Posts by Collection

Portfolio item number 1

Short description of portfolio item number 1

Portfolio item number 2

Short description of portfolio item number 2

EMID: An Emotional Aligned Dataset in Audio-Visual Modality

Published in ACM MM Workshop 2023, 2023

A high-quality music-image cross-modal matching dataset (30K+ pairs) using emotional consistency as the primary basis for cross-modal alignment.

Recommended citation: Jialing Zou*, Jiahao Mei*, Guangze Ye, et al. "EMID: An Emotional Aligned Dataset in Audio-Visual Modality." ACM MM Workshop, 2023.
Download Paper

TEAdapter: Supply Vivid Guidance for Controllable Text-to-Music Generation

Published in IEEE ICME 2024, 2024

A lightweight plugin for diffusion models enabling fine-grained controllable music generation through chord, melody, and instrument features extraction.

Recommended citation: Jialing Zou*, Jiahao Mei*, Xudong Nan, et al. "TEAdapter: Supply Vivid Guidance for Controllable Text-to-Music Generation." IEEE ICME, 2024.
Download Paper

DashengTokenizer: One Layer is Enough for Unified Audio Understanding and Generation

Published in arXiv preprint, 2025

A unified continuous audio tokenizer for audio understanding and generation, outperforming mainstream codec/tokenizer baselines on speech, music, and environmental sound tasks.

Recommended citation: Heinrich Dinkel, Xingwei Sun, Gang Li, Jiahao Mei, et al. "DashengTokenizer: One Layer is Enough for Unified Audio Understanding and Generation." arXiv, 2025.
Download Paper

MM-StoryAgent: Immersive Narrated Storybook Video Generation with a Multi-Agent Paradigm across Text, Image and Audio

Published in NAACL 2025, 2025

An open-source multi-modal multi-agent story video generation framework achieving automated immersive narrated storybook video generation. 85K+ visits on ModelScope.

Recommended citation: Xuenan Xu, Jiahao Mei, Chenliang Li, et al. "MM-StoryAgent: Immersive Narrated Storybook Video Generation with a Multi-Agent Paradigm across Text, Image and Audio." NAACL, 2025.
Download Paper

UniFlow-Audio: Unified Flow Matching for Audio Generation from Omni-Modalities

Published in arXiv preprint, 2025

The first fully open-source unified audio generation framework based on Flow Matching, with a novel Dual-Fusion mechanism supporting text, audio, and video inputs across 7 tasks.

Recommended citation: Xuenan Xu*, Jiahao Mei*, Zihao Zheng, et al. "UniFlow-Audio: Unified Flow Matching for Audio Generation from Omni-Modalities." arXiv, 2025.
Download Paper

LARA-Gen: Enabling Continuous Emotion Control for Music Generation Models via Latent Affective Representation Alignment

Published in arXiv preprint, 2025

A latent affective representation alignment mechanism for continuous fine-grained emotion control in music generation using valence-arousal values.

Recommended citation: Jiahao Mei, Xuenan Xu, Zeyu Xie, et al. "LARA-Gen: Enabling Continuous Emotion Control for Music Generation Models via Latent Affective Representation Alignment." arXiv, 2025.
Download Paper

Dasheng AudioGen: A Unified Model for Generating Coherent Audio Scenes from Text

Published in arXiv preprint, 2025

A unified text-to-audio framework for general audio scene generation, enabling end-to-end collaborative generation of speech, music, sound effects, and environmental sounds.

Recommended citation: Jiahao Mei, Heinrich Dinkel, Yadong Niu, et al. "Dasheng AudioGen: A Unified Model for Generating Coherent Audio Scenes from Text." arXiv, 2025.
Download Paper

WritingBench: A Comprehensive Benchmark for Generative Writing

Published in NeurIPS 2025, 2025

A comprehensive benchmark covering 6 domains, 100 sub-domains, and 1239 queries for long-form creative writing evaluation, with a dynamic evaluation framework achieving 83% human agreement.

Recommended citation: Yuning Wu, Jiahao Mei, Ming Yan, et al. "WritingBench: A Comprehensive Benchmark for Generative Writing." NeurIPS, 2025.
Download Paper

Moyun: A Diffusion-Based Model for Style-Specific Chinese Calligraphy Generation

Published in ACM MM Workshop 2025, 2025

A Vision Mamba and TripleLabel mechanism-based Chinese calligraphy generation model trained on 1.9M+ calligraphy images.

Recommended citation: Kaiyuan Liu, Jiahao Mei, Hengyu Zhang, et al. "Moyun: A Diffusion-Based Model for Style-Specific Chinese Calligraphy Generation." ACM MM Workshop, 2025.
Download Paper

Teaching experience 1

Undergraduate course, University 1, Department, 2014

This is a description of a teaching experience. You can use markdown like any other post.

Teaching experience 2

Workshop, University 1, Department, 2015

This is a description of a teaching experience. You can use markdown like any other post.