Zixuan Jiang

Research archive

Publications

Selected work, connected research series, and the broader publication record. * Equal contribution. † Corresponding author.

Research Series 01

Remote Sensing Image Interpretation

Open-vocabulary perception and detailed language grounding for understanding the Earth at scale.
5papers
ISPRS P&RS
DescribeEarth method and sample results

DescribeEarth method and sample results

Esc to close · Use + to inspect the original image

Featured work

SCI District I Top · IF 12.4

DescribeEarth: Describe Anything for Remote Sensing Images

Kaiyu Li*, Zixuan Jiang*, Xiangyong Cao†, Jiayu Wang, Yuchen Xiao, Deyu Meng, Zhi Wang

PaperCodeStarsDatasetBenchmark
  • Introduces geo-spatial detailed localized captioning.
  • Builds the first describe-anything model for remote sensing.
  • Releases the associated dataset and evaluation benchmark.
Media遥感与深度学习码科智能CV炼丹术

Research Series 02

Agentic Speech Recognition

Interactive speech systems and evaluation methods that move recognition beyond a single pass.
3papers
arXiv 2026
AgenticASR task overview

AgenticASR task overview

Esc to close · Use + to inspect the original image

AgenticASR method overview

AgenticASR method overview

Esc to close · Use + to inspect the original image

Featured work

AAAI 2027 · Under review

AgenticASR: Refining Speech Recognition in Real-World Scenarios via an Agentic Approach

Zixuan Jiang*, Binghao Qiang*, Jiaying Chi*, Yanqiao Zhu, Kai Yu, Xie Chen†

PaperProject PageCodeStarsAASR-Bench
  • Defines Agentic Speech Recognition as final-intent-preserving audio-to-clean-text recognition.
  • Introduces a decoupled ASR–Refiner system for bilingual online and offline refinement.
  • Releases AASR-Bench with 6,637 atomic rubrics and leads the evaluated ASR families.
arXiv 2026
Interactive ASR system overview

Interactive ASR system overview

Esc to close · Use + to inspect the original image

Featured work

TASLP · Under review

Towards Human-Like Interactive Speech Recognition With Agentic Correction and Semantic Evaluation

Zixuan Jiang*, Yanqiao Zhu*, Peng Wang*, Qinyuan Chen, Xinjian Zhao, Xipeng Qiu, Wupeng Wang, Zhifu Gao, Xiangang Li, Kai Yu, Xie Chen†

PaperProject PageLive Demo
  • Extends one-pass ASR into an interactive system with user feedback and semantic correction.
  • Introduces an agent-based framework for interactive speech recognition.
  • Develops the S²ER metric and ISS simulation framework for semantic evaluation.

Research Series 03

Audio Generation & Editing

Unified models for audio generation and editing.
2papers
arXiv 2026
AuK overview: speech generation, acoustic and paralinguistic editing, content editing, enhancement and separation

AuK overview: speech generation, acoustic and paralinguistic editing, content editing, enhancement and separation

Esc to close · Use + to inspect the original image

Featured work

AuK Technical Report: An Open-Source Foundational Model for Speech Generation and Editing

Ziyang Ma, Zhikang Niu, Wenming Tu, Tianrui Wang, Ruiqi Yan, Junxi Liu, Yanru Huo, Nickk Huang, Yang Liu, Qicong Xie, Zeyu Xie, Hui Wang, Haitao Li, Zixuan Jiang, Yalin Li, Jie Fang, Yifan Duan, Zeyue Tian, Guangzheng Li, Haina Zhu, Shuyi Wang, Jinwen Wang, Mingyu Cui, Tian Tan, Auden, Sen Liang, Steve Yves, Shan Yang, Liefeng Bo, Zilong Zheng, Kai Yu, Eng-Siong Chng, Xie Chen

PaperCodeStarsProject Page
  • Unifies speech generation and editing with natural-language instructions and reference audio.
  • AuK-Flash: 4-step inference, approximately 4.5× faster under matched conditions.
  • AuK-VAE: jointly trained on speech, music, and sound.