Encoder to capture long-term dependencies, supervised by an auxiliary Overlap-Aware Loss that explicitly guides the model to discern acoustic states. Furthermore, we design a Holistic Gating Mechanism to arbitrate expert selection by jointly evaluating global context and local details. Experiments on LibriSpeechMix demonstrate that H-SAGE achieves consistent improvements over strong baselines, particularly in complex scenarios, validating that explicit acoustic guidance effectively enhances expert collaboration. Our code can be found at https://github.com/NKU-HLT/H-SAGE.

cs.SD↗

arXiv2026

PointSplat: Compact Gaussian Splatting via Human-Centric Prediction

Producing 3D human representations from input views on the fly is essential for immersive live streaming systems, where representation compactness is as critical as high fidelity given limited computational power and transmission bandwidth. Although recent feed-forward reconstruction methods achieve impressive quality through the view-centric prediction of 3D representations, they repeatedly encode the same subject content across multiple views, leading to significant inter-view redundancy. Our key insight is to perform predictions directly in 3D space, enabling the network to learn and produce a highly compact representation. To this end, we propose PointSplat, a novel human-centric approach that directly infers Gaussian primitives from an input point set. The proposed method first estimates a coarse geometric proxy and performs ray casting to prune redundant points and establish explicit 2D--3D correspondences. Subsequently, it employs a Point-Image Transformer to fuse appearance and geometry features, predicting Gaussian attributes in a single forward pass. This design restricts predictions to foreground regions of interest, substantially reducing the total number of Gaussians while improving novel-view rendering quality. Extensive experiments demonstrate that PointSplat achieves higher efficiency and quality while exhibiting strong robustness to variations in view count and image resolution across multiple datasets.

cs.CV↗

arXiv2025

GLAD: Global-Local Aware Dynamic Mixture-of-Experts for Multi-Talker ASR

End-to-end multi-talker automatic speech recognition (MTASR) faces significant challenges in accurately transcribing overlapping speech. A critical bottleneck is that speaker-specific acoustic characteristics, which are essential for distinguishing overlapping speech, are often diluted in deep network layers. To address this, we propose the Global-Local Aware Dynamic Mixture-of-Experts (GLAD) architecture. GLAD introduces a novel routing mechanism that dynamically fuses speaker-aware global context with fine-grained local acoustic details to adaptively guide expert selection. Experiments on the LibriSpeechMix and CH109 datasets demonstrate that GLAD significantly outperforms existing Serialized Output Training (SOT)-based MTASR approaches, exhibiting exceptional robustness in challenging, high-overlap scenarios. To the best of our knowledge, this is the first work to apply a global-local fusion MoE strategy to MTASR.

cs.SD↗

arXiv2025

Towards Automatic Evaluation and High-Quality Pseudo-Parallel Dataset Construction for Audio Editing: A Human-in-the-Loop Method

Audio editing aims to manipulate audio content based on textual descriptions, supporting tasks such as adding, removing, or replacing audio events. Despite recent progress, the lack of high-quality benchmark datasets and comprehensive evaluation metrics remains a major challenge for both assessing audio editing quality and improving the task itself. In this work, we propose a novel approach for audio editing task by incorporating expert knowledge into both the evaluation and dataset construction processes: 1) First, we establish AuditScore, the first comprehensive dataset for subjective evaluation of audio editing, consisting of over 6,300 edited samples generated from 7 representative audio editing frameworks and 23 system configurations. Each sample is annotated by professional raters on three key aspects of audio editing quality: overall Quality, Relevance to editing intent, and Faithfulness to original features. 2) Based on this dataset, we systematically propose AuditEval, a family of automatic MOS-style evaluators tailored for audio editing, covering both SSL-based and LLM-based approaches. It addresses the lack of effective objective metrics and the prohibitive cost of subjective evaluation in this field. 3) We further leverage AuditEval to evaluate and filter a large amount of synthetically mixed editing pairs, mining a high-quality pseudo-parallel subset by selecting the most plausible samples. Comprehensive experiments validate that our expert-informed filtering strategy effectively yields higher-quality data, while also exposing the limitations of traditional objective metrics and the advantages of AuditEval. The dataset, codes and tools can be found at: https://github.com/NKU-HLT/AuditEval.

cs.SD↗

arXiv2025

From Contrast to Commonality: Audio Commonality Captioning for Enhanced Audio-Text Cross-modal Understanding in Multimodal LLMs

Audio Captioning (AC) plays a pivotal role in enhancing audio-text cross-modal understanding during the pretraining and finetuning of Multimodal LLMs (MLLMs). To strengthen this alignment, recent works propose Audio Difference Captioning (ADC), which takes multiple audio inputs and encourages the model to describe their differences, thereby promoting fine-grained discrimination. However, despite its effectiveness, ADC introduces a semantic gap between input audios-often rich in diverse events-and the brief, difference-focused short caption. This deviation from AC-style task causes a mismatch with the pretraining objective, leading to catastrophic forgetting. To address this, we propose Audio Commonality Captioning (ACC), a comparably challenging but gentler alternative that guides the model to capture shared semantics across audio clips rather than detailed differences. Experiments show that ACC not only improves audio-text understanding on captioning benchmarks but also better preserves general capabilities across diverse speech and music tasks, confirming its ability to enable more robust cross-modal understanding and achieve a better balance between generalization and task-specific performance in MLLMs.

cs.SD↗

arXiv2025

DIFFA: Large Language Diffusion Models Can Listen and Understand

Recent advances in large language models (LLMs) have shown remarkable capabilities across textual and multimodal domains. In parallel, diffusion-based language models have emerged as a promising alternative to the autoregressive paradigm, offering improved controllability, bidirectional context modeling, and robust generation. However, their application to the audio modality remains underexplored. In this work, we introduce \textbf{DIFFA}, the first diffusion-based large audio-language model designed to perform spoken language understanding. DIFFA integrates a frozen diffusion language model with a lightweight dual-adapter architecture that bridges speech understanding and natural language reasoning. We employ a two-stage training pipeline: first, aligning semantic representations via an ASR objective; then, learning instruction-following abilities through synthetic audio-caption pairs automatically generated by prompting LLMs. Despite being trained on only 960 hours of ASR and 127 hours of synthetic instruction data, DIFFA demonstrates competitive performance on major benchmarks, including MMSU, MMAU, and VoiceBench, outperforming several autoregressive open-source baselines. Our results reveal the potential of diffusion-based language models for efficient and scalable audio understanding, opening a new direction for speech-driven AI. Our code will be available at https://github.com/NKU-HLT/DIFFA.git.

cs.SD↗

arXiv2025

Photonics in Flatland: Challenges and Opportunities for Nanophotonics with 2D Semiconductors

Two-dimensional (2D) semiconductors are emerging as a versatile platform for nanophotonics, offering unprecedented tunability in optical properties through exciton resonance engineering, van der Waals heterostructuring, and external field control. These materials enable active optical modulation, single-photon emission, quantum photonics, and valleytronic functionalities, paving the way for next-generation optoelectronic and quantum photonic devices. However, key challenges remain in achieving large-area integration, maintaining excitonic coherence, and optimizing amplitude-phase modulation for efficient light manipulation. Advances in fabrication, strain engineering, and computational modelling will be crucial to overcoming these limitations. This perspective highlights recent progress in 2D semiconductor-based nanophotonics, emphasizing opportunities for scalable integration into photonics.

physics.optics↗

arXiv2025

ChildMandarin: A Comprehensive Mandarin Speech Dataset for Young Children Aged 3-5

Automatic speech recognition (ASR) systems have advanced significantly with models like Whisper, Conformer, and self-supervised frameworks such as Wav2vec 2.0 and HuBERT. However, developing robust ASR models for young children's speech remains challenging due to differences in pronunciation, tone, and pace compared to adult speech. In this paper, we introduce a new Mandarin speech dataset focused on children aged 3 to 5, addressing the scarcity of resources in this area. The dataset comprises 41.25 hours of speech with carefully crafted manual transcriptions, collected from 397 speakers across various provinces in China, with balanced gender representation. We provide a comprehensive analysis of speaker demographics, speech duration distribution and geographic coverage. Additionally, we evaluate ASR performance on models trained from scratch, such as Conformer, as well as fine-tuned pre-trained models like HuBERT and Whisper, where fine-tuning demonstrates significant performance improvements. Furthermore, we assess speaker verification (SV) on our dataset, showing that, despite the challenges posed by the unique vocal characteristics of young children, the dataset effectively supports both ASR and SV tasks. This dataset is a valuable contribution to Mandarin child speech research. The dataset is now open-source and freely available for all academic purposes on https://github.com/flageval-baai/ChildMandarin.

cs.SD↗

arXiv2025

CS-Dialogue: A 104-Hour Dataset of Spontaneous Mandarin-English Code-Switching Dialogues for Speech Recognition

Code-switching (CS), the alternation between two or more languages within a single conversation, presents significant challenges for automatic speech recognition (ASR) systems. Existing Mandarin-English code-switching datasets often suffer from limitations in size, spontaneity, and the lack of full-length dialogue recordings with transcriptions, hindering the development of robust ASR models for real-world conversational scenarios. This paper introduces CS-Dialogue, a novel large-scale Mandarin-English code-switching speech dataset comprising 104 hours of spontaneous conversations from 200 speakers. Unlike previous datasets, CS-Dialogue provides full-length dialogue recordings with complete transcriptions, capturing naturalistic code-switching patterns in continuous speech. We describe the data collection and annotation processes, present detailed statistics of the dataset, and establish benchmark ASR performance using state-of-the-art models. Our experiments, using Transformer, Conformer, and Branchformer, demonstrate the challenges of code-switching ASR, and show that existing pre-trained models such as Whisper still have the space to improve. The CS-Dialogue dataset will be made freely available for all academic purposes.

cs.CL↗

arXiv2024

A silicon photonics waveguide-coupled colloidal quantum dot photodiode sensitive beyond 1.6 um

Silicon photonics faces a persistent challenge in extending photodetection capabilities beyond the 1.6 um wavelength range, primarily due to the lack of appropriate epitaxial materials. Colloidal quantum dots (QDs) present a promising solution here, offering distinct advantages such as infrared wavelength tunability, cost-effectiveness, and facile deposition. Their unique properties position them as a potential candidate for enabling photodetection in silicon photonics beyond the conventional telecom wavelength, thereby expanding the potential applications and capabilities within this domain. In this study, we have successfully integrated lead sulfide (PbS) colloidal quantum dot photodiodes (QDPDs) onto silicon waveguides using standard process techniques. The integrated photodiodes exhibit a remarkable responsivity of 1.3 A/W (with an external quantum efficiency of 74.8%) at a wavelength of 2.1 um, a low dark current of only 106 nA and a bandwidth of 1.1 MHz under a -3 V bias. To demonstrate the scalability of our integration approach, we have developed a compact 8-channel spectrometer incorporating an array of QDPDs. This achievement marks a significant step toward realizing a cost-effective photodetector solution for silicon photonics, particularly tailored for a wide range of sensing applications around the 2 um wavelength range.

physics.optics↗

arXiv2014

Electromagnetially-induced-transparency-like ground-state cooling in a double-cavity optomechanical system

We propose to cool a mechanical resonator close to its ground state via an electromagnetically-induced-transparency- (EIT-) like cooling mechanism in a double-cavity optomechanical system, where an additional cavity couples to the original one in the standard optomechanical system. By choosing optimal parameters such that the cooling process of the mechanical resonator corresponds to the maximum value of the optical fluctuation spectrum and the heating process to the minimum one, the mechanical resonator can be cooled with the final mean phonon number less than that at the absence of the additional cavity. And we show the mechanical resonator may be cooled close to its ground state via such an EIT-like cooling mechanism even when the original resolved sideband condition is not fulfilled at the absence of the additional cavity.

quant-ph↗