Shaobo Cui

I am a fourth-year PhD student in Computer Science at EPFL, supervised by Prof. Boi Faltings (AAAI and EurAI Fellow), with co-supervision from Prof. Antoine Bosselut and mentorship by Prof. Michael C. Gastpar. I received honors including Rising Star(明日之星) in Alibaba Group in 2020, DAAD AInet 2024 Fellow in AI for Science. I also hold 9 patents for industrial applications.

Previously, I worked as a senior NLP engineer at DAMO Academy, Alibaba Group, where I led projects in machine comprehension, dialogue systems, and recommendation algorithms for digital humans. Notably, I developed the recommendation algorithm for Alibaba’s Virtual Influencer from 0-1, showcased at the Winter Olympic Games, Beijing 2022.

I hold a Bachelor’s degree from Beihang University and a Master’s degree from Tsinghua University with distinctions.

Research Highlight: Human-Machine Interaction

I am passionate about enhancing machines’ ability to understand and interact with humans:

• Industrial Dialogue Systems:
  • Enhancing retrieval dialogue systems using session information(QSM paper @ CIKM 2020, CSM paper @ KDD 2020, patent, patent).
  • Developing methods for generative dialogue systems, including teacher-student(MTSS) framework (AAAI 2020 paper) and dual adversarial learning(DAL) framework (workshop paper@NAACL).
  • Detecting optimal response timing to in industrial dialogue systems (TASLP paper).
  • Knowledge extraction via a OneStop approach (workshop paper@WWW).
  • Developing methods and platforms for the automated testing of dialogue systems (patents).
• Virtual Influencer:
  • Building the recommendation algorithm of Alibaba Virtual Influencer from 0-1, leading to its debut at Olympic Games.
  • Applying generative methods to assist virtual influencers in constructing their manuscript (CIKM paper).
• Machine Learning: Greedy coordinate descent, regularized optimal transport, and Lipschitz constrainted WGAN(NeurIPS paper, ITCAI paper).
• Communication Enhancement: Enhances heading in academia, news, SEO, and advertising by balancing summarization, creativity, and algorithmic structure(ACL Findings paper, also nominated as a *best paper candidate from reviewer and meta-reviewer⭐).

Research Highlight: Causal Reasoning

My PhD thesis aims to advance machine understanding of causality, with a focus on:

• Commonsense Causality: We synthesize insights on commonsense causality, reviewing benchmarks, acquisition, and reasoning methods under a novel and clear taxonomy (EMNLP paper).
• Defeasibility in Causal Reasoning: We defined and quantified defeasibility, an often-overlooked aspect of causality (ACL(Findings) paper).
• LLMs’ Consistency on Causal Reasoning: We empirically study whether LLMs can keep self-consistency over nuances hidden inside causality (nuance paper: arxiv).
• Causality for Science: Ongoing. Will release it later.

Research Highlight: Open-Source Software

I actively contribute to the open-source community, developing tools that help advance research and real-world applications in natural language processing and machine learning.

• fancy-title: Python package for evaluating and optimizing titles in academia, SEO, and more.
• causal-strength: Python package for measuring and visualizing causal strength.

Combination of Academic Ideas with Industry Products 🌍

I am proud to have transitioned innovative ideas into real-world applications reaching millions, with highlights including contributions to the Olympic Winter Games, the Taobao APP, and Alibaba Cloud.

• Generative AI: Designed frameworks like MTSS (AAAI paper) and DAL (NAACL paper) for dialogue generation and OneStop QA Maker for QA generation.
• Virtual Influencers: Built the recommendation algorithm for Alibaba’s Virtual Influencer showcased at the Winter Olympic Games, Beijing 2022, and developed generative methods for VI manuscripts (CIKM paper).
• Dialogue Systems: Enhanced retrieval-based KB-QA systems (CIKM paper) and automated testing platforms for dialogue systems (patent).

Patents 📑

I have contributed to 9 patents in AI, machine learning, and dialogue systems, focusing on model training, human-computer interaction, and data processing. Highlights include:

• Cui, Shaobo; Lian, Rongzhong; Jiang, Di; He, Jingzhou. 2018. “Training method and device as well as computer equipment and readable storage medium.” CN108846124B, filed on June 29, 2018, and granted on May 17, 2022.
• Ji, Feng; Peng, Shuke; Cui, Shaobo; Chen, Haiqing. 2020. “Network model obtaining method and device, information processing method and device and electronic equipment.” CN113450786A, filed on March 25, 2020, and granted on July 26, 2024.
• Cui, Shaobo; Ji, Feng; Tang, Dahuai, Kang Xiaoming. 2020. “Dialogue system testing method and device and statement rewriting method.” CN113742461A, filed on May 28, 2020. Patent Pending.
• Kang, Xiaoming; Cui, Shaobo; Ji, Feng. 2020. “A man-machine interaction dialogue method and device.” CN113901181A, filed on July 6, 2020. Patent Pending.
• Fu, Zhenxin; Cui, Shaobo; Ji, Feng; Yan, Rui; Zhang, Ji; Chen, Haiqing. 2020. “Dialogue method and system.” CN114036267A, filed on July 21, 2020. Patent Pending.
• Cui, Shaobo; Ji, Feng; Lin, Zehao; Zhao, Zhongzhou; Chen, Haiqing; Zhang, Yin. 2020. “Dialogue model training method, dialogue method and dialogue device.” CN114254086A, filed on September 25, 2020. Patent Pending.
• Cui, Shaobo; Zu, Xinxing; Bao, Xintong; Zhao, Zhongzhou; Zhang, Ji; Chen, Haiqing. 2021. “A question-answer pair generation method, device, electronic device, and computer storage medium” CN115114416A, filed on March 23, 2021. Patent Pending.
• Cui, Shaobo; Fu, Zhenxin; Ji, Feng; Yan, Rui; Zhao, Zhongzhou; Chen, Haiqing. 2021. “Data processing method, device, electronic equipment and computer storage medium” CN115129831A, filed on March 29, 2021. Patent Pending.
• Cui, Shaobo; Tang, Dahuai; Zhao, Zhongzhou; Zhang,Ji; Chen, Haiqing. 2021. “Inquiry statement rewriting method and device.” CN113673256A, filed on July 21, 2021 and granted on August 27, 2024.

For Prospective Students

I have reserved time slots to supervise master students for thesis and semester projects. Undergraduate and master’s students interested in research projects are encouraged to read this page for more information. Feel free to drop me an email via myfirstname.mylastname[at]epfl.ch