论文分享 | 多模态大模型相关研究进展

我们从2024-12-25到2024-12-31的44篇文章中精选出5篇优秀的工作分享给读者。

1. ETTA: Elucidating the Design Space of Text-to-Audio Models
2. Zero-resource Speech Translation and Recognition with LLMs
3. CrossSpeech++: Cross-lingual Speech Synthesis with Decoupled Language and Speaker Generation
4. MBQ: Modality-Balanced Quantization for Large Vision-Language Models
5. Are audio DeepFake detection models polyglots?

1.ETTA: Elucidating the Design Space of Text-to-Audio Models

Authors: Sang-gil Lee, Zhifeng Kong, Arushi Goel, Sungwon Kim, Rafael Valle, Bryan Catanzaro

https://arxiv.org/abs/2412.19351

论文摘要

Recent years have seen significant progress in Text-To-Audio (TTA) synthesis, en abling users to enrich their creative workflows with synthetic audio generated from natural language prompts. Despite this progress, the effects of data, model archi tecture, training objective functions, and sampling strategies on target benchmarks are not well understood. With the purpose of providing a holistic understanding of the design space of TTA models, we set up a large-scale empirical experiment focused on diffusion and flow matching models. Our contributions include: 1) AF-Synthetic, a large dataset of high quality synthetic captions obtained from an audio understanding model; 2) a systematic comparison of different architectural, training, and inference design choices for TTA models; 3) an analysis of sampling methods and their Pareto curves with respect to generation quality and inference speed. We leverage the knowledge obtained from this extensive analysis to pro pose our best model dubbed Elucidated Text-To-Audio (ETTA). When evaluated on AudioCaps and MusicCaps, ETTA provides improvements over the baselines trained on publicly available data, while being competitive with models trained on proprietary data. Finally, we show ETTA’s improved ability to generate creative audio following complex and imaginative captions– a task that is more challenging than current benchmarks.

论文简评

本篇论文通过系统性的研究，探讨了文本到音频（Text-to-Audio）模型的设计空间。通过对影响模型性能的各种因素进行全面的实证分析，提出了AF-Synthetic大型合成文本集，并对不同模型架构、训练目标以及采样方法进行了详尽比较。最终，提出的模型E-ETTA在多个基准数据集上表现出了显著提升，表明其具有实际应用价值和潜在影响力，对于推动该领域的发展起到了积极作用。这一系列研究成果不仅为未来TTS技术提供了新方向，也为人工智能领域的研究开辟了新思路。

2.Zero-resource Speech Translation and Recognition with LLMs

Authors: Karel Mundnich, Xing Niu, Prashant Mathur, Srikanth Ronanki, Brady Houston, Veera Raghavendra Elluru, Nilaksh Das, Zejiang Hou, Goeric Huybrechts, Anshu Bhatia, Daniel Garcia-Romero, Kyu J. Han, Katrin Kirchhoff

https://arxiv.org/abs/2412.18566

论文摘要

Despite recent advancements in speech processing, zero-resource speech translation (ST) and automatic speech recognition (ASR) remain challenging problems. In this work, we propose to leverage a multilingual Large Language Model (LLM) to perform ST and ASR in languages for which the model has never seen paired audio-text data. We achieve this by using a pre-trained multilingual speech encoder, a multilingual LLM, and a lightweight adaptation module that maps the audio representations to the token embedding space of the LLM. We perform several experiments both in ST and ASR to understand how to best train the model and what data has the most impact on performance in previously unseen languages. In ST, our best model is capable of achieving BLEU scores over 23 in CoVoST2 for two previously unseen languages, while in ASR, we achieve WERs of up to 28.2%. We finally show that the performance of our system is bounded by the ability of the LLM to output text in the desired language.

论文简评

本文提出了一种基于多语言大型语言模型（LLMs）的零资源语音翻译（ST）和自动语音识别（ASR）方法，并通过介绍一个轻量级适应模块来连接预训练的语音编码器到LLM的词嵌入空间。实验结果显示，在未见的语言上取得了令人满意的性能指标，表明了所提出方案的有效性。该研究解决了语音处理中的一个重大挑战，特别是对于没有配对音频-文本数据的语言。此外，利用一种新颖的方法，该方案结合了多语言语音编码器和LLMs，可能有助于推动语音处理领域的最新进展。实验结果展示了使用所提出的策略在语音翻译和语音识别任务上的可行性。

3.CrossSpeech++: Cross-lingual Speech Synthesis with Decoupled Language and Speaker Generation

Authors: Ji-Hoon Kim, Hong-Sun Yang, Yoon-Cheol Ju, Il-Hwan Kim, Byeong-Yeol Kim, Joon Son Chung

https://arxiv.org/abs/2412.20048

论文摘要

The goal of this work is to generate natural speech in multiple languages while maintaining the same speaker identity, a task known as cross-lingual speech synthesis. A key challenge of cross-lingual speech synthesis is the language-speaker entanglement problem, which causes the quality of cross-lingual systems to lag behind that of intra-lingual systems. In this paper, we propose CrossSpeech++, which effectively disentangles language and speaker information and significantly improves the quality of cross-lingual speech synthesis. To this end, we break the complex speech generation pipeline into two simple components: language-dependent and speaker-dependent generators. The language-dependent generator produces linguistic variations that are not biased by specific speaker attributes. The speaker-dependent generator models acoustic variations that characterize speaker identity. By handling each type of information in separate modules, our method can effectively disentangle language and speaker representation. We conduct extensive experiments using various metrics, and demonstrate that CrossSpeech++ achieves significant improvements in cross-lingual speech synthesis, outperforming existing methods by a large margin.

论文简评

这篇关于跨语言语音合成的论文提出了CrossSpeech++框架，旨在解决跨语言语音合成中存在的一种挑战：语言与说话者之间的耦合问题。该文提出了一种新颖的方法，即分离语言依赖和说话者依赖特征生成器，通过大量的实验验证了这种方法的有效性，显著提高了合成语音的质量。

论文的核心亮点在于其提出的架构，它能够有效地解决语言与说话者之间的耦合问题，并且可以独立地处理两个维度的信息。这种创新的设计使得CrossSpeech++在解决跨语言语音合成中的关键技术问题上取得了突破性进展。

此外，该文还进行了广泛的实验测试，以评估不同方法的效果，证明了CrossSpeech++在性能上的优越性。这些实证研究不仅为相关领域的研究人员提供了宝贵参考，也为实际应用提供了一个可行的技术方案。

总的来说，CrossSpeech++是一个具有重要理论意义和实用价值的研究成果。它解决了跨语言语音合成领域的一个长期存在的难题，同时展示了强大的可扩展性和灵活性，有望在未来的发展中继续发挥重要作用。

4.MBQ: Modality-Balanced Quantization for Large Vision-Language Models

Authors: Shiyao Li, Yingchun Hu, Xuefei Ning, Xihui Liu, Ke Hong, Xiaotao Jia, Xiuhong Li, Yaqi Yan, Pei Ran, Guohao Dai, Shengen Yan, Huazhong Yang, Yu Wang

https://arxiv.org/abs/2412.19509

论文摘要

Vision-Language Models (VLMs) have enabled a variety of real-world applications. The large parameter size of VLMs brings significant memory and computation overhead, posing challenges for deployment. Post-Training Quantization (PTQ) is an effective technique to reduce memory and computation overhead. Existing PTQ methods mainly focus on large language models (LLMs), without considering the differences across other modalities. In this paper, we discover that there is a significant difference in sensitivity between language and vision tokens in large VLMs. Therefore, treating tokens from different modalities equally, as in existing PTQ methods, may over-emphasize the insensitive modalities, leading to significant accuracy loss. To address this issue, we propose a simple yet effective method, Modality-Balanced Quantization, for large VLMs. Specifically, this method incorporates the different sensitivities across modalities during the calibration process to minimize the reconstruction loss for better quantization parameters. Extensive experiments show that this method can significantly improve task accuracy by up to 4.4% and 11.6% under W3 and W4A8 quantization for 7B to 70B VLMs, compared to SOTA baselines. Additionally, we implement a W3 GPU kernel that fuses the dequantization and GEMV operators, achieving a 1.4× speedup on LLaVA-onevision-7B on the RTX 4090. The code is available at https://github.com/thu-nics/MBQ.

论文简评

这篇论文主要探讨了大规模视觉语言模型（VLM）中视觉与语言符号之间的敏感性差异，并提出了一种名为Modality-Balanced Quantization（MBQ）的方法来提升这些模型的性能。该方法通过调整对视觉和语言符号敏感性的处理，来优化量化过程中的参数选择，从而提高了任务的准确率。实验结果表明，在多种量化方案下，这种方法都能显著提高VLM的性能，尤其是在不同架构和量化条件下。因此，这篇论文为如何有效处理视觉与语言数据之间复杂关系提供了新思路，对于推动视觉语言模型的发展具有重要意义。

5.Are audio DeepFake detection models polyglots?

Authors: Bartłomiej Marek, Piotr Kawa, Piotr Syga

https://arxiv.org/abs/2412.17924

论文摘要

Since the majority of audio DeepFake (DF) detection methods are trained on English-centric datasets, their applicability to non-English languages remains largely unexplored. In this work, we present a benchmark for the multilingual audio DF detection challenge by evaluating various adaptation strategies. Our experiments focus on analyzing models trained on English benchmark datasets, as well as intra-linguistic (same-language) and cross-linguistic adaptation approaches. Our results indicate considerable variations in detection efficacy, highlighting the difficulties of multilingual settings. We show that limiting the dataset to English negatively impacts efficacy, while stressing the importance of data in the target language.

论文简评

该论文针对音频深度伪造检测领域的一个重要问题——如何在非英语语言中有效检测音频深度伪造进行了深入研究。通过比较和分析现有模型的表现，该文揭示了多语境下音频深度伪造检测的有效性，并强调了语言特定训练的重要性和跨语言设置带来的挑战。此外，论文还提出了一项多语言音频深度伪造检测基准，旨在为解决这一问题提供参考框架。整体来看，该文不仅对音频深度伪造检测领域具有重要的理论意义，也为实际应用提供了实践指导。因此，可以认为该论文是一个有价值的学术成果。

我们欢迎您在评论区中留下宝贵的建议！包括但不限于：

• 可以提出推文中论文简评的不足！
• 可以分享最近更值得推荐的论文并给出理由！

#论文分享 #论文 #人工智能 #ai #算法 #深度学习 #大模型 #多模态 #多模态大模型 #MLLM

推荐阅读