arXiv subjects

Haoyang Zhang

Publications and source records attributed to Haoyang Zhang.

At least 55 records · Page 3Linked to original sources

arXiv2025

Step-Audio-R1 Technical Report

Recent advances in reasoning models have demonstrated remarkable success in text and vision domains through extended chain-of-thought deliberation. However, a perplexing phenomenon persists in audio language models: they consistently perform better with minimal or no reasoning, raising a fundamental question - can audio intelligence truly benefit from deliberate thinking? We introduce Step-Audio-R1, the first audio reasoning model that successfully unlocks reasoning capabilities in the audio domain. Through our proposed Modality-Grounded Reasoning Distillation (MGRD) framework, Step-Audio-R1 learns to generate audio-relevant reasoning chains that genuinely ground themselves in acoustic features rather than hallucinating disconnected deliberations. Our model exhibits strong audio reasoning capabilities, surpassing Gemini 2.5 Pro and achieving performance comparable to the state-of-the-art Gemini 3 Pro across comprehensive audio understanding and reasoning benchmarks spanning speech, environmental sounds, and music. These results demonstrate that reasoning is a transferable capability across modalities when appropriately anchored, transforming extended deliberation from a liability into a powerful asset for audio intelligence. By establishing the first successful audio reasoning model, Step-Audio-R1 opens new pathways toward building truly multimodal reasoning systems that think deeply across all sensory modalities.

cs.AI↗

arXiv2025

Aligning Speech to Languages to Enhance Code-switching Speech Recognition

Code-switching (CS) refers to the switching of languages within a speech signal and results in language confusion for automatic speech recognition (ASR). To address language confusion, we propose a language alignment loss (LAL) that aligns acoustic features to pseudo-language labels learned from the ASR decoder during ASR training. This approach enables frame-level language identification without the need for frame-level language annotations. To further tackle the complex token alternatives for language modeling in bilingual scenarios, we propose to employ large language models via a generative error correction method. A linguistic hint, derived from LAL outputs and decoded hypotheses, is introduced to guide the prompting and enhance the LLM-based generative error correction for CS-ASR. The proposed methods are evaluated on the SEAME dataset and data from the ASRU 2019 Mandarin-English code-switching speech recognition challenge. The incorporation of the proposed language alignment loss improves CS-ASR performance for both hybrid CTC/attention and Whisper models on both datasets, with only a negligible increase in the number of parameters. This work also highlights the efficacy of language alignment loss in balancing primary-language-dominant bilingual data during training, with an 8.6% relative improvement on the ASRU dataset compared to the baseline model. Performance evaluation using large language models reveals the advantage of the linguistic hint by achieving 14.1% and 5.5% relative improvement on test sets of the ASRU and SEAME datasets, respectively.

eess.AS↗

arXiv2025

MULTI-Bench: A Multi-Turn Interactive Benchmark for Assessing Emotional Intelligence ability of Spoken Dialogue Models

Spoken Dialogue Models (SDMs) have advanced rapidly, yet their ability to sustain genuinely interactive multi-turn conversations remains underexplored, as most benchmarks focus on single-turn exchanges. We introduce Multi-Bench, the first benchmark explicitly designed to evaluate SDMs in multi-turn interactive dialogue with an emphasis on emotional intelligence. Multi-Bench employs a hierarchical structure with a basic track for emotion understanding and reasoning and an advanced track for emotion support and application. It comprises five carefully designed tasks and about 3.2K samples, ranging from emotion recognition to complex reasoning and interactive dialogue, supported by a reproducible evaluation framework. We evaluate six representative SDMs on eight subsets of Multi-Bench. Results show that while current SDMs achieve good performance on basic understanding tasks, they still have room for improvement in advanced multi-turn interactive dialogue and reasoning-related tasks, particularly in emotion awareness and application.

eess.AS↗

arXiv2025

MCIHN: A Hybrid Network Model Based on Multi-path Cross-modal Interaction for Multimodal Emotion Recognition

Multimodal emotion recognition is crucial for future human-computer interaction. However, accurate emotion recognition still faces significant challenges due to differences between different modalities and the difficulty of characterizing unimodal emotional information. To solve these problems, a hybrid network model based on multipath cross-modal interaction (MCIHN) is proposed. First, adversarial autoencoders (AAE) are constructed separately for each modality. The AAE learns discriminative emotion features and reconstructs the features through a decoder to obtain more discriminative information about the emotion classes. Then, the latent codes from the AAE of different modalities are fed into a predefined Cross-modal Gate Mechanism model (CGMM) to reduce the discrepancy between modalities, establish the emotional relationship between interacting modalities, and generate the interaction features between different modalities. Multimodal fusion using the Feature Fusion module (FFM) for better emotion recognition. Experiments were conducted on publicly available SIMS and MOSI datasets, demonstrating that MCIHN achieves superior performance.

cs.CV↗

arXiv2025

Detection of Quasi-periodic Oscillations in the $γ$-Ray Light Curve of 4FGL J0309.9-6058

In this work, we report, for the first time, a quasi-periodic oscillation (QPO) in the $γ$-ray band of 4FGL J0309.9-6058, also known as PKS 0308-611. We employed three analytical methods (the Lomb-Scargle periodogram, REDFIT, and the weighted wavelet Z-transform) to analyze the QPO signal using \textit{Fermi} $γ$-ray light curve data. The analysis reveals a potential QPO during MJD 57983$-$60503, with a period of approximately 550 days and a maximum local significance of 3.72$σ$ and global significance of 2.72$σ$ derived from the WWZ analysis. To validate this result, we applied Gaussian Process (GP) to the same light curve, which independently confirms the presence of QPO signal consistent with our Fourier-based results. We further extended the analysis to the full duration of the \textit{Fermi} observations, and the results consistently support and strengthen the presence of this QPO signal. Additionally, a time lag between the optical and $γ$-ray bands indicates separate emission regions for these two bands. Given the year-like timescale of the QPO signal and the fact that a QPO signal with local significance over 3$σ$ for full \textit{Fermi}-LAT observed time, we suggest that the QPO is most likely caused by a precessing jet.

astro-ph.HE↗

arXiv2025

Real-time Measurement-based Optimization for Distribution System Operation Considering Battery Voltage and Thermal Constraints

The secure operation of power distribution systems is challenged by the growing integration of distributed energy resources. Leveraging the flexibility of battery storage offers a cost-effective alternative to measures like generation curtailment, which results in energy losses. However, developing an effective operational model for battery storage is hindered by inaccurate grid models, unavailability of load data, nonlinear relationship between power injections and network states, intertemporal constraints, and complex electrochemical and thermal dynamics. To address these challenges, this paper proposes a data-driven operational control scheme for battery storage in distribution systems. Linear and convex quadratic operational constraints are constructed based on real-time distribution system and battery storage measurements. Lyapunov optimization decouples multi-period battery operation, enabling a real-time, forecast-free control strategy with low computational complexity. Numerical studies using nonlinear distribution system and battery storage simulators validate the effectiveness of the approach in ensuring secure distribution system operation and satisfaction of voltage and thermal constraints of battery storage.

eess.SY↗

arXiv2025

User to Video: A Model for Spammer Detection Inspired by Video Classification Technology

This article is inspired by video classification technology. If the user behavior subspace is viewed as a frame image, consecutive frame images are viewed as a video. Following this novel idea, a model for spammer detection based on user videoization, called UVSD, is proposed. Firstly, a user2piexl algorithm for user pixelization is proposed. Considering the adversarial behavior of user stances, the user is viewed as a pixel, and the stance is quantified as the pixel's RGB. Secondly, a behavior2image algorithm is proposed for transforming user behavior subspace into frame images. Low-rank dense vectorization of subspace user relations is performed using representation learning, while cutting and diffusion algorithms are introduced to complete the frame imageization. Finally, user behavior videos are constructed based on temporal features. Subsequently, a video classification algorithm is combined to identify the spammers. Experiments using publicly available datasets, i.e., WEIBO and TWITTER, show an advantage of the UVSD model over state-of-the-art methods.

cs.CV↗

arXiv2025

GameGPT: Multi-agent Collaborative Framework for Game Development

The large language model (LLM) based agents have demonstrated their capacity to automate and expedite software development processes. In this paper, we focus on game development and propose a multi-agent collaborative framework, dubbed GameGPT, to automate game development. While many studies have pinpointed hallucination as a primary roadblock for deploying LLMs in production, we identify another concern: redundancy. Our framework presents a series of methods to mitigate both concerns. These methods include dual collaboration and layered approaches with several in-house lexicons, to mitigate the hallucination and redundancy in the planning, task identification, and implementation phases. Furthermore, a decoupling approach is also introduced to achieve code generation with better precision.

cs.AI↗

arXiv2025

Step-Audio 2 Technical Report

This paper presents Step-Audio 2, an end-to-end multi-modal large language model designed for industry-strength audio understanding and speech conversation. By integrating a latent audio encoder and reasoning-centric reinforcement learning (RL), Step-Audio 2 achieves promising performance in automatic speech recognition (ASR) and audio understanding. To facilitate genuine end-to-end speech conversation, Step-Audio 2 incorporates the generation of discrete audio tokens into language modeling, significantly enhancing its responsiveness to paralinguistic information such as speaking styles and emotions. To effectively leverage the rich textual and acoustic knowledge in real-world data, Step-Audio 2 integrates retrieval-augmented generation (RAG) and is able to call external tools such as web search to mitigate hallucination and audio search to switch timbres. Trained on millions of hours of speech and audio data, Step-Audio 2 delivers intelligence and expressiveness across diverse conversational scenarios. Evaluation results demonstrate that Step-Audio 2 achieves state-of-the-art performance on various audio understanding and conversational benchmarks compared to other open-source and commercial solutions. Please visit https://github.com/stepfun-ai/Step-Audio2 for more information.

cs.CL↗

arXiv2025

Arbitrage Tactics in the Local Markets via Hierarchical Multi-agent Reinforcement Learning

Strategic bidding tactics employed by prosumers in local markets, including the Local Electricity Market (LEM) and Local Flexibility Market (LFM), have attracted significant attention due to their potential to enhance economic benefits for market participants through optimized energy management and bidding. While existing research has explored strategic bidding in a single market with multi-agent reinforcement learning (MARL) algorithms, arbitrage opportunities across local markets remain unexplored. This paper introduces a hierarchical MARL (HMARL) algorithm designed to enable aggregator arbitrage across multiple local markets. The strategic behavior of these aggregators in local markets is modeled as a two-stage Markov game: the first stage involves the LEM, while the second stage encompasses both the LFM and the balancing market. To solve this two-stage Markov game, the HMARL framework assigns two sub-agents to each aggregator, a primary sub-agent and a secondary sub-agent. Without the arbitrage strategy, these sub-agents operate in silos, with the primary sub-agent focusing on first-stage profits and the secondary sub-agent on second-stage profits, each employing independent MARLs. On the contrary, when implementing the arbitrage strategy with the proposed HMARL, the sub-agents communicate and coordinate to perform arbitrage across multiple local markets, enhancing overall efficiency. The case study, conducted under a scenario where all aggregators employ the arbitrage strategy, shows that despite higher initial costs in the LEM, this strategy generates substantial savings in the LFM and the balancing market, resulting in a total profit increase of $40.6\%$ on average. This highlights the capability of the proposed HMARL to address the two-stage Markov game and facilitate arbitrage across local markets, thereby enhancing profitability for participants.

eess.SY↗

arXiv2025

NTU Speechlab LLM-Based Multilingual ASR System for Interspeech MLC-SLM Challenge 2025

This report details the NTU Speechlab system developed for the Interspeech 2025 Multilingual Conversational Speech and Language Model (MLC-SLM) Challenge (Task I), where we achieved 5th place. We present comprehensive analyses of our multilingual automatic speech recognition system, highlighting key advancements in model architecture, data selection, and training strategies. In particular, language-specific prompts and model averaging techniques were instrumental in boosting system performance across diverse languages. Compared to the initial baseline system, our final model reduced the average Mix Error Rate from 20.2% to 10.6%, representing an absolute improvement of 9.6% (a relative improvement of 48%) on the evaluation set. Our results demonstrate the effectiveness of our approach and offer practical insights for future Speech Large Language Models.

cs.CL↗

arXiv2025

Impact of Frame Rates on Speech Tokenizer: A Case Study on Mandarin and English

The speech tokenizer plays a crucial role in recent speech tasks, generally serving as a bridge between speech signals and language models. While low-frame-rate codecs are widely employed as speech tokenizers, the impact of frame rates on speech tokens remains underexplored. In this study, we investigate how varying frame rates affect speech tokenization by examining Mandarin and English, two typologically distinct languages. We encode speech at different frame rates and evaluate the resulting semantic tokens in the speech recognition task. Our findings reveal that frame rate variations influence speech tokenization differently for each language, highlighting the interplay between frame rates, phonetic density, and language-specific acoustic features. The results provide insights into optimizing frame rate selection for speech tokenizers, with implications for automatic speech recognition, text-to-speech, and other speech-related applications.

cs.CL↗

arXiv2025

Cost-Efficient LLM Training with Lifetime-Aware Tensor Offloading via GPUDirect Storage

We present the design and implementation of a new lifetime-aware tensor offloading framework for GPU memory expansion using low-cost PCIe-based solid-state drives (SSDs). Our framework, TERAIO, is developed explicitly for large language model (LLM) training with multiple GPUs and multiple SSDs. Its design is driven by our observation that the active tensors take only a small fraction (1.7% on average) of allocated GPU memory in each LLM training iteration, the inactive tensors are usually large and will not be used for a long period of time, creating ample opportunities for offloading/prefetching tensors to/from slow SSDs without stalling the GPU training process. TERAIO accurately estimates the lifetime (active period of time in GPU memory) of each tensor with the profiling of the first few iterations in the training process. With the tensor lifetime analysis, TERAIO will generate an optimized tensor offloading/prefetching plan and integrate it into the compiled LLM program via PyTorch. TERAIO has a runtime tensor migration engine to execute the offloading/prefetching plan via GPUDirect storage, which allows direct tensor migration between GPUs and SSDs for alleviating the CPU bottleneck and maximizing the SSD bandwidth utilization. In comparison with state-of-the-art studies such as ZeRO-Offload and ZeRO-Infinity, we show that TERAIO improves the training performance of various LLMs by 1.47x on average, and achieves 80.7% of the ideal performance assuming unlimited GPU memory.

cs.DC↗

arXiv2025

Generating Multimodal Driving Scenes via Next-Scene Prediction

Generative models in Autonomous Driving (AD) enable diverse scene creation, yet existing methods fall short by only capturing a limited range of modalities, restricting the capability of generating controllable scenes for comprehensive evaluation of AD systems. In this paper, we introduce a multimodal generation framework that incorporates four major data modalities, including a novel addition of map modality. With tokenized modalities, our scene sequence generation framework autoregressively predicts each scene while managing computational demands through a two-stage approach. The Temporal AutoRegressive (TAR) component captures inter-frame dynamics for each modality while the Ordered AutoRegressive (OAR) component aligns modalities within each scene by sequentially predicting tokens in a fixed order. To maintain coherence between map and ego-action modalities, we introduce the Action-aware Map Alignment (AMA) module, which applies a transformation based on the ego-action to maintain coherence between these modalities. Our framework effectively generates complex, realistic driving scenes over extended sequences, ensuring multimodal consistency and offering fine-grained control over scene elements. Project page: https://yanhaowu.github.io/UMGen/

cs.CV↗

arXiv2025

Revisiting the observation-theory confrontation in high-frequency QPOs:a new QPO in NGC 5506, intermediate mass black holes, and the crucial role of accretion state

The scale invariance of accretion processes (SIAP) is crucial for understanding the physical processes of black hole accretion systems at different scales. When applying this rule to high-frequency quasi-periodic oscillations (HFQPOs), there is an observation-theory confrontation in active galactic nuclei (AGNs). By compiling an updated X-ray HFQPO catalog, we found that the ultraluminous X-ray sources support the HFQPO models, similar to black hole X-ray binaries. More importantly, we identified two supermassive black hole (SMBH) sources (Sgr A and NGC 1365) with possible advection-dominated accretion flow (ADAF) configurations that support existing HFQPO models, even though many AGNs still do not. Furthermore, we report a new HFQPO candidate in NGC 5506. This source exhibits an accretion state similar to that of Sgr A* and NGC 1365, and it also supports the HFQPO models. Our results are consistent with previous numerical simulations and suggest that the accretion state of HFQPOs in SMBHs may differ from that of stellar-mass black holes (SBHs). To reconcile the sources that do not support the models, either a global general-relativistic HFQPO model based on magnetohydrodynamics needs to be considered, or the HFQPOs in these sources may originate from entirely different physical processes. This discovery significantly extends the SIAP rule to a broader scale, confirming that the paradigm of accretion scale invariance remains consistent from SBHs to SMBHs.

astro-ph.HE↗

arXiv2025

Step-Audio: Unified Understanding and Generation in Intelligent Speech Interaction

Real-time speech interaction, serving as a fundamental interface for human-machine collaboration, holds immense potential. However, current open-source models face limitations such as high costs in voice data collection, weakness in dynamic control, and limited intelligence. To address these challenges, this paper introduces Step-Audio, the first production-ready open-source solution. Key contributions include: 1) a 130B-parameter unified speech-text multi-modal model that achieves unified understanding and generation, with the Step-Audio-Chat version open-sourced; 2) a generative speech data engine that establishes an affordable voice cloning framework and produces the open-sourced lightweight Step-Audio-TTS-3B model through distillation; 3) an instruction-driven fine control system enabling dynamic adjustments across dialects, emotions, singing, and RAP; 4) an enhanced cognitive architecture augmented with tool calling and role-playing abilities to manage complex tasks effectively. Based on our new StepEval-Audio-360 evaluation benchmark, Step-Audio achieves state-of-the-art performance in human evaluations, especially in terms of instruction following. On open-source benchmarks like LLaMA Question, shows 9.3% average performance improvement, demonstrating our commitment to advancing the development of open-source multi-modal language technologies. Our code and models are available at https://github.com/stepfun-ai/Step-Audio.

cs.CL↗

arXiv2025

SkyByte: Architecting an Efficient Memory-Semantic CXL-based SSD with OS and Hardware Co-design

The CXL-based solid-state drive (CXL-SSD) provides a promising approach towards scaling the main memory capacity at low cost. However, the CXL-SSD faces performance challenges due to the long flash access latency and unpredictable events such as garbage collection in the SSD device, stalling the host processor and wasting compute cycles. Although the CXL interface enables the byte-granular data access to the SSD, accessing flash chips is still at page granularity due to physical limitations. The mismatch of access granularity causes significant unnecessary I/O traffic to flash chips, worsening the suboptimal end-to-end data access performance. In this paper, we present SkyByte, an efficient CXL-based SSD that employs a holistic approach to address the aforementioned challenges by co-designing the host operating system (OS) and SSD controller. To alleviate the long memory stall when accessing the CXL-SSD, SkyByte revisits the OS context switch mechanism and enables opportunistic context switches upon the detection of long access delays. To accommodate byte-granular data accesses, SkyByte architects the internal DRAM of the SSD controller into a cacheline-level write log and a page-level data cache, and enables data coalescing upon log cleaning to reduce the I/O traffic to flash chips. SkyByte also employs optimization techniques that include adaptive page migration for exploring the performance benefits of fast host memory by promoting hot pages in CXL-SSD to the host. We implement SkyByte with a CXL-SSD simulator and evaluate its efficiency with various data-intensive applications. Our experiments show that SkyByte outperforms current CXL-based SSD by 6.11X, and reduces the I/O traffic to flash chips by 23.08X on average. SkyByte also reaches 75% of the performance of the ideal case that assumes unlimited DRAM capacity in the host, which offers an attractive cost-effective solution.

cs.AR↗

arXiv2024

Discovering the Mass-Scaled Damping Timescale from Microquasars to Blazars

Studying the variability of the accretion disks of black holes and jets is important to identify their internal physical processes. In this letter, we obtain the characteristic damping timescale of 34 blazars and seven microquasars from the Fermi-Large Area Telescope and the XMM-Newton X-ray telescope, respectively. We found that the mass-scaled characteristic timescales, ranging from the microquasars of stellar-mass black holes to the blazars of supermassive black holes, exhibited a linear relationship with a slope of $\sim$0.57. Given the fact the damping timescales of the $γ$-ray in the blazars are associated with the jet, we propose that the timescales of the X-ray in these microquasars are also related with the jet. The mass-scaled damping timescale that we found was consistent with the radiation of the optical accretion disk. This can be attributed to the viscous timescale at the ultraviolet-emitting radii of the disk, which can affect the jet. Our study provides a new perspective on the origin of the region of radiation and the possible disk--jet connection based on time-domain analysis.

astro-ph.HE↗