Yao Teng滕 尧

02Selected Research [ Google Scholar ]

Preprint · 2026 survey NEW

Towards Next-Generation Architectures of Large Language Models: A Survey

Yao Teng†, Guoxuan Chen†, Yimeng Chen*, Han Shi*, Piotr Piękos, Jie Ran, Yutao Zeng, Chuanyang Zheng, Jingyao Li, Li Nanbo, Haoli Bai, Chao Huang, Xihui Liu, Di He, Weiyang Liu and Jürgen Schmidhuber

A systematic overview of how LLM architectures are evolving beyond the canonical Transformer: efficient attention, linear-complexity sequence models, MoE, multi-token prediction and more.

Paper Code

arXiv 2026video tokenizer

Adaptive 1D Video Diffusion Autoencoder

Yao Teng, Minxuan Lin, Xian Liu, Shuai Wang, Xiao Yang and Xihui Liu

Diffusion-based video generation on a one-dimensional video diffusion tokenizer.

Paper

IEEE T-PAMIautoregressive T2I

SJD++: Improved Speculative Jacobi Decoding for Training-free Acceleration of Discrete Auto-regressive Text-to-Image Generation

Yao Teng, Zhihuan Jiang, Han Shi, Xian Liu, Xuefei Ning, Guohao Dai, Yu Wang, Zhenguo Li and Xihui Liu

Improvement over SJD in speed.

Paper

NeurIPS 2025autoregressive T2I

Speculative Jacobi-Denoising Decoding for Accelerating Autoregressive Text-to-image Generation

Yao Teng, Fuyun Wang, Xian Liu, Zhekai Chen, Han Shi, Yu Wang, Zhenguo Li, Weiyang Liu, Difan Zou and Xihui Liu

Unifying denoising and autoregressive decoding.

Paper

ICLR 2025autoregressive T2I

Accelerating Auto-regressive Text-to-Image Generation with Training-free Speculative Jacobi Decoding

Yao Teng, Han Shi, Xian Liu, Xuefei Ning, Guohao Dai, Yu Wang, Zhenguo Li and Xihui Liu

The first training-free acceleration method for T2I AR models.

Paper Code

arXiv 2024diffusion

DiM: Diffusion Mamba for Efficient High-Resolution Image Synthesis

Yao Teng, Yue Wu, Han Shi, Xuefei Ning, Guohao Dai, Yu Wang, Zhenguo Li and Xihui Liu

A Mamba-based diffusion model for image generation.

Paper Code

arXiv 2023video editing

Drag-A-Video: Non-rigid Video Editing with Point-based Interaction

Yao Teng, Enze Xie, Yue Wu, Haoyu Han, Zhenguo Li and Xihui Liu

The “drag” in video editing.

Paper

ICCV 2023detection

StageInteractor: Query-based Object Detector with Cross-stage Interaction

Yao Teng, Haisong Liu, Sheng Guo and Limin Wang

New label assignment and feature interaction for query-based object detection.

Paper Code

IJCVdetection

Logit Normalization for Long-tail Object Detection

Liang Zhao*, Yao Teng* and Limin Wang

A test-time batch normalization to calibrate logits for long-tail object detection.

Paper Code

CVPR 2022scene graph

Structured Sparse R-CNN for Direct Scene Graph Generation

Yao Teng and Limin Wang

The first one-stage end-to-end scene graph generation framework without an extra object detector at inference.

Paper Code Zhihu

ICCV 2021scene graph

Target Adaptive Context Aggregation for Video Scene Graph Generation

Yao Teng, Limin Wang, Zhifeng Li and Gangshan Wu

The first framework unifying frame-level and video-level scene graph generation.

Paper Code

03Selected Awards

• 2021National Scholarship for PostgraduatesNanjing University
• 2021Pacemaker to Outstanding Graduate StudentNanjing University
• 20211st Prize, Scholarship for Postgraduate StudentsNanjing University
• 2019Silver Medal, The 2019 ICPC China Shaanxi Provincial Programming ContestXidian University
• 2016 – 20182nd Award, University ScholarshipXidian University