语音合成论文趋势 2023–2026

主要结论

• 语音合成论文从 2023 年的 125 篇增长到 2026 年的 200 篇，三年间扩大到约 1.6 倍。
• NeurIPS 是发语音合成论文最多的会议，2023–2026 年共 204 篇，占四个会议总量的 33%。
• 交叉方向里涨得最快的是目标检测：同时属于两个主题的论文从 2023 年的 1 篇增至 2026 年的 10 篇。
• 在「语音与音频」的 5 个主题里，它的 2026 年论文数排第 2；同组增长最快的音频生成是 4.3 倍，语音合成是 1.6 倍。

代表论文

按年份和会议均匀挑选，方便快速看到每一届的样貌

• NeurIPS 2026CAPER: Clause-Aligned Process Supervision for Text-to-SQL
• ICML 2026Black-Box Detection of LLM-Generated Text Using Generalized Jensen Shannon Divergence
• ICLR 2026CSRv2: Unlocking Ultra-Sparse Embeddings
• CVPR 2026Aligning Text Images and 3D Structure Token-by-Token
• NeurIPS 2025AdaDetectGPT: Adaptive Detection of LLM-Generated Text with Statistical Guarantees
• ICML 2025How to Synthesize Text Data without Model Collapse?
• ICLR 2025AI as Humanity’s Salieri: Quantifying Linguistic Creativity of Language Models via Systematic Attribution of Machine Text against Web Text
• CVPR 2025AMO Sampler: Enhancing Text Rendering with Overshooting
• NeurIPS 2024SM3-Text-to-Query: Synthetic Multi-Model Medical Text-to-Query Benchmark
• ICML 2024Adaptive Text Watermark for Large Language Models
• ICLR 2024Are Models Biased on Text without Gender-related Language?
• CVPR 2024Kernel Adaptive Convolution for Scene Text Detection via Distance Map Prediction
• NeurIPS 2023[Re] Pure Noise to the Rescue of Insufficient Data
• ICML 2023Approximation and Estimation Ability of Transformers for Sequence-to-Sequence Functions with Infinite Dimensional Input
• ICLR 2023Bag of Tricks for Unsupervised Text-to-Speech
• CVPR 20233D Highlighter: Localizing Regions on 3D Shapes via Text Descriptions
• NeurIPS 2026DiffScore: Text Evaluation Beyond Autoregressive Likelihood
• ICML 2026CoCoEmo: Composable and Controllable Human-Like Emotional TTS via Activation Steering
• ICLR 2026DrVoice: Parallel Speech-Text Voice Conversation Model via Dual-Resolution Speech Representations
• CVPR 2026Learning complete and explainable visual representations from itemized text supervision

全部语音合成论文

共 123 篇，第 2 / 3 页

51. ICLR 2025Continuous Autoregressive Modeling with Stochastic Monotonic Alignment for Speech Synthesis
52. ICLR 2025Deep Kernel Relative Test for Machine-generated Text Detection
53. ICLR 2025DiTTo-TTS: Diffusion Transformers for Scalable Text-to-Speech without Domain-Specific Factors
54. ICLR 2025Effective post-training embedding compression via temperature control in contrastive training
55. ICLR 2025FIRING-Net: A filtered feature recycling network for speech enhancement
56. ICLR 2025Fluid: Scaling Autoregressive Text-to-image Generative Models with Continuous Tokens
57. ICLR 2025HALL-E: Hierarchical Neural Codec Language Model for Minute-Long Zero-Shot Text-to-Speech Synthesis
58. ICLR 2025ImpScore: A Learnable Metric For Quantifying The Implicitness Level of Sentences
59. ICLR 2025Joint Fine-tuning and Conversion of Pretrained Speech and Language Models towards Linear Complexity
60. ICLR 2025Making Text Embedders Few-Shot Learners
61. ICLR 2025MaskGCT: Zero-Shot Text-to-Speech with Masked Generative Codec Transformer
62. ICLR 2025Mining your own secrets: Diffusion Classifier Scores for Continual Personalization of Text-to-Image Diffusion Models
63. ICLR 2025MMTEB: Massive Multilingual Text Embedding Benchmark
64. ICLR 2025Monte Carlo Planning with Large Language Model for Text-Based Game Agents
65. ICLR 2025Not All LLM-Generated Data Are Equal: Rethinking Data Weighting in Text Classification
66. ICLR 2025Not All Prompts Are Made Equal: Prompt-based Pruning of Text-to-Image Diffusion Models
67. ICLR 2025PaLD: Detection of Text Partially Written by Large Language Models
68. ICLR 2025Regularization by Texts for Latent Diffusion Inverse Solvers
69. ICLR 2025Rethinking Artistic Copyright Infringements In the Era Of Text-to-Image Generative Models
70. ICLR 2025ROUTE: Robust Multitask Tuning and Collaboration for Text-to-SQL
71. ICLR 2025Scaling Speech-Text Pre-training with Synthetic Interleaved Data
72. ICLR 2025Scaling Transformers for Low-Bitrate High-Quality Speech Coding
73. ICLR 2025SyllableLM: Learning Coarse Semantic Units for Speech Language Models
74. ICLR 2025T2V2: A Unified Non-Autoregressive Model for Speech Recognition and Synthesis via Multitask Learning
75. ICLR 2025Training-free LLM-generated Text Detection by Mining Token Probability Sequences
76. ICLR 2024An Efficient Membership Inference Attack for the Diffusion Model by Proximal Initialization
77. ICLR 2024Are Models Biased on Text without Gender-related Language?
78. ICLR 2024CLaM-TTS: Improving Neural Codec Language Model for Zero-Shot Text-to-Speech
79. ICLR 2024DIAGNOSIS: Detecting Unauthorized Data Usages in Text-to-image Diffusion Models
80. ICLR 2024DiffAR: Denoising Diffusion Autoregressive Model for Raw Speech Waveform Generation
81. ICLR 2024Exposing Text-Image Inconsistency Using Diffusion Models
82. ICLR 2024Generative Pre-training for Speech with Flow Matching
83. ICLR 2024Harnessing Explanations: LLM-to-LM Interpreter for Enhanced Text-Attributed Graph Representation Learning
84. ICLR 2024Interpreting CLIP's Image Representation via Text-Based Decomposition
85. ICLR 2024LipVoicer: Generating Speech from Silent Videos Guided by Lip Reading
86. ICLR 2024Mega-TTS 2: Boosting Prompting Mechanisms for Zero-Shot Speech Synthesis
87. ICLR 2024NaturalSpeech 2: Latent Diffusion Models are Natural and Zero-Shot Speech and Singing Synthesizers
88. ICLR 2024Out-of-Distribution Detection by Leveraging Between-Layer Transformation Smoothness
89. ICLR 2024PolyVoice: Language Models for Speech to Speech Translation
90. ICLR 2024PromptTTS 2: Describing and Generating Voices with Text Prompt
91. ICLR 2024Provable Robust Watermarking for AI-Generated Text
92. ICLR 2024Self-Supervised Speech Quality Estimation and Enhancement Using Only Clean Speech
93. ICLR 2024SpeechTokenizer: Unified Speech Tokenizer for Speech Language Models
94. ICLR 2024Spoken Question Answering and Speech Continuation Using Spectrogram-Powered LLM
95. ICLR 2024SweetDreamer: Aligning Geometric Priors in 2D diffusion for Consistent Text-to-3D
96. ICLR 2024TEST: Text Prototype Aligned Embedding to Activate LLM's Ability for Time Series
97. ICLR 2024Text-to-3D with Classifier Score Distillation
98. ICLR 2024TOSS: High-quality Text-guided Novel View Synthesis from a Single Image
99. ICLR 2024Towards 3D Molecule-Text Interpretation in Language Models
100. ICLR 2024Würstchen: An Efficient Architecture for Large-Scale Text-to-Image Diffusion Models

常见问题

2023–2026 年四大会议一共有多少篇语音合成论文？

共 621 篇，按年份分别是 2023 年 125 篇、2024 年 145 篇、2025 年 151 篇、2026 年 200 篇。

NeurIPS 2026 有多少篇语音合成论文？

NeurIPS 2026 有 67 篇语音合成论文，上一届是 45 篇。

ICLR 2026 有多少篇语音合成论文？

ICLR 2026 有 43 篇语音合成论文，上一届是 32 篇。

ICML 2026 有多少篇语音合成论文？

ICML 2026 有 36 篇语音合成论文，上一届是 16 篇。

CVPR 2026 有多少篇语音合成论文？

CVPR 2026 有 54 篇语音合成论文，上一届是 58 篇。

哪个会议的语音合成论文最多？

NeurIPS，2023–2026 年共 204 篇。