About me

Welcome to my site!

My name is Yinpei Dai, and I am a Ph.D. candidate at Situated Language and Embodied Dialogue (SLED) Lab, University of Michigan. Previously, I was a senior algorithm engineer at Alibaba DAMO Academy. I received my M.S. and B.S. degrees from Tsinghua University.

My main interest is embodied AI. I am especially interested in human-robot dialogue systems, interactive navigation and language-conditioned manipulation.

News

  • [Oct 2024] 🚀 We released RACER, a framework to train visuomotor policies with rich language instructions and failure recovery behaviors.
  • [Jan. 2024] 🎉 Our papar ORION is accepted to ICRA 2024! It’s an agent-based system that enables human-robot interactive personalized navigation. Check out more details here.
  • [June 2023] 🏆 We won 🥇 First Place ($500,000) in the first Amazon Alexa Prize SimBot Challenge! 🎉 Read our technical report here. (Media coverage: U-M, Amazon Science)
  • [Aug. 2022] I started my Ph.D. at University of Michigan!
  • [Jun. 2022] 🎉 Our SPACE series on pre-trained dialogue models have been open-sourced. It’s a model family including GALAXY (AAAI 2022), SPACE-2 (COLING 2022) and SPACE-3 (SIGIR 2022). Check out our code, here.

Publications

* indicates equal contribution

Preprints

RACER: Rich Language-Guided Failure Recovery Policies for Imitation Learning
arXiv preprint arXiv:2409.14674 [website]
Yinpei Dai*, Jayjun Lee*, Nima Fazeli, Joyce Chai

Toward satisfactory public accessibility: A crowdsourcing approach through online reviews to inclusive urban design
arXiv preprint arXiv:2409.08459
Lingyao Li, Songhua Hu, Yinpei Dai, et al.

A Survey on Dialog Management: Recent Advances and Challenges
arXiv preprint arXiv:2005.02233
Yinpei Dai, Huihua Yu, Yixuan Jiang, Chengguang Tang, Yongbin Li, Jian Sun

Conference

Teaching Embodied Reinforcement Learning Agents: Informativeness and Diversity of Language Use
EMNLP 2024 (To be appear)
Jiajun Xi, Yinong He, Jianing Yang, Yinpei Dai, Joyce Chai

Think, Act, and Ask: Open-World Interactive Personalized Robot Navigation
ICRA 2024 [code]
Yinpei Dai, Run Peng, Sikai Li, Joyce Chai

SpokenWOZ: A large-scale speech-text benchmark for spoken task-oriented dialogue agents
NeurIPS 2023 (Datasets and Benchmarks Track) [website]
Shuzheng Si, Wentao Ma, Haoyu Gao, Yuchuan Wu, Ting-En Lin, Yinpei Dai, et al.

Task-Oriented Dialogue System as Natural Language Generation
SIGIR 2022
Weizhi Wang, Zhirui Zhang, Junliang Guo, Yinpei Dai, Boxing Chen, et al.

SPACE-3: Unified Dialog Model Pre-training for Task-Oriented Dialog Understanding and Generation
SIGIR 2022 [code]
Wanwei He*, Yinpei Dai*, Min Yang, Jian Sun, et al.

CGoDial: A large-scale benchmark for chinese goal-oriented dialog evaluation
EMNLP 2022 [data]
Yinpei Dai, Wanwei He, Bowen Li, et al.

SPACE-2: Tree-Structured Semi-Supervised Contrastive Pre-training for Task-Oriented Dialog Understanding
COLING 2022 [code]
Wanwei He*, Yinpei Dai*, Binyuan Hui, et al.

GALAXY: A Generative Pre-trained Model for Task-Oriented Dialog with Semi-Supervised Learning and Explicit Policy Injection
AAAI 2022 [code]
Wanwei He*, Yinpei Dai*, Yinhe Zheng, Yuchuan Wu, Zheng Cao, et al.

Preview, Attend and Review: Schema-Aware Curriculum Learning for Multi-Domain Dialog State Tracking
ACL-IJCNLP 2021
Yinpei Dai, Hangyu Li, Yongbin Li, Jian Sun, Fei Huang, Luo Si, Xiaodan Zhu

Transferable Dialogue Systems and User Simulators
ACL-IJCNLP 2021
Bo-Hsiang Tseng, Yinpei Dai, Florian Kreyssig, Bill Byrne

Unsupervised Learning of Deterministic Dialogue Structure with Edge-Enhanced Graph Auto-Encoder
AAAI 2021
Yajing Sun, Yong Shan, Chengguang Tang, Yue Hu, Yinpei Dai, Jing Yu, et al.

Learning Low-Resource End-To-End Goal-Oriented Dialog for Fast and Reliable System Deployment
ACL 2020
Yinpei Dai, Hangyu Li, Chengguang Tang, Yongbin Li, Jian Sun, Xiaodan Zhu

Improved Learning of Word Embeddings with Word Definitions and Semantic Injection
INTERSPEECH 2020
Yichi Zhang, Yinpei Dai, Zhijian Ou, Huixin Wang, Junlan Feng

Tracking of Enriched Dialog States for Flexible Conversational Information Access
ICASSP 2018
Yinpei Dai, Zhijian Ou, Dawei Ren, Pengfei Yu

Workshops and other venues

SEAGULL: An Embodied Agent for Instruction Following through Situated Dialog
Alexa Prize SimBot Challenge Proceedings (1st Place 🏆)
Yichi Zhang, Jianing Yang, Keunwoo Yu, Yinpei Dai, Shane Storks, et al.

Towards Generalized Models for Beyond Domain API Task-oriented Dialogue
DSTC Workshop, AAAI 2021
Haitao Mi, Qiyu Ren, Yinpei Dai, Yifan He, Jian Sun, Yongbin Li, et al.

Education

  • Aug. 2022 - now. Ph.D. Candidate, Department of Computer Science, University of Michigan, MI, USA.
  • Aug. 2016 - Jun. 2019. M.S., Department of Electronic Engineering, Tsinghua University, Beijing, China.
  • Sep. 2012 - Jul. 2016. B.S., Department of Physics, Tsinghua University, Beijing, China.

Experience

  • Jul. 2019 - Jul. 2022. Senior Algorithm Engineer, DAMO Academy, Alibaba Group, Beijing, China.
  • Jun. 2018 - Sep. 2018. Research Intern, Dialogue Systems Group, University of Cambridge, Cambridge, UK.
  • Oct. 2014 - Jul. 2019. Research Assistant, Speech Processing and Machine Intelligence Lab, Tsinghua University, Beijing, China.

Services

  • Reviewer for Journals - IEEE Access, TASLP, JAIR, RA-L, ACL Rolling Review
  • Reviewer for Conference - AAAI, ACL, EMNLP, HRI, COLM, NeurIPS, ICRA

Honors and Awards

  • First place in Alexa Prize SimBot Challenge, 2023
  • First place in track1 objective evaluation, the Ninth Dialog System Technology Challenge (DSTC9), 2020
  • First prize of National Scholarship, 2014
  • Samsung Scholarship, 2013