Xiong Wang

Chaoyou Fu, Haojia Lin, Zuwei Long et al.

The remarkable multimodal capabilities and interactive experience of GPT-4o underscore their necessity in practical applications, yet open-source models rarely excel in both areas. In this paper, we introduce VITA, the first-ever open-source Multimodal Large Language Model (MLLM) adept at simultaneous processing and analysis of Video, Image, Text, and Audio modalities, and meanwhile has an advanced multimodal interactive experience. Starting from Mixtral 8x7B as a language foundation, we expand its Chinese vocabulary followed by bilingual instruction tuning. We further endow the language model with visual and audio capabilities through two-stage multi-task learning of multimodal alignment and instruction tuning. VITA demonstrates robust foundational capabilities of multilingual, vision, and audio understanding, as evidenced by its strong performance across a range of both unimodal and multimodal benchmarks. Beyond foundational capabilities, we have made considerable progress in enhancing the natural multimodal human-computer interaction experience. VITA is the first step for the open-source community to explore the seamless integration of multimodal understanding and interaction. While there is still lots of work to be done on VITA to get close to close-source counterparts, we hope that its role as a pioneer can serve as a cornerstone for subsequent research. Project Page: https://vita-home.github.io.

Matthias Eisenmann, Annika Reinke, Vivienn Weru et al. · utoronto

The number of international benchmarking competitions is steadily increasing in various fields of machine learning (ML) research and practice. So far, however, little is known about the common practice as well as bottlenecks faced by the community in tackling the research questions posed. To shed light on the status quo of algorithm development in the specific field of biomedical imaging analysis, we designed an international survey that was issued to all participants of challenges conducted in conjunction with the IEEE ISBI 2021 and MICCAI 2021 conferences (80 competitions in total). The survey covered participants' expertise and working environments, their chosen strategies, as well as algorithm characteristics. A median of 72% challenge participants took part in the survey. According to our results, knowledge exchange was the primary incentive (70%) for participation, while the reception of prize money played only a minor role (16%). While a median of 80 working hours was spent on method development, a large portion of participants stated that they did not have enough time for method development (32%). 25% perceived the infrastructure to be a bottleneck. Overall, 94% of all solutions were deep learning-based. Of these, 84% were based on standard architectures. 43% of the respondents reported that the data samples (e.g., images) were too large to be processed at once. This was most commonly addressed by patch-based training (69%), downsampling (37%), and solving 3D analysis tasks as a series of 2D tasks. K-fold cross-validation on the training set was performed by only 37% of the participants and only 50% of the participants performed ensembling based on multiple identical models (61%) or heterogeneous models (39%). 48% of the respondents applied postprocessing steps.

Xian Shi, Xiong Wang, Zhifang Guo et al.

In this report, we introduce Qwen3-ASR family, which includes two powerful all-in-one speech recognition models and a novel non-autoregressive speech forced alignment model. Qwen3-ASR-1.7B and Qwen3-ASR-0.6B are ASR models that support language identification and ASR for 52 languages and dialects. Both of them leverage large-scale speech training data and the strong audio understanding ability of their foundation model Qwen3-Omni. We conduct comprehensive internal evaluation besides the open-sourced benchmarks as ASR models might differ little on open-sourced benchmark scores but exhibit significant quality differences in real-world scenarios. The experiments reveal that the 1.7B version achieves SOTA performance among open-sourced ASR models and is competitive with the strongest proprietary APIs while the 0.6B version offers the best accuracy-efficiency trade-off. Qwen3-ASR-0.6B can achieve an average TTFT as low as 92ms and transcribe 2000 seconds speech in 1 second at a concurrency of 128. Qwen3-ForcedAligner-0.6B is an LLM based NAR timestamp predictor that is able to align text-speech pairs in 11 languages. Timestamp accuracy experiments show that the proposed model outperforms the three strongest force alignment models and takes more advantages in efficiency and versatility. To further accelerate the community research of ASR and audio understanding, we release these models under the Apache 2.0 license.

Zhanheng Yang, Sining Sun, Xiong Wang et al.

It is difficult for an E2E ASR system to recognize words such as entities appearing infrequently in the training data. A widely used method to mitigate this issue is feeding contextual information into the acoustic model. Previous works have proven that a compact and accurate contextual list can boost the performance significantly. In this paper, we propose an efficient approach to obtain a high quality contextual list for a unified streaming/non-streaming based E2E model. Specifically, we make use of the phone-level streaming output to first filter the predefined contextual word list then fuse it into non-casual encoder and decoder to generate the final recognition results. Our approach improve the accuracy of the contextual ASR system and speed up the inference process. Experiments on two datasets demonstrates over 20% CER reduction comparing to the baseline system. Meanwhile, the RTF of our system can be stabilized within 0.15 when the size of the contextual word list grows over 6,000.

Zheng Lin, Zhe Chen, Zihan Fang et al.

Recently, a large number of Low Earth Orbit (LEO) satellites have been launched and deployed successfully in space by commercial companies, such as SpaceX. Due to multimodal sensors equipped by the LEO satellites, they serve not only for communication but also for various machine learning applications, such as space modulation recognition, remote sensing image classification, etc. However, the ground station (GS) may be incapable of downloading such a large volume of raw sensing data for centralized model training due to the limited contact time with LEO satellites (e.g. 5 minutes). Therefore, federated learning (FL) has emerged as the promising solution to address this problem via on-device training. Unfortunately, to enable FL on LEO satellites, we still face three critical challenges that are i) heterogeneous computing and memory capabilities, ii) limited uplink rate, and iii) model staleness. To this end, we propose FedSN as a general FL framework to tackle the above challenges, and fully explore data diversity on LEO satellites. Specifically, we first present a novel sub-structure scheme to enable heterogeneous local model training considering different computing, memory, and communication constraints on LEO satellites. Additionally, we propose a pseudo-synchronous model aggregation strategy to dynamically schedule model aggregation for compensating model staleness. To further demonstrate the effectiveness of the FedSN, we evaluate it using space modulation recognition and remote sensing image classification tasks by leveraging the data from real-world satellite networks. Extensive experimental results demonstrate that FedSN framework achieves higher accuracy, lower computing, and communication overhead than the state-of-the-art benchmarks and the effectiveness of each components in FedSN.

Chaoyou Fu, Haojia Lin, Xiong Wang et al.

Recent Multimodal Large Language Models (MLLMs) have typically focused on integrating visual and textual modalities, with less emphasis placed on the role of speech in enhancing interaction. However, speech plays a crucial role in multimodal dialogue systems, and implementing high-performance in both vision and speech tasks remains a significant challenge due to the fundamental modality differences. In this paper, we propose a carefully designed multi-stage training methodology that progressively trains LLM to understand both visual and speech information, ultimately enabling fluent vision and speech interaction. Our approach not only preserves strong vision-language capacity, but also enables efficient speech-to-speech dialogue capabilities without separate ASR and TTS modules, significantly accelerating multimodal end-to-end response speed. By comparing our method against state-of-the-art counterparts across benchmarks for image, video, and speech tasks, we demonstrate that our model is equipped with both strong visual and speech capabilities, making near real-time vision and speech interaction. Code has been released at https://github.com/VITA-MLLM/VITA.

Hangrui Hu, Xinfa Zhu, Ting He et al.

In this report, we present the Qwen3-TTS series, a family of advanced multilingual, controllable, robust, and streaming text-to-speech models. Qwen3-TTS supports state-of-the-art 3-second voice cloning and description-based control, allowing both the creation of entirely novel voices and fine-grained manipulation over the output speech. Trained on over 5 million hours of speech data spanning 10 languages, Qwen3-TTS adopts a dual-track LM architecture for real-time synthesis, coupled with two speech tokenizers: 1) Qwen-TTS-Tokenizer-25Hz is a single-codebook codec emphasizing semantic content, which offers seamlessly integration with Qwen-Audio and enables streaming waveform reconstruction via a block-wise DiT. 2) Qwen-TTS-Tokenizer-12Hz achieves extreme bitrate reduction and ultra-low-latency streaming, enabling immediate first-packet emission ($97\,\mathrm{ms}$) through its 12.5 Hz, 16-layer multi-codebook design and a lightweight causal ConvNet. Extensive experiments indicate state-of-the-art performance across diverse objective and subjective benchmark (e.g., TTS multilingual test set, InstructTTSEval, and our long speech test set). To facilitate community research and development, we release both tokenizers and models under the Apache 2.0 license.

Jin Xu, Zhifang Guo, Hangrui Hu et al. · pku

We present Qwen3-Omni, a single multimodal model that, for the first time, maintains state-of-the-art performance across text, image, audio, and video without any degradation relative to single-modal counterparts. Qwen3-Omni matches the performance of same-sized single-modal models within the Qwen series and excels particularly on audio tasks. Across 36 audio and audio-visual benchmarks, Qwen3-Omni achieves open-source SOTA on 32 benchmarks and overall SOTA on 22, outperforming strong closed-source models such as Gemini-2.5-Pro, Seed-ASR, and GPT-4o-Transcribe. Qwen3-Omni adopts a Thinker-Talker MoE architecture that unifies perception and generation across text, images, audio, and video, yielding fluent text and natural real-time speech. It supports text interaction in 119 languages, speech understanding in 19 languages, and speech generation in 10 languages. To reduce first-packet latency in streaming synthesis, Talker autoregressively predicts discrete speech codecs using a multi-codebook scheme. Leveraging the representational capacity of these codebooks, we replace computationally intensive block-wise diffusion with a lightweight causal ConvNet, enabling streaming from the first codec frame. In cold-start settings, Qwen3-Omni achieves a theoretical end-to-end first-packet latency of 234 ms. To further strengthen multimodal reasoning, we introduce a Thinking model that explicitly reasons over inputs from any modality. Since the research community currently lacks a general-purpose audio captioning model, we fine-tuned Qwen3-Omni-30B-A3B to obtain Qwen3-Omni-30B-A3B-Captioner, which produces detailed, low-hallucination captions for arbitrary audio inputs. Qwen3-Omni-30B-A3B, Qwen3-Omni-30B-A3B-Thinking, and Qwen3-Omni-30B-A3B-Captioner are publicly released under the Apache 2.0 license.

Yunhang Shen, Chaoyou Fu, Shaoqi Dong et al.

We introduce Long-VITA, a simple yet effective large multi-modal model for long-context visual-language understanding tasks. It is adept at concurrently processing and analyzing modalities of image, video, and text over 4K frames or 1M tokens while delivering advanced performances on short-context multi-modal tasks. We propose an effective multi-modal training schema that starts with large language models and proceeds through vision-language alignment, general knowledge learning, and two sequential stages of long-sequence fine-tuning. We further implement context-parallelism distributed inference and logits-masked language modeling head to scale Long-VITA to infinitely long inputs of images and texts during model inference. Regarding training data, Long-VITA is built on a mix of 17M samples from public datasets only and demonstrates state-of-the-art performance on various multi-modal benchmarks, compared against recent cutting-edge models with internal data. Long-VITA is fully open-source and reproducible.. By leveraging our inference designs, Long-VITA models achieve a remarkable 2x prefill speedup and 4x context length extension in a single node with 8 GPUs. We hope Long-VITA can serve as a competitive baseline and offer valuable insights for the open-source community in advancing long-context multi-modal understanding.

Neil K. Chada, Quanjun Lang, Fei Lu et al.

Kernels are efficient in representing nonlocal dependence and they are widely used to design operators between function spaces. Thus, learning kernels in operators from data is an inverse problem of general interest. Due to the nonlocal dependence, the inverse problem can be severely ill-posed with a data-dependent singular inversion operator. The Bayesian approach overcomes the ill-posedness through a non-degenerate prior. However, a fixed non-degenerate prior leads to a divergent posterior mean when the observation noise becomes small, if the data induces a perturbation in the eigenspace of zero eigenvalues of the inversion operator. We introduce a data-adaptive prior to achieve a stable posterior whose mean always has a small noise limit. The data-adaptive prior's covariance is the inversion operator with a hyper-parameter selected adaptive to data by the L-curve method. Furthermore, we provide a detailed analysis on the computational practice of the data-adaptive prior, and demonstrate it on Toeplitz matrices and integral operators. Numerical tests show that a fixed prior can lead to a divergent posterior mean in the presence of any of the four types of errors: discretization error, model error, partial observation and wrong noise assumption. In contrast, the data-adaptive prior always attains posterior means with small noise limits.

Jin Xu, Zhifang Guo, Jinzheng He et al.

In this report, we present Qwen2.5-Omni, an end-to-end multimodal model designed to perceive diverse modalities, including text, images, audio, and video, while simultaneously generating text and natural speech responses in a streaming manner. To enable the streaming of multimodal information inputs, both audio and visual encoders utilize a block-wise processing approach. To synchronize the timestamps of video inputs with audio, we organize the audio and video sequentially in an interleaved manner and propose a novel position embedding approach, named TMRoPE(Time-aligned Multimodal RoPE). To concurrently generate text and speech while avoiding interference between the two modalities, we propose \textbf{Thinker-Talker} architecture. In this framework, Thinker functions as a large language model tasked with text generation, while Talker is a dual-track autoregressive model that directly utilizes the hidden representations from the Thinker to produce audio tokens as output. Both the Thinker and Talker models are designed to be trained and inferred in an end-to-end manner. For decoding audio tokens in a streaming manner, we introduce a sliding-window DiT that restricts the receptive field, aiming to reduce the initial package delay. Qwen2.5-Omni is comparable with the similarly sized Qwen2.5-VL and outperforms Qwen2-Audio. Furthermore, Qwen2.5-Omni achieves state-of-the-art performance on multimodal benchmarks like Omni-Bench. Notably, Qwen2.5-Omni's performance in end-to-end speech instruction following is comparable to its capabilities with text inputs, as evidenced by benchmarks such as MMLU and GSM8K. As for speech generation, Qwen2.5-Omni's streaming Talker outperforms most existing streaming and non-streaming alternatives in robustness and naturalness.

Yian Liu, Xiong Wang, Ping Xu et al.

Real-time lane detection in embedded systems encounters significant challenges due to subtle and sparse visual signals in RGB images, often constrained by limited computational resources and power consumption. Although deep learning models for lane detection categorized into segmentation-based, anchor-based, and curve-based methods there remains a scarcity of universally applicable optimization techniques tailored for low-power embedded environments. To overcome this, we propose an innovative Covariance Distribution Optimization (CDO) module specifically designed for efficient, real-time applications. The CDO module aligns lane feature distributions closely with ground-truth labels, significantly enhancing detection accuracy without increasing computational complexity. Evaluations were conducted on six diverse models across all three method categories, including two optimized for real-time applications and four state-of-the-art (SOTA) models, tested comprehensively on three major datasets: CULane, TuSimple, and LLAMAS. Experimental results demonstrate accuracy improvements ranging from 0.01% to 1.5%. The proposed CDO module is characterized by ease of integration into existing systems without structural modifications and utilizes existing model parameters to facilitate ongoing training, thus offering substantial benefits in performance, power efficiency, and operational flexibility in embedded systems.

Linfeng Fan, Ziwei Li, Yuan Tian et al.

Tool-using LLM agents must act on untrusted webpages, emails, files, and API outputs while issuing privileged tool calls. Existing defenses often mediate trust at the granularity of an entire tool invocation, forcing a brittle choice in mixed-trust workflows: allow external content to influence a call and risk hijacked destinations or commands, or quarantine the call and block benign retrieval-then-act behavior. The key observation behind this paper is that indirect prompt injection becomes dangerous not when untrusted content appears in context, but when it determines an authority-bearing argument. We present \textsc{PACT} (\emph{Provenance-Aware Capability Contracts}), a runtime monitor that assigns semantic roles to tool arguments, tracks value provenance across replanning steps, and checks whether each argument's origin satisfies its role-specific trust contract. Under oracle provenance, \textsc{PACT} achieves 100\% utility and 100\% security on mixed-trust diagnostic suites, while flat invocation-level monitors incur false positives or false negatives. In full AgentDojo deployments across five models, \textsc{PACT} reaches 100\% security on the three strongest models while recovering 38.1--46.4\% utility, 8--16 percentage points above CaMeL at the same security level. Ablations show that both semantic roles and cross-step provenance are necessary. \textsc{PACT} reframes agent security as authority binding, and isolates the remaining deployment bottleneck to provenance inference and contract synthesis.

Zhuoyuan Yao, Di Wu, Xiong Wang et al.

In this paper, we propose an open source, production first, and production ready speech recognition toolkit called WeNet in which a new two-pass approach is implemented to unify streaming and non-streaming end-to-end (E2E) speech recognition in a single model. The main motivation of WeNet is to close the gap between the research and the production of E2E speechrecognition models. WeNet provides an efficient way to ship ASR applications in several real-world scenarios, which is the main difference and advantage to other open source E2E speech recognition toolkits. In our toolkit, a new two-pass method is implemented. Our method propose a dynamic chunk-based attention strategy of the the transformer layers to allow arbitrary right context length modifies in hybrid CTC/attention architecture. The inference latency could be easily controlled by only changing the chunk size. The CTC hypotheses are then rescored by the attention decoder to get the final result. Our experiments on the AISHELL-1 dataset using WeNet show that, our model achieves 5.03\% relative character error rate (CER) reduction in non-streaming ASR compared to a standard non-streaming transformer. After model quantification, our model perform reasonable RTF and latency.

Yihui Fu, Zhuoyuan Yao, Weipeng He et al.

The IEEE Spoken Language Technology Workshop (SLT) 2021 Alpha-mini Speech Challenge (ASC) is intended to improve research on keyword spotting (KWS) and sound source location (SSL) on humanoid robots. Many publications report significant improvements in deep learning based KWS and SSL on open source datasets in recent years. For deep learning model training, it is necessary to expand the data coverage to improve the robustness of model. Thus, simulating multi-channel noisy and reverberant data from single-channel speech, noise, echo and room impulsive response (RIR) is widely adopted. However, this approach may generate mismatch between simulated data and recorded data in real application scenarios, especially echo data. In this challenge, we open source a sizable speech, keyword, echo and noise corpus for promoting data-driven methods, particularly deep-learning approaches on KWS and SSL. We also choose Alpha-mini, a humanoid robot produced by UBTECH equipped with a built-in four-microphone array on its head, to record development and evaluation sets under the actual Alpha-mini robot application scenario, including noise as well as echo and mechanical noise generated by the robot itself for model evaluation. Furthermore, we illustrate the rules, evaluation methods and baselines for researchers to quickly assess their achievements and optimize their models.

Xiong Wang, Yangze Li, Chaoyou Fu et al.

Rapidly developing large language models (LLMs) have brought tremendous intelligent applications. Especially, the GPT-4o's excellent duplex speech interaction ability has brought impressive experience to users. Researchers have recently proposed several multi-modal LLMs in this direction that can achieve user-agent speech-to-speech conversations. This paper proposes a novel speech-text multimodal LLM architecture called Freeze-Omni. Our main contribution is that the speech input and output modalities can be easily connected to a textual LLM while keeping the LLM's parameters frozen throughout the training process. We design a three-stage training strategy for modeling both the speech input and output, enabling Freeze-Omni to obtain speech-to-speech conversation ability using text-speech paired data (such as ASR and TTS data) and only 60,000 multi-round text Q&A data on 8 GPUs. Moreover, we can effectively ensure that the intelligence of the Freeze-Omni in the speech modality is at the same level compared with that in the text modality of its backbone LLM, while achieving low latency end-to-end spoken response. In addition, we also designed a method to achieve duplex dialogue ability through multi-task training, giving Freeze-Omni a more natural style of dialogue ability between users and agents. In summary, Freeze-Omni holds great potential to conduct speech-to-speech dialogue based on a multimodal LLM under the condition of a frozen LLM, avoiding the catastrophic forgetting problem caused by limited data and training resources.

Heting Gao, Hang Shao, Xiong Wang et al.

The film Her features Samantha, a sophisticated AI audio agent who is capable of understanding both linguistic and paralinguistic information in human speech and delivering real-time responses that are natural, informative and sensitive to emotional subtleties. Moving one step toward more sophisticated audio agent from recent advancement in end-to-end (E2E) speech systems, we propose LUCY, a E2E speech model that (1) senses and responds to user's emotion, (2) deliver responses in a succinct and natural style, and (3) use external tool to answer real-time inquiries. Experiment results show that LUCY is better at emotion control than peer models, generating emotional responses based on linguistic emotional instructions and responding to paralinguistic emotional cues. Lucy is also able to generate responses in a more natural style, as judged by external language models, without sacrificing much performance on general question answering. Finally, LUCY can leverage function calls to answer questions that are out of its knowledge scope.

Dingdong Wang, Jin Xu, Ruihang Chu et al.

Recent advancements in speech large language models (SpeechLLMs) have attracted considerable attention. Nonetheless, current methods exhibit suboptimal performance in adhering to speech instructions. Notably, the intelligence of models significantly diminishes when processing speech-form input as compared to direct text-form input. Prior work has attempted to mitigate this semantic inconsistency between speech and text representations through techniques such as representation and behavior alignment, which involve the meticulous design of data pairs during the post-training phase. In this paper, we introduce a simple and scalable training method called InSerter, which stands for Interleaved Speech-Text Representation Pre-training. InSerter is designed to pre-train large-scale unsupervised speech-text sequences, where the speech is synthesized from randomly selected segments of an extensive text corpus using text-to-speech conversion. Consequently, the model acquires the ability to generate textual continuations corresponding to the provided speech segments, obviating the need for intensive data design endeavors. To systematically evaluate speech instruction-following capabilities, we introduce SpeechInstructBench, the first comprehensive benchmark specifically designed for speech-oriented instruction-following tasks. Our proposed InSerter achieves SOTA performance in SpeechInstructBench and demonstrates superior or competitive results across diverse speech processing tasks.

He Wang, Linhan Ma, Dake Guo et al.

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.

Quanjun Lang, Xiong Wang, Fei Lu et al.

Modeling multi-agent systems on networks is a fundamental challenge in a wide variety of disciplines. We jointly infer the weight matrix of the network and the interaction kernel, which determine respectively which agents interact with which others and the rules of such interactions from data consisting of multiple trajectories. The estimator we propose leads naturally to a non-convex optimization problem, and we investigate two approaches for its solution: one is based on the alternating least squares (ALS) algorithm; another is based on a new algorithm named operator regression with alternating least squares (ORALS). Both algorithms are scalable to large ensembles of data trajectories. We establish coercivity conditions guaranteeing identifiability and well-posedness. The ALS algorithm appears statistically efficient and robust even in the small data regime but lacks performance and convergence guarantees. The ORALS estimator is consistent and asymptotically normal under a coercivity condition. We conduct several numerical experiments ranging from Kuramoto particle systems on networks to opinion dynamics in leader-follower models.

Fei Wang, Yutong Zhang, Xiong Wang

Learning interpretable multimodal representations inherently relies on uncovering the conditional dependencies between heterogeneous features. However, sparse graph estimation techniques, such as Graphical Lasso (GLasso), to visual-linguistic domains is severely bottlenecked by high-dimensional noise, modality misalignment, and the confounding of shared versus category-specific topologies. In this paper, we propose Cross-Modal Graphical Lasso (CM-GLasso) that overcomes these fundamental limitations. By coupling a novel text-visualization strategy with a unified vision-language encoder, we strictly align multimodal features into a shared latent space. We introduce a cross-attention distillation mechanism that condenses high-dimensional patches into explicit semantic nodes, naturally extracting spatial-aware cross-modal priors. Furthermore, we unify tailored GLasso estimation and Common-Specific Structure Learning (CSSL) into a joint objective optimized via the Alternating Direction Method of Multiplier (ADMM). This formulation guarantees the simultaneous disentanglement of invariant and class-specific precision matrices without multi-step error accumulation. Extensive experiments across eight benchmarks covering both natural and medical domains demonstrate that CM-GLasso establishes a new state-of-the-art in generative classification and dense semantic segmentation tasks.

Quanjun Lang, Xiong Wang, Fei Lu et al.

We propose a framework for the joint inference of network topology, multi-type interaction kernels, and latent type assignments in heterogeneous interacting particle systems from multi-trajectory data. This learning task is a challenging non-convex mixed-integer optimization problem, which we address through a novel three-stage approach. First, we leverage shared structure across agent interactions to recover a low-rank embedding of the system parameters via matrix sensing. Second, we identify discrete interaction types by clustering within the learned embedding. Third, we recover the network weight matrix and kernel coefficients through matrix factorization and a post-processing refinement. We provide theoretical guarantees with estimation error bounds under a Restricted Isometry Property (RIP) assumption and establish conditions for the exact recovery of interaction types based on cluster separability. Numerical experiments on synthetic datasets, including heterogeneous predator-prey systems, demonstrate that our method yields an accurate reconstruction of the underlying dynamics and is robust to noise.

Shai Zucker, Xiong Wang, Fei Lu et al.

We study the convergence rate of learning pairwise interactions in single-layer attention-style models, where tokens interact through a weight matrix and a non-linear activation function. We prove that the minimax rate is $M^{-\frac{2β}{2β+1}}$ with $M$ being the sample size, depending only on the smoothness $β$ of the activation, and crucially independent of token count, ambient dimension, or rank of the weight matrix. These results highlight a fundamental dimension-free statistical efficiency of attention-style nonlocal models, even when the weight matrix and activation are not separately identifiable and provide a theoretical understanding of the attention mechanism and its training.

Ruiqi Wang, Wenjun Li, Jing Ren et al.

The deployment of large-scale LoRaWAN networks requires jointly optimizing conflicting metrics like Packet Delivery Ratio (PDR) and Energy Efficiency (EE) by dynamically allocating transmission parameters, including Carrier Frequency, Spreading Factor, and Transmission Power. Existing methods often oversimplify this challenge, focusing on a single metric or lacking the adaptability needed for dynamic channel environments, leading to suboptimal performance. To address this, we propose two online learning-based resource allocation frameworks that intelligently navigate the PDR-EE trade-off. Our foundational proposal, D-LoRa, is a fully distributed framework that models the problem as a Combinatorial Multi-Armed Bandit. By decomposing the joint parameter selection and employing specialized, disaggregated reward functions, D-LoRa dramatically reduces learning complexity and enables nodes to autonomously adapt to network dynamics. To further enhance performance in LoRaWAN networks, we introduce CD-LoRa, a hybrid framework that integrates a lightweight, centralized initialization phase to perform a one-time, quasi-optimal channel assignment and action space pruning, thereby accelerating subsequent distributed learning. Extensive simulations and real-world field experiments demonstrate the superiority of our frameworks, showing that D-LoRa excels in non-stationary environments while CD-LoRa achieves the fastest convergence in stationary conditions. In physical deployments, our methods outperform state-of-the-art baselines, improving PDR by up to 10.8% and EE by 26.1%, confirming their practical effectiveness for scalable and efficient LoRaWAN networks.

Jiaxu Zheng, Meiman He, Xuhui Tang et al.

Magnetic Resonance (MR) imaging plays an essential role in contemporary clinical diagnostics. It is increasingly integrated into advanced therapeutic workflows, such as hybrid Positron Emission Tomography/Magnetic Resonance (PET/MR) imaging and MR-only radiation therapy. These integrated approaches are critically dependent on accurate estimation of radiation attenuation, which is typically facilitated by synthesizing Computed Tomography (CT) images from MR scans to generate attenuation maps. However, existing MR-to-CT synthesis methods for whole-body imaging often suffer from poor spatial alignment between the generated CT and input MR images, and insufficient image quality for reliable use in downstream clinical tasks. In this paper, we present a novel 3D Wavelet Latent Diffusion Model (3D-WLDM) that addresses these limitations by performing modality translation in a learned latent space. By incorporating a Wavelet Residual Module into the encoder-decoder architecture, we enhance the capture and reconstruction of fine-scale features across image and latent spaces. To preserve anatomical integrity during the diffusion process, we disentangle structural and modality-specific characteristics and anchor the structural component to prevent warping. We also introduce a Dual Skip Connection Attention mechanism within the diffusion model, enabling the generation of high-resolution CT images with improved representation of bony structures and soft-tissue contrast.

Jiasheng Wu, Jingjing Zhang, Zheng Lin et al.

Recently, the rapid development of LEO satellite networks spurs another widespread concern-data processing at satellites. However, achieving efficient computation at LEO satellites in highly dynamic satellite networks is challenging and remains an open problem when considering the constrained computation capability of LEO satellites. For the first time, we propose a novel distributed learning framework named SFL-LEO by combining Federated Learning (FL) with Split Learning (SL) to accommodate the high dynamics of LEO satellite networks and the constrained computation capability of LEO satellites by leveraging the periodical orbit traveling feature. The proposed scheme allows training locally by introducing an asynchronous training strategy, i.e., achieving local update when LEO satellites disconnect with the ground station, to provide much more training space and thus increase the training performance. Meanwhile, it aggregates client-side sub-models at the ground station and then distributes them to LEO satellites by borrowing the idea from the federated learning scheme. Experiment results driven by satellite-ground bandwidth measured in Starlink demonstrate that SFL-LEO provides a similar accuracy performance with the conventional SL scheme because it can perform local training even within the disconnection duration.

Hongwei Bran Li, Fernando Navarro, Ivan Ezhov et al.

Uncertainty in medical image segmentation tasks, especially inter-rater variability, arising from differences in interpretations and annotations by various experts, presents a significant challenge in achieving consistent and reliable image segmentation. This variability not only reflects the inherent complexity and subjective nature of medical image interpretation but also directly impacts the development and evaluation of automated segmentation algorithms. Accurately modeling and quantifying this variability is essential for enhancing the robustness and clinical applicability of these algorithms. We report the set-up and summarize the benchmark results of the Quantification of Uncertainties in Biomedical Image Quantification Challenge (QUBIQ), which was organized in conjunction with International Conferences on Medical Image Computing and Computer-Assisted Intervention (MICCAI) 2020 and 2021. The challenge focuses on the uncertainty quantification of medical image segmentation which considers the omnipresence of inter-rater variability in imaging datasets. The large collection of images with multi-rater annotations features various modalities such as MRI and CT; various organs such as the brain, prostate, kidney, and pancreas; and different image dimensions 2D-vs-3D. A total of 24 teams submitted different solutions to the problem, combining various baseline models, Bayesian neural networks, and ensemble model techniques. The obtained results indicate the importance of the ensemble models, as well as the need for further research to develop efficient 3D methods for uncertainty quantification methods in 3D segmentation tasks.

Shaohua Pan, Qi Ma, Xinyu Yi et al.

Either RGB images or inertial signals have been used for the task of motion capture (mocap), but combining them together is a new and interesting topic. We believe that the combination is complementary and able to solve the inherent difficulties of using one modality input, including occlusions, extreme lighting/texture, and out-of-view for visual mocap and global drifts for inertial mocap. To this end, we propose a method that fuses monocular images and sparse IMUs for real-time human motion capture. Our method contains a dual coordinate strategy to fully explore the IMU signals with different goals in motion capture. To be specific, besides one branch transforming the IMU signals to the camera coordinate system to combine with the image information, there is another branch to learn from the IMU signals in the body root coordinate system to better estimate body poses. Furthermore, a hidden state feedback mechanism is proposed for both two branches to compensate for their own drawbacks in extreme input cases. Thus our method can easily switch between the two kinds of signals or combine them in different cases to achieve a robust mocap. %The two divided parts can help each other for better mocap results under different conditions. Quantitative and qualitative results demonstrate that by delicately designing the fusion method, our technique significantly outperforms the state-of-the-art vision, IMU, and combined methods on both global orientation and local pose estimation. Our codes are available for research at https://shaohua-pan.github.io/robustcap-page/.

Xiong Wang, Jiancheng Ye, John C. S. Lui

Mobile edge computing facilitates users to offload computation tasks to edge servers for meeting their stringent delay requirements. Previous works mainly explore task offloading when system-side information is given (e.g., server processing speed, cellular data rate), or centralized offloading under system uncertainty. But both generally fall short to handle task placement involving many coexisting users in a dynamic and uncertain environment. In this paper, we develop a multi-user offloading framework considering unknown yet stochastic system-side information to enable a decentralized user-initiated service placement. Specifically, we formulate the dynamic task placement as an online multi-user multi-armed bandit process, and propose a decentralized epoch based offloading (DEBO) to optimize user rewards which are subjected under network delay. We show that DEBO can deduce the optimal user-server assignment, thereby achieving a close-to-optimal service performance and tight O(log T) offloading regret. Moreover, we generalize DEBO to various common scenarios such as unknown reward gap, dynamic entering or leaving of clients, and fair reward distribution, while further exploring when users' offloaded tasks require heterogeneous computing resources. Particularly, we accomplish a sub-linear regret for each of these instances. Real measurements based evaluations corroborate the superiority of our offloading schemes over state-of-the-art approaches in optimizing delay-sensitive rewards.

Xiong Wang, Sining Sun, Lei Xie et al.

End-to-end models are favored in automatic speech recognition (ASR) because of their simplified system structure and superior performance. Among these models, Transformer and Conformer have achieved state-of-the-art recognition accuracy in which self-attention plays a vital role in capturing important global information. However, the time and memory complexity of self-attention increases squarely with the length of the sentence. In this paper, a prob-sparse self-attention mechanism is introduced into Conformer to sparse the computing process of self-attention in order to accelerate inference speed and reduce space consumption. Specifically, we adopt a Kullback-Leibler divergence based sparsity measurement for each query to decide whether we compute the attention function on this query. By using the prob-sparse attention mechanism, we achieve impressively 8% to 45% inference speed-up and 15% to 45% memory usage reduction of the self-attention module of Conformer Transducer while maintaining the same level of error rate.

Xiong Wang, Riheng Jia

Mean field game facilitates analyzing multi-armed bandit (MAB) for a large number of agents by approximating their interactions with an average effect. Existing mean field models for multi-agent MAB mostly assume a binary reward function, which leads to tractable analysis but is usually not applicable in practical scenarios. In this paper, we study the mean field bandit game with a continuous reward function. Specifically, we focus on deriving the existence and uniqueness of mean field equilibrium (MFE), thereby guaranteeing the asymptotic stability of the multi-agent system. To accommodate the continuous reward function, we encode the learned reward into an agent state, which is in turn mapped to its stochastic arm playing policy and updated using realized observations. We show that the state evolution is upper semi-continuous, based on which the existence of MFE is obtained. As the Markov analysis is mainly for the case of discrete state, we transform the stochastic continuous state evolution into a deterministic ordinary differential equation (ODE). On this basis, we can characterize a contraction mapping for the ODE to ensure a unique MFE for the bandit game. Extensive evaluations validate our MFE characterization, and exhibit tight empirical regret of the MAB problem.

Binbin Zhang, Di Wu, Zhuoyuan Yao et al.

In this paper, we present a novel two-pass approach to unify streaming and non-streaming end-to-end (E2E) speech recognition in a single model. Our model adopts the hybrid CTC/attention architecture, in which the conformer layers in the encoder are modified. We propose a dynamic chunk-based attention strategy to allow arbitrary right context length. At inference time, the CTC decoder generates n-best hypotheses in a streaming way. The inference latency could be easily controlled by only changing the chunk size. The CTC hypotheses are then rescored by the attention decoder to get the final result. This efficient rescoring process causes very little sentence-level latency. Our experiments on the open 170-hour AISHELL-1 dataset show that, the proposed method can unify the streaming and non-streaming model simply and efficiently. On the AISHELL-1 test set, our unified model achieves 5.60% relative character error rate (CER) reduction in non-streaming ASR compared to a standard non-streaming transformer. The same model achieves 5.42% CER with 640ms latency in a streaming ASR system.

Zhichao Wang, Wenshuo Ge, Xiong Wang et al.

This paper proposes an interesting voice and accent joint conversion approach, which can convert an arbitrary source speaker's voice to a target speaker with non-native accent. This problem is challenging as each target speaker only has training data in native accent and we need to disentangle accent and speaker information in the conversion model training and re-combine them in the conversion stage. In our recognition-synthesis conversion framework, we manage to solve this problem by two proposed tricks. First, we use accent-dependent speech recognizers to obtain bottleneck features for different accented speakers. This aims to wipe out other factors beyond the linguistic information in the BN features for conversion model training. Second, we propose to use adversarial training to better disentangle the speaker and accent information in our encoder-decoder based conversion model. Specifically, we plug an auxiliary speaker classifier to the encoder, trained with an adversarial loss to wipe out speaker information from the encoder output. Experiments show that our approach is superior to the baseline. The proposed tricks are quite effective in improving accentedness and audio quality and speaker similarity are well maintained.

Xiong Wang, Zhuoyuan Yao, Xian Shi et al.

End-to-end models are favored in automatic speech recognition (ASR) because of its simplified system structure and superior performance. Among these models, recurrent neural network transducer (RNN-T) has achieved significant progress in streaming on-device speech recognition because of its high-accuracy and low-latency. RNN-T adopts a prediction network to enhance language information, but its language modeling ability is limited because it still needs paired speech-text data to train. Further strengthening the language modeling ability through extra text data, such as shallow fusion with an external language model, only brings a small performance gain. In view of the fact that Mandarin Chinese is a character-based language and each character is pronounced as a tonal syllable, this paper proposes a novel cascade RNN-T approach to improve the language modeling ability of RNN-T. Our approach firstly uses an RNN-T to transform acoustic feature into syllable sequence, and then converts the syllable sequence into character sequence through an RNN-T-based syllable-to-character converter. Thus a rich text repository can be easily used to strengthen the language model ability. By introducing several important tricks, the cascade RNN-T approach surpasses the character-based RNN-T by a large margin on several Mandarin test sets, with much higher recognition quality and similar latency.

Fan Yu, Zhuoyuan Yao, Xiong Wang et al.

Automatic speech recognition (ASR) has been significantly advanced with the use of deep learning and big data. However improving robustness, including achieving equally good performance on diverse speakers and accents, is still a challenging problem. In particular, the performance of children speech recognition (CSR) still lags behind due to 1) the speech and language characteristics of children's voice are substantially different from those of adults and 2) sizable open dataset for children speech is still not available in the research community. To address these problems, we launch the Children Speech Recognition Challenge (CSRC), as a flagship satellite event of IEEE SLT 2021 workshop. The challenge will release about 400 hours of Mandarin speech data for registered teams and set up two challenge tracks and provide a common testbed to benchmark the CSR performance. In this paper, we introduce the datasets, rules, evaluation method as well as baselines.