arXiv subjects

Lei Xie

Publications and source records attributed to Lei Xie.

At least 163 records · Page 9Linked to original sources

arXiv2025

ContextASR-Bench: A Massive Contextual Speech Recognition Benchmark

Automatic Speech Recognition (ASR) has been extensively investigated, yet prior benchmarks have largely focused on assessing the acoustic robustness of ASR models, leaving evaluations of their linguistic capabilities relatively underexplored. This largely stems from the limited parameter sizes and training corpora of conventional ASR models, leaving them with insufficient world knowledge, which is crucial for accurately recognizing named entities across diverse domains. For instance, drug and treatment names in medicine or specialized technical terms in engineering. Recent breakthroughs in Large Language Models (LLMs) and corresponding Large Audio Language Models (LALMs) have markedly enhanced the visibility of advanced context modeling and general artificial intelligence capabilities. Leveraging LLMs, we envision a unified system capable of robust speech recognition across diverse real-world domains, yet existing benchmarks are inadequate for evaluating this objective. To address this gap, we propose ContextASR-Bench: a comprehensive, large-scale benchmark designed to assess the linguistic competence of ASR systems using corpora that feature numerous named entities across multiple domains. It encompasses up to 40,000 data entries with more than 300,000 named entities across over 10 domains. Beyond the audio and its transcription, each sample provides the domain it belongs to and a list of named entities it contains, which are referred to as the context. Based on this, we introduce three evaluation modes to assess how effectively models can exploit such context to improve ASR accuracy. Extensive evaluation on ContextASR-Bench highlights that LALMs outperform conventional ASR models by a large margin thanks to the strong world knowledge and context modeling of LLMs, yet there remains ample room for further improvement. The dataset and evaluation code have been released.

eess.AS↗

arXiv2025

Efficient Scaling for LLM-based ASR

Large language model (LLM)-based automatic speech recognition (ASR) achieves strong performance but often incurs high computational costs. This work investigates how to obtain the best LLM-ASR performance efficiently. Through comprehensive and controlled experiments, we find that pretraining the speech encoder before integrating it with the LLM leads to significantly better scaling efficiency than the standard practice of joint post-training of LLM-ASR. Based on this insight, we propose a new multi-stage LLM-ASR training strategy, EFIN: Encoder First Integration. Among all training strategies evaluated, EFIN consistently delivers better performance (relative to 21.1% CERR) with significantly lower computation budgets (49.9% FLOPs). Furthermore, we derive a scaling law that approximates ASR error rates as a computation function, providing practical guidance for LLM-ASR scaling.

cs.SD↗

arXiv2025

Residual Koopman Model Predictive Control for Enhanced Vehicle Dynamics with Small On-Track Data Input

In vehicle trajectory tracking tasks, the simplest approach is the Pure Pursuit (PP) Control. However, this single-point preview tracking strategy fails to consider vehicle model constraints, compromising driving safety. Model Predictive Control (MPC) as a widely adopted control method, optimizes control actions by incorporating mechanistic models and physical constraints. While its control performance critically depends on the accuracy of vehicle modeling. Traditional vehicle modeling approaches face inherent trade-offs between capturing nonlinear dynamics and maintaining computational efficiency, often resulting in reduced control performance. To address these challenges, this paper proposes Residual Koopman Model Predictive Control (RKMPC) framework. This method uses two linear MPC architecture to calculate control inputs: a Linear Model Predictive Control (LMPC) computes the baseline control input based on the vehicle kinematic model, and a neural network-based RKMPC calculates the compensation input. The final control command is obtained by adding these two components. This design preserves the reliability and interpretability of traditional mechanistic model while achieving performance optimization through residual modeling. This method has been validated on the Carsim-Matlab joint simulation platform and a physical 1:10 scale F1TENTH racing car. Experimental results show that RKMPC requires only 20% of the training data needed by traditional Koopman Model Predictive Control (KMPC) while delivering superior tracking performance. Compared to traditional LMPC, RKMPC reduces lateral error by 11.7%-22.1%, decreases heading error by 8.9%-15.8%, and improves front-wheel steering stability by up to 27.6%. The implementation code is available at: https://github.com/ZJU-DDRX/Residual Koopman.

cs.RO↗

arXiv2025

Tractography-Guided Dual-Label Collaborative Learning for Multi-Modal Cranial Nerves Parcellation

The parcellation of Cranial Nerves (CNs) serves as a crucial quantitative methodology for evaluating the morphological characteristics and anatomical pathways of specific CNs. Multi-modal CNs parcellation networks have achieved promising segmentation performance, which combine structural Magnetic Resonance Imaging (MRI) and diffusion MRI. However, insufficient exploration of diffusion MRI information has led to low performance of existing multi-modal fusion. In this work, we propose a tractography-guided Dual-label Collaborative Learning Network (DCLNet) for multi-modal CNs parcellation. The key contribution of our DCLNet is the introduction of coarse labels of CNs obtained from fiber tractography through CN atlas, and collaborative learning with precise labels annotated by experts. Meanwhile, we introduce a Modality-adaptive Encoder Module (MEM) to achieve soft information swapping between structural MRI and diffusion MRI. Extensive experiments conducted on the publicly available Human Connectome Project (HCP) dataset demonstrate performance improvements compared to single-label network. This systematic validation underscores the effectiveness of dual-label strategies in addressing inherent ambiguities in CNs parcellation tasks.

eess.IV↗

arXiv2025

Automated Mapping the Pathways of Cranial Nerve II, III, V, and VII/VIII: A Multi-Parametric Multi-Stage Diffusion Tractography Atlas

Cranial nerves (CNs) play a crucial role in various essential functions of the human brain, and mapping their pathways from diffusion MRI (dMRI) provides valuable preoperative insights into the spatial relationships between individual CNs and key tissues. However, mapping a comprehensive and detailed CN atlas is challenging because of the unique anatomical structures of each CN pair and the complexity of the skull base environment.In this work, we present what we believe to be the first study to develop a comprehensive diffusion tractography atlas for automated mapping of CN pathways in the human brain. The CN atlas is generated by fiber clustering by using the streamlines generated by multi-parametric fiber tractography for each pair of CNs. Instead of disposable clustering, we explore a new strategy of multi-stage fiber clustering for multiple analysis of approximately 1,000,000 streamlines generated from the 50 subjects from the Human Connectome Project (HCP). Quantitative and visual experiments demonstrate that our CN atlas achieves high spatial correspondence with expert manual annotations on multiple acquisition sites, including the HCP dataset, the Multi-shell Diffusion MRI (MDM) dataset and two clinical cases of pituitary adenoma patients. The proposed CN atlas can automatically identify 8 fiber bundles associated with 5 pairs of CNs, including the optic nerve CN II, oculomotor nerve CN III, trigeminal nerve CN V and facial-vestibulocochlear nerve CN VII/VIII, and its robustness is demonstrated experimentally. This work contributes to the field of diffusion imaging by facilitating more efficient and automated mapping the pathways of multiple pairs of CNs, thereby enhancing the analysis and understanding of complex brain structures through visualization of their spatial relationships with nearby anatomy.

cs.CV↗

arXiv2025

Programmable Virtual Humans Toward Human Physiologically-Based Drug Discovery

Artificial intelligence (AI) has sparked immense interest in drug discovery, but most current approaches only digitize existing high-throughput experiments. They remain constrained by conventional pipelines. As a result, they do not address the fundamental challenges of predicting drug effects in humans. Similarly, biomedical digital twins, largely grounded in real-world data and mechanistic models, are tailored for late-phase drug development and lack the resolution to model molecular interactions or their systemic consequences, limiting their impact in early-stage discovery. This disconnect between early discovery and late development is one of the main drivers of high failure rates in drug discovery. The true promise of AI lies not in augmenting current experiments but in enabling virtual experiments that are impossible in the real world: testing novel compounds directly in silico in the human body. Recent advances in AI, high-throughput perturbation assays, and single-cell and spatial omics across species now make it possible to construct programmable virtual humans: dynamic, multiscale models that simulate drug actions from molecular to phenotypic levels. By bridging the translational gap, programmable virtual humans offer a transformative path to optimize therapeutic efficacy and safety earlier than ever before. This perspective introduces the concept of programmable virtual humans, explores their roles in a new paradigm of drug discovery centered on human physiology, and outlines key opportunities, challenges, and roadmaps for their realization.

cs.CY↗

arXiv2025

DiffRhythm+: Controllable and Flexible Full-Length Song Generation with Preference Optimization

Songs, as a central form of musical art, exemplify the richness of human intelligence and creativity. While recent advances in generative modeling have enabled notable progress in long-form song generation, current systems for full-length song synthesis still face major challenges, including data imbalance, insufficient controllability, and inconsistent musical quality. DiffRhythm, a pioneering diffusion-based model, advanced the field by generating full-length songs with expressive vocals and accompaniment. However, its performance was constrained by an unbalanced model training dataset and limited controllability over musical style, resulting in noticeable quality disparities and restricted creative flexibility. To address these limitations, we propose DiffRhythm+, an enhanced diffusion-based framework for controllable and flexible full-length song generation. DiffRhythm+ leverages a substantially expanded and balanced training dataset to mitigate issues such as repetition and omission of lyrics, while also fostering the emergence of richer musical skills and expressiveness. The framework introduces a multi-modal style conditioning strategy, enabling users to precisely specify musical styles through both descriptive text and reference audio, thereby significantly enhancing creative control and diversity. We further introduce direct performance optimization aligned with user preferences, guiding the model toward consistently preferred outputs across evaluation metrics. Extensive experiments demonstrate that DiffRhythm+ achieves significant improvements in naturalness, arrangement complexity, and listener satisfaction over previous systems.

eess.AS↗

arXiv2025

Mixture of LoRA Experts with Multi-Modal and Multi-Granularity LLM Generative Error Correction for Accented Speech Recognition

Despite improvements in automatic speech recognition, performance drops with accented speech. Generative error correction (GER) leverages the linguistic knowledge of large language models (LLMs), outperforming typical language model methods. However, it lacks specificity in accented speech scenarios. Accents represent deviations from standard pronunciation, making multi-granularity pronunciation and semantic information essential for accented speech recognition. Moreover, accents exhibit considerable diversity, with each accent possessing distinct characteristics. In this study, we leverage GER to improve transcription accuracy by addressing the two primary features. We propose the multi-modal GER, which integrates pronunciation information from the speech modality, and the multi-granularity GER, which incorporates fine-grained phoneme-level pronunciation information. These methods enable the LLM to utilize the pronunciation information of accented speech and the semantic information from word-level hypotheses for accurate transcription predictions through low-rank adaptation (LoRA) fine-tuning. We employ a three-stage strategy to train separate multi-modal GER models for each accent to obtain mono-accent LoRA experts. By adopting our proposed HDMoLE method, which incorporates hierarchical routing and dynamic thresholds within the mixture of LoRA experts, we effectively merge mono-accent LoRA experts within a single multi-modal GER to overcome accent diversity challenges. Furthermore, multi-granularity GER leverages N-best word-level and phoneme-level hypotheses from the HDMoLE model to predict final transcriptions. Experiments on a multi-accent English dataset show that our methods reduce word error rate by 67.35% compared to the baseline vanilla Whisper-large-v3 model.

cs.SD↗

arXiv2025

DualDub: Video-to-Soundtrack Generation via Joint Speech and Background Audio Synthesis

While recent video-to-audio (V2A) models can generate realistic background audio from visual input, they largely overlook speech, an essential part of many video soundtracks. This paper proposes a new task, video-to-soundtrack (V2ST) generation, which aims to jointly produce synchronized background audio and speech within a unified framework. To tackle V2ST, we introduce DualDub, a unified framework built on a multimodal language model that integrates a multimodal encoder, a cross-modal aligner, and dual decoding heads for simultaneous background audio and speech generation. Specifically, our proposed cross-modal aligner employs causal and non-causal attention mechanisms to improve synchronization and acoustic harmony. Besides, to handle data scarcity, we design a curriculum learning strategy that progressively builds the multimodal capability. Finally, we introduce DualBench, the first benchmark for V2ST evaluation with a carefully curated test set and comprehensive metrics. Experimental results demonstrate that DualDub achieves state-of-the-art performance, generating high-quality and well-synchronized soundtracks with both speech and background audio.

cs.MM↗

arXiv2025

A Learning-based Planning and Control Framework for Inertia Drift Vehicles

Inertia drift is a transitional maneuver between two sustained drift stages in opposite directions, which provides valuable insights for navigating consecutive sharp corners for autonomous racing.However, this can be a challenging scenario for the drift controller to handle rapid transitions between opposing sideslip angles while maintaining accurate path tracking. Moreover, accurate drift control depends on a high-fidelity vehicle model to derive drift equilibrium points and predict vehicle states, but this is often compromised by the strongly coupled longitudinal-lateral drift dynamics and unpredictable environmental variations. To address these challenges, this paper proposes a learning-based planning and control framework utilizing Bayesian optimization (BO), which develops a planning logic to ensure a smooth transition and minimal velocity loss between inertia and sustained drift phases. BO is further employed to learn a performance-driven control policy that mitigates modeling errors for enhanced system performance. Simulation results on an 8-shape reference path demonstrate that the proposed framework can achieve smooth and stable inertia drift through sharp corners.

cs.RO↗

arXiv2025

Convergent and divergent connectivity patterns of the arcuate fasciculus in macaques and humans

The organization and connectivity of the arcuate fasciculus (AF) in nonhuman primates remain contentious, especially concerning how its anatomy diverges from that of humans. Here, we combined cross-scale single-neuron tracing - using viral-based genetic labeling and fluorescence micro-optical sectioning tomography in macaques (n = 4; age 3 - 11 years) - with whole-brain tractography from 11.7T diffusion MRI. Complemented by spectral embedding analysis of 7.0T MRI in humans, we performed a comparative connectomic analysis of the AF across species. We demonstrate that the macaque AF originates in the temporal-parietal cortex, traverses the auditory cortex and parietal operculum, and projects into prefrontal regions. In contrast, the human AF exhibits greater expansion into the middle temporal gyrus and stronger prefrontal and parietal operculum connectivity - divergences quantified by Kullback-Leibler analysis that likely underpin the evolutionary specialization of human language networks. These interspecies differences - particularly the human AF's broader temporal integration and strengthened frontoparietal linkages - suggest a connectivity-based substrate for the emergence of advanced language processing unique to humans. Furthermore, our findings offer a neuroanatomical framework for understanding AF-related disorders such as aphasia and dyslexia, where aberrant connectivity disrupts language function.

q-bio.NC↗

arXiv2025

StreamFlow: Streaming Flow Matching with Block-wise Guided Attention Mask for Speech Token Decoding

Recent advancements in discrete token-based speech generation have highlighted the importance of token-to-waveform generation for audio quality, particularly in real-time interactions. Traditional frameworks integrating semantic tokens with flow matching (FM) struggle with streaming capabilities due to their reliance on a global receptive field. Additionally, directly implementing token-by-token streaming speech generation often results in degraded audio quality. To address these challenges, we propose StreamFlow, a novel neural architecture that facilitates streaming flow matching with diffusion transformers (DiT). To mitigate the long-sequence extrapolation issues arising from lengthy historical dependencies, we design a local block-wise receptive field strategy. Specifically, the sequence is first segmented into blocks, and we introduce block-wise attention masks that enable the current block to receive information from the previous or subsequent block. These attention masks are combined hierarchically across different DiT-blocks to regulate the receptive field of DiTs. Both subjective and objective experimental results demonstrate that our approach achieves performance comparable to non-streaming methods while surpassing other streaming methods in terms of speech quality, all the while effectively managing inference time during long-sequence generation. Furthermore, our method achieves a notable first-packet latency of only 180 ms.\footnote{Speech samples: https://dukguo.github.io/StreamFlow/}

cs.SD↗

arXiv2025

Safe Reinforcement Learning with a Predictive Safety Filter for Motion Planning and Control: A Drifting Vehicle Example

Autonomous drifting is a complex and crucial maneuver for safety-critical scenarios like slippery roads and emergency collision avoidance, requiring precise motion planning and control. Traditional motion planning methods often struggle with the high instability and unpredictability of drifting, particularly when operating at high speeds. Recent learning-based approaches have attempted to tackle this issue but often rely on expert knowledge or have limited exploration capabilities. Additionally, they do not effectively address safety concerns during learning and deployment. To overcome these limitations, we propose a novel Safe Reinforcement Learning (RL)-based motion planner for autonomous drifting. Our approach integrates an RL agent with model-based drift dynamics to determine desired drift motion states, while incorporating a Predictive Safety Filter (PSF) that adjusts the agent's actions online to prevent unsafe states. This ensures safe and efficient learning, and stable drift operation. We validate the effectiveness of our method through simulations on a Matlab-Carsim platform, demonstrating significant improvements in drift performance, reduced tracking errors, and computational efficiency compared to traditional methods. This strategy promises to extend the capabilities of autonomous vehicles in safety-critical maneuvers.

cs.RO↗

arXiv2025

LLaSE-G1: Incentivizing Generalization Capability for LLaMA-based Speech Enhancement

Recent advancements in language models (LMs) have demonstrated strong capabilities in semantic understanding and contextual modeling, which have flourished in generative speech enhancement (SE). However, many LM-based SE approaches primarily focus on semantic information, often neglecting the critical role of acoustic information, which leads to acoustic inconsistency after enhancement and limited generalization across diverse SE tasks. In this paper, we introduce LLaSE-G1, a LLaMA-based language model that incentivizes generalization capabilities for speech enhancement. LLaSE-G1 offers the following key contributions: First, to mitigate acoustic inconsistency, LLaSE-G1 employs continuous representations from WavLM as input and predicts speech tokens from X-Codec2, maximizing acoustic preservation. Second, to promote generalization capability, LLaSE-G1 introduces dual-channel inputs and outputs, unifying multiple SE tasks without requiring task-specific IDs. Third, LLaSE-G1 outperforms prior task-specific discriminative and generative SE models, demonstrating scaling effects at test time and emerging capabilities for unseen SE tasks. Additionally, we release our code and models to support further research in this area.

eess.AS↗

arXiv2025

Learning to Drift in Extreme Turning with Active Exploration and Gaussian Process Based MPC

Extreme cornering in racing often leads to large sideslip angles, presenting a significant challenge for vehicle control. Conventional vehicle controllers struggle to manage this scenario, necessitating the use of a drifting controller. However, the large sideslip angle in drift conditions introduces model mismatch, which in turn affects control precision. To address this issue, we propose a model correction drift controller that integrates Model Predictive Control (MPC) with Gaussian Process Regression (GPR). GPR is employed to correct vehicle model mismatches during both drift equilibrium solving and the MPC optimization process. Additionally, the variance from GPR is utilized to actively explore different cornering drifting velocities, aiming to minimize trajectory tracking errors. The proposed algorithm is validated through simulations on the Simulink-Carsim platform and experiments with a 1:10 scale RC vehicle. In the simulation, the average lateral error with GPR is reduced by 52.8% compared to the non-GPR case. Incorporating exploration further decreases this error by 27.1%. The velocity tracking Root Mean Square Error (RMSE) also decreases by 10.6% with exploration. In the RC car experiment, the average lateral error with GPR is 36.7% lower, and exploration further leads to a 29.0% reduction. Moreover, the velocity tracking RMSE decreases by 7.2% with the inclusion of exploration.

cs.RO↗

arXiv2025

EASY: Emotion-aware Speaker Anonymization via Factorized Distillation

Emotion plays a significant role in speech interaction, conveyed through tone, pitch, and rhythm, enabling the expression of feelings and intentions beyond words to create a more personalized experience. However, most existing speaker anonymization systems employ parallel disentanglement methods, which only separate speech into linguistic content and speaker identity, often neglecting the preservation of the original emotional state. In this study, we introduce EASY, an emotion-aware speaker anonymization framework. EASY employs a novel sequential disentanglement process to disentangle speaker identity, linguistic content, and emotional representation, modeling each speech attribute in distinct subspaces through a factorized distillation approach. By independently constraining speaker identity and emotional representation, EASY minimizes information leakage, enhancing privacy protection while preserving original linguistic content and emotional state. Experimental results on the VoicePrivacy Challenge official datasets demonstrate that our proposed approach outperforms all baseline systems, effectively protecting speaker privacy while maintaining linguistic content and emotional state.

eess.AS↗

arXiv2025

AISHELL-5: The First Open-Source In-Car Multi-Channel Multi-Speaker Speech Dataset for Automatic Speech Diarization and Recognition

This paper delineates AISHELL-5, the first open-source in-car multi-channel multi-speaker Mandarin automatic speech recognition (ASR) dataset. AISHLL-5 includes two parts: (1) over 100 hours of multi-channel speech data recorded in an electric vehicle across more than 60 real driving scenarios. This audio data consists of four far-field speech signals captured by microphones located on each car door, as well as near-field signals obtained from high-fidelity headset microphones worn by each speaker. (2) a collection of 40 hours of real-world environmental noise recordings, which supports the in-car speech data simulation. Moreover, we also provide an open-access, reproducible baseline system based on this dataset. This system features a speech frontend model that employs speech source separation to extract each speaker's clean speech from the far-field signals, along with a speech recognition module that accurately transcribes the content of each individual speaker. Experimental results demonstrate the challenges faced by various mainstream ASR models when evaluated on the AISHELL-5. We firmly believe the AISHELL-5 dataset will significantly advance the research on ASR systems under complex driving scenarios by establishing the first publicly available in-car ASR benchmark.

cs.SD↗

arXiv2025

Contextualized Automatic Speech Recognition with Dynamic Vocabulary Prediction and Activation

Deep biasing improves automatic speech recognition (ASR) performance by incorporating contextual phrases. However, most existing methods enhance subwords in a contextual phrase as independent units, potentially compromising contextual phrase integrity, leading to accuracy reduction. In this paper, we propose an encoder-based phrase-level contextualized ASR method that leverages dynamic vocabulary prediction and activation. We introduce architectural optimizations and integrate a bias loss to extend phrase-level predictions based on frame-level outputs. We also introduce a confidence-activated decoding method that ensures the complete output of contextual phrases while suppressing incorrect bias. Experiments on Librispeech and Wenetspeech datasets demonstrate that our approach achieves relative WER reductions of 28.31% and 23.49% compared to baseline, with the WER on contextual phrases decreasing relatively by 72.04% and 75.69%.

cs.SD↗