arXiv subjects

Heng Lu

Publications and source records attributed to Heng Lu.

At least 19 recordsLinked to original sources

arXiv2026

REAL-Q: E2E LLM Quantization via Dynamic Gradient Descent

Post-training quantization (PTQ) is essential for deploying large language models (LLMs) under strict resource constraints. State-of-the-art PTQ methods quantize each layer with a single closed-form second-order solver: to remain analytically tractable, they heavily approximate the global loss (dropping cross-channel coupling, pooling output rows into groups), and they then freeze the resulting Hessian across the entire layer, with no way to refresh it as the loss landscape shifts column by column--a phenomenon we call information misalignment. We propose REAL-Q (Real-time E2E-loss Aligned LLM Quantization), a novel PTQ paradigm that breaks this compromise: instead of diluting the objective for the sake of analytic tractability, REAL-Q targets an end-to-end-aligned surrogate of the global loss and refines it via fine-grained, dynamic Block-wise Gradient Descent applied after every column block (128 columns). By coupling this fine-grained correction with a sliding window mechanism for smooth cross-layer transitions, REAL-Q effectively mitigates error propagation across the network. On LLaMA-3.1 (8B and 70B) and Qwen3 (0.6B-32B) at W4A16, REAL-Q reduces end-to-end KL divergence by up to ~49% relative to state-of-the-art globally-guided methods.

cs.LG↗

arXiv2026

Preserving Speech-to-Text LLM Capabilities in Speech-to-Speech Generation

Strong speech-to-text (S2T) LLMs already provide robust speech perception and text reasoning, but adding speech-to-speech (S2S) output is challenging: fine-tuning the backbone can degrade the original S2T performance, while attaching a downstream talker reintroduces a serial text-to-speech bottleneck. We present PRIME-Speech, a frozen-backbone S2S conversion framework that trains only speech-generation modules. PRIME-Speech synchronizes a causal audio post-decoder with intermediate hidden states of the frozen backbone, so codec tokens are generated from the model's evolving reasoning trajectory rather than from completed text chunks. The post-decoder uses mixed hidden-state, text, and audio-history conditioning, and a training-time packing strategy with turn-level audio KV-cache and position reset stabilizes multi-turn spoken interaction without additional multi-turn S2S training data. Multi-token prediction further reduces the effective codec prediction rate and improves first-audio latency without modifying the reasoning path. Across speech translation, spoken QA, speech understanding, and multi-turn dialogue, PRIME-Speech preserves the S2T behavior of the frozen backbone while producing accurate, low-WER spoken responses.

eess.AS↗

arXiv2026

Closing the Modality Reasoning Gap for Speech Large Language Models

Although Speech Large Language Models have achieved notable progress, a substantial modality reasoning gap remains: their reasoning performance on speech inputs is markedly weaker than on text. This gap could be associated with representational drift across Transformer layers and behavior deviations in long-chain reasoning. To address this issue, we introduce TARS, a reinforcement-learning framework that aligns text-conditioned and speech-conditioned trajectories through an asymmetric reward design. The framework employs two dense and complementary signals: representation alignment, which measures layer-wise hidden-state similarity between speech- and text-conditioned trajectories, and behavior alignment, which evaluates semantic consistency between generated outputs and reference text completions. Experiments on challenging reasoning benchmarks, including MMSU and OBQA, show that our approach significantly narrows the modality reasoning gap and achieves state-of-the-art performance among 7B-scale Speech LLMs.

cs.CL↗

arXiv2025

FlexiCodec: A Dynamic Neural Audio Codec for Low Frame Rates

Neural audio codecs are foundational to speech language models. It is expected to have a low frame rate and decoupled semantic and acoustic information. A lower frame rate codec can reduce the computational cost of speech language models by shortening the sequence length. Recent studies have developed 12.5Hz low-frame-rate audio codecs, but even lower frame rate codecs remain underexplored. We find that a major challenge for very low frame rate tokens is missing semantic information. This paper introduces FlexiCodec to address this limitation. FlexiCodec improves semantic preservation with a dynamic frame rate approach and introduces a novel architecture featuring an ASR feature-assisted dual stream encoding and Transformer bottlenecks. With dynamic frame rates, it uses less frames at information-sparse regions through adaptively merging semantically similar frames. A dynamic frame rate also allows FlexiCodec to support inference-time controllable frame rates between 3Hz and 12.5Hz. Experiments on 6.25Hz, 8.3Hz and 12.5Hz average frame rates confirm that FlexiCodec excels over baseline systems in semantic information preservation and delivers a high audio reconstruction quality. We also validate the effectiveness of FlexiCodec in language model-based TTS. Demos are available at: https://flexicodec.github.io. Code is available at: https://github.com/amphionteam/flexicodec.

cs.SD↗

arXiv2025

Long-Context Speech Synthesis with Context-Aware Memory

In long-text speech synthesis, current approaches typically convert text to speech at the sentence-level and concatenate the results to form pseudo-paragraph-level speech. These methods overlook the contextual coherence of paragraphs, leading to reduced naturalness and inconsistencies in style and timbre across the long-form speech. To address these issues, we propose a Context-Aware Memory (CAM)-based long-context Text-to-Speech (TTS) model. The CAM block integrates and retrieves both long-term memory and local context details, enabling dynamic memory updates and transfers within long paragraphs to guide sentence-level speech synthesis. Furthermore, the prefix mask enhances the in-context learning ability by enabling bidirectional attention on prefix tokens while maintaining unidirectional generation. Experimental results demonstrate that the proposed method outperforms baseline and state-of-the-art long-context methods in terms of prosody expressiveness, coherence and context inference cost across paragraph-level speech.

eess.AS↗

arXiv2025

Force sensing with a graphene nanomechanical resonator coupled to photonic crystal guided resonances

Achieving optimal force sensitivity with nanomechanical resonators requires the ability to resolve their thermal vibrations. In two-dimensional resonators, this can be done by measuring the energy they absorb while vibrating in an optical standing wave formed between a light source and a mirror. However, the responsivity of this method -- the change in optical energy per unit displacement of the resonator -- is modest, fundamentally limited by the physics of propagating plane waves. We present simulations showing that replacing the mirror with a photonic crystal supporting guided resonances increases the responsivity of graphene resonators by an order of magnitude. The steep optical energy gradients enable efficient transduction of flexural vibrations using low optical power, thereby reducing heating. Furthermore, the presence of two guided resonances at different wavelengths allows thermal vibrations to be resolved with a high signal-to-noise ratio across a wide range of membrane positions in free space. Our approach provides a simple optical method for implementing ultrasensitive force detection using a graphene nanomechanical resonator.

cond-mat.mes-hall↗

arXiv2025

SLM-S2ST: A multimodal language model for direct speech-to-speech translation

Speech-aware language models (LMs) have demonstrated capabilities in understanding spoken language while generating text-based responses. However, enabling them to produce speech output efficiently and effectively remains a challenge. In this paper, we present SLM-S2ST, a multimodal LM for direct speech-to-speech translation (S2ST), built on the open-source Phi4-MM model. SLM-S2ST extends its predecessor by generating translated speech using an audio transformer head that predicts audio tokens with a delay relative to text tokens, followed by a streaming vocoder for waveform synthesis. Our experimental results on the CVSS-C dataset demonstrate SLM-S2ST's superior performance, significantly surpassing existing baseline models trained on the same dataset. Furthermore, when we scale up the training data and the model size, SLM-S2ST reaches on-par performance with the current SOTA model.

eess.AS↗

arXiv2025

Towards Efficient Speech-Text Jointly Decoding within One Speech Language Model

Speech language models (Speech LMs) enable end-to-end speech-text modeling within a single model, offering a promising direction for spoken dialogue systems. The choice of speech-text jointly decoding paradigm plays a critical role in performance, efficiency, and alignment quality. In this work, we systematically compare representative joint speech-text decoding strategies, including the interleaved, and parallel generation paradigms, under a controlled experimental setup using the same base language model, speech tokenizer and training data. Our results show that the interleaved approach achieves the best alignment. However it suffers from slow inference due to long token sequence length. To address this, we propose a novel early-stop interleaved (ESI) pattern that not only significantly accelerates decoding but also yields slightly better performance. Additionally, we curate high-quality question answering (QA) datasets to further improve speech QA performance.

eess.AS↗

arXiv2025

Hysteretic responses of nanomechanical resonators based on crumpled few-layer graphene

Manipulating two-dimensional materials occasionally results in crumpled membranes. Their complicated morphologies feature an abundance of folds, creases and wrinkles that make each crumpled membrane unique. Here, we prepare four nanomechanical resonators based on crumpled membranes of few-layer graphene and measure their static response and the spectrum of their dynamic response. We tune both responses with a dc voltage applied between the membrane and an underlying gate electrode. Surprisingly, we find that all four resonators exhibit hysteretic responses as the gate voltage is increased and then decreased. Concomitant discontinuities in the static response and in the vibrational resonant frequencies indicate a sudden change in the shape and in the tensile strain of the membranes. We also find that the hystereses can be removed and regular responses can be restored by annealing the resonators. We hypothesize that the hysteretic nature of the responses may originate from an interplay between the rugged morphology of the membranes and adsorbates trapped within the confine of the folds.

cond-mat.mes-hall↗

arXiv2025

Symmetry breaking of large-amplitude parametric oscillations in few-layer graphene nanomechanical resonators

Graphene nanomechanical resonators are well suited for the study of parametric oscillations. Their large frequency tunability and their pronounced nonlinearities enable an efficient modulation of their resonant frequencies. Here, we present measurements of the response of few-layer graphene nanomechanical resonators, each driven by a large parametric pump at frequency $2\omega$ and a weak external drive at $\omega$, where $\omega$ is set near the mechanical resonant frequency $\omega_0$. The pump actuates the resonator beyond the threshold for large-amplitude parametric oscillations, while the drive breaks the symmetry between the parametric phase states. By increasing and decreasing a gate voltage to detune $\omega_0$ in the presence of the pump and the drive, we observe a double hysteresis in the response. The double hysteresis reveals the existence of two possible large-amplitude vibrational states whose phase difference is nearly $\pi$ radians. We deterministically prepare the resonator in either one of these states by cycling the gate voltage. We measure the stationary occupation probabilities of the two states in the presence of a white Gaussian force noise, and find that they strongly depend on the amplitude and on the phase of the external drive. The phase states of parametric oscillations with broken amplitude symmetry can be mapped to biased bi-modal degrees of freedom, such as Ising spins in an external magnetic field. Therefore, they hold promise as units of binary information.

cond-mat.mes-hall↗

arXiv2025

A few-layer graphene nanomechanical resonator driven by digitally modulated video signals

Nanomechanical resonators driven by multifrequency signals combine the physics of mesoscopic vibrations and the technologies of radio communication. Their simplest property stems from their resonant response: they behave as filters, responding only to driving signals whose frequency range is contained within that of the mechanical response. While the response is routinely probed with a single tone drive, a multifrequency drive offers the possibility of inducing richer vibrational dynamics. In this case, all the frequency components of the drive are simultaneously transduced into vibrations with different amplitudes and phases that superimpose and interfere. Here, we employ a few-layer graphene nanomechanical resonator as a filter for broadband, digitally modulated video signals. We transduce the modulated drive into modulated vibrations, which we demodulate into a nanomechanical video. Our work reveals distinctive features in vibrations driven by a coherent, multifrequency drive unseen in vibrations actuated by single tone drives or by a force noise.

cond-mat.mes-hall↗

arXiv2025

Unispeaker: A Unified Approach for Multimodality-driven Speaker Generation

Recent advancements in personalized speech generation have brought synthetic speech increasingly close to the realism of target speakers' recordings, yet multimodal speaker generation remains on the rise. This paper introduces UniSpeaker, a unified approach for multimodality-driven speaker generation. Specifically, we propose a unified voice aggregator based on KV-Former, applying soft contrastive loss to map diverse voice description modalities into a shared voice space, ensuring that the generated voice aligns more closely with the input descriptions. To evaluate multimodality-driven voice control, we build the first multimodality-based voice control (MVC) benchmark, focusing on voice suitability, voice diversity, and speech quality. UniSpeaker is evaluated across five tasks using the MVC benchmark, and the experimental results demonstrate that UniSpeaker outperforms previous modality-specific models. Speech samples are available at \url{https://UniSpeaker.github.io}.

cs.SD↗

arXiv2024

FlashAudio: Rectified Flows for Fast and High-Fidelity Text-to-Audio Generation

Recent advancements in latent diffusion models (LDMs) have markedly enhanced text-to-audio generation, yet their iterative sampling processes impose substantial computational demands, limiting practical deployment. While recent methods utilizing consistency-based distillation aim to achieve few-step or single-step inference, their one-step performance is constrained by curved trajectories, preventing them from surpassing traditional diffusion models. In this work, we introduce FlashAudio with rectified flows to learn straight flow for fast simulation. To alleviate the inefficient timesteps allocation and suboptimal distribution of noise, FlashAudio optimizes the time distribution of rectified flow with Bifocal Samplers and proposes immiscible flow to minimize the total distance of data-noise pairs in a batch vias assignment. Furthermore, to address the amplified accumulation error caused by the classifier-free guidance (CFG), we propose Anchored Optimization, which refines the guidance scale by anchoring it to a reference trajectory. Experimental results on text-to-audio generation demonstrate that FlashAudio's one-step generation performance surpasses the diffusion-based models with hundreds of sampling steps on audio quality and enables a sampling speed of 400x faster than real-time on a single NVIDIA 4090Ti GPU. Code will be available at https://github.com/liuhuadai/FlashAudio.

eess.AS↗

arXiv2024

IntrinsicVoice: Empowering LLMs with Intrinsic Real-time Voice Interaction Abilities

Current methods of building LLMs with voice interaction capabilities rely heavily on explicit text autoregressive generation before or during speech response generation to maintain content quality, which unfortunately brings computational overhead and increases latency in multi-turn interactions. To address this, we introduce IntrinsicVoic,e an LLM designed with intrinsic real-time voice interaction capabilities. IntrinsicVoice aims to facilitate the transfer of textual capabilities of pre-trained LLMs to the speech modality by mitigating the modality gap between text and speech. Our novelty architecture, GroupFormer, can reduce speech sequences to lengths comparable to text sequences while generating high-quality audio, significantly reducing the length difference between speech and text, speeding up inference, and alleviating long-text modeling issues. Additionally, we construct a multi-turn speech-to-speech dialogue dataset named \method-500k which includes nearly 500k turns of speech-to-speech dialogues, and a cross-modality training strategy to enhance the semantic alignment between speech and text. Experimental results demonstrate that IntrinsicVoice can generate high-quality speech response with latency lower than 100ms in multi-turn dialogue scenarios. Demos are available at https://instrinsicvoice.github.io/.

cs.SD↗

arXiv2024

Deep Learning Meets OBIA: Tasks, Challenges, Strategies, and Perspectives

Deep learning has gained significant attention in remote sensing, especially in pixel- or patch-level applications. Despite initial attempts to integrate deep learning into object-based image analysis (OBIA), its full potential remains largely unexplored. In this article, as OBIA usage becomes more widespread, we conducted a comprehensive review and expansion of its task subdomains, with or without the integration of deep learning. Furthermore, we have identified and summarized five prevailing strategies to address the challenge of deep learning's limitations in directly processing unstructured object data within OBIA, and this review also recommends some important future research directions. Our goal with these endeavors is to inspire more exploration in this fascinating yet overlooked area and facilitate the integration of deep learning into OBIA processing workflows.

cs.CV↗

arXiv2024

CosyVoice: A Scalable Multilingual Zero-shot Text-to-speech Synthesizer based on Supervised Semantic Tokens

Recent years have witnessed a trend that large language model (LLM) based text-to-speech (TTS) emerges into the mainstream due to their high naturalness and zero-shot capacity. In this paradigm, speech signals are discretized into token sequences, which are modeled by an LLM with text as prompts and reconstructed by a token-based vocoder to waveforms. Obviously, speech tokens play a critical role in LLM-based TTS models. Current speech tokens are learned in an unsupervised manner, which lacks explicit semantic information and alignment to the text. In this paper, we propose to represent speech with supervised semantic tokens, which are derived from a multilingual speech recognition model by inserting vector quantization into the encoder. Based on the tokens, we further propose a scalable zero-shot TTS synthesizer, CosyVoice, which consists of an LLM for text-to-token generation and a conditional flow matching model for token-to-speech synthesis. Experimental results show that supervised semantic tokens significantly outperform existing unsupervised tokens in terms of content consistency and speaker similarity for zero-shot voice cloning. Moreover, we find that utilizing large-scale data further improves the synthesis performance, indicating the scalable capacity of CosyVoice. To the best of our knowledge, this is the first attempt to involve supervised speech tokens into TTS models.

cs.SD↗

arXiv2024

The Impact of Quantization and Pruning on Deep Reinforcement Learning Models

Deep reinforcement learning (DRL) has achieved remarkable success across various domains, such as video games, robotics, and, recently, large language models. However, the computational costs and memory requirements of DRL models often limit their deployment in resource-constrained environments. The challenge underscores the urgent need to explore neural network compression methods to make RDL models more practical and broadly applicable. Our study investigates the impact of two prominent compression methods, quantization and pruning on DRL models. We examine how these techniques influence four performance factors: average return, memory, inference time, and battery utilization across various DRL algorithms and environments. Despite the decrease in model size, we identify that these compression techniques generally do not improve the energy efficiency of DRL models, but the model size decreases. We provide insights into the trade-offs between model compression and DRL performance, offering guidelines for deploying efficient DRL models in resource-constrained settings.

cs.LG↗

arXiv2024

FunAudioLLM: Voice Understanding and Generation Foundation Models for Natural Interaction Between Humans and LLMs

This report introduces FunAudioLLM, a model family designed to enhance natural voice interactions between humans and large language models (LLMs). At its core are two innovative models: SenseVoice, which handles multilingual speech recognition, emotion recognition, and audio event detection; and CosyVoice, which facilitates natural speech generation with control over multiple languages, timbre, speaking style, and speaker identity. SenseVoice-Small delivers exceptionally low-latency ASR for 5 languages, and SenseVoice-Large supports high-precision ASR for over 50 languages, while CosyVoice excels in multi-lingual voice generation, zero-shot in-context learning, cross-lingual voice cloning, and instruction-following capabilities. The models related to SenseVoice and CosyVoice have been open-sourced on Modelscope and Huggingface, along with the corresponding training, inference, and fine-tuning codes released on GitHub. By integrating these models with LLMs, FunAudioLLM enables applications such as speech-to-speech translation, emotional voice chat, interactive podcasts, and expressive audiobook narration, thereby pushing the boundaries of voice interaction technology. Demos are available at https://fun-audio-llm.github.io, and the code can be accessed at https://github.com/FunAudioLLM.

cs.SD↗