Endong Sun

🔬 Research Interests

My research focuses on the intersection of artificial intelligence and robotics, with particular emphasis on:

• Robot Learning: Developing algorithms that enable robots to learn complex behaviors through machine learning, deep learning, and reinforcement learning.
• Human Demonstration for Robot Learning: Training users to provide high-quality demonstrations and data, enabling robots to learn more effectively from human input. My work explores how to design intuitive processes and guidance so that non-expert users can generate better training data for robot learning systems.

I believe that the future of robotics lies in creating systems that can learn, adapt, and work alongside humans effectively. My research aims to advance the era of human-robot collaboration, enabling seamless and effective interaction between people and intelligent robotic systems.

🔥 Latest News

• 2025: 🎉 Paper "Training People to Reward Robots" accepted at IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2025
• March 2024: 🎉 Successfully passed the KCL 9-month review of my PhD project, marking a significant milestone in my research journey.
• June 2023: 🎓 Started my PhD journey at King's College London, beginning an exciting new chapter in Robot Learning and HRI research.
• March 2023: 💼 Completed my internship at SearchPilot, gaining valuable industry experience in London.

📝 Publications

Training People to Reward Robots

Authors: Endong Sun*, Yuqing Zhu, Matthew Howard

Conference: IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2025

Abstract: This paper presents novel approaches to training humans to provide effective rewards for robot learning systems, improving the human-robot interaction loop in reinforcement learning scenarios.

Links: 📄 Paper

🎥 Video

Knowledge Distilled Ensemble Model for sEMG-based Silent Speech Interface

Authors: Wenqiang Lai, Qihan Yang, Ye Mao, Endong Sun, Jiangnan Ye*

Conference: IEEE EUROCON 2023 - 20th International Conference on Smart Technologies

Abstract: Our innovative model successfully classifies 26 NATO phonetic alphabet datasets with 3900 data samples, enabling the unambiguous generation of any English word through spelling. Extensive experiments validate the effectiveness of KDE-SSI, achieving an impressive test accuracy of 85.9%.

Technical Details: Processed time-series signals and applied trained robust ML and DL models (SVM/KNN and 1D-CNN, CNN-LSTM, etc.) to classify 26 phonetic alphabet words using Python. Came up with a better baseline (ResNet-1D) for our teamwork to enhance the performance of our DL model.

Links: 📄 Paper

🎥 Demo

🎖 Honors and Awards

• 🎓 2023: PhD Studentships in Faculty of Natural, Mathematics & Engineering Sciences (Full-fund)
• 💰 2023: ICRA@40 Grant ($2000)
• 💰 2023: Research, Training and Support Grant (£8750)
• 🥇 2022: The Dick Poortvliet Award (1st) in the IEEE student branch paper contest
• 💰 2019: International Excellence Scholarship (£2,000 awarded)
• 🌟 2020: Outstanding Academic Achievement Award

📖 Education

• 2023–2027 (Expected): PhD in Robot Learning and HRI, King’s College London
  Research Focus: Robot Learning and Human-Robot Interaction
• 2021–2022: MSc in Applied Machine Learning, Imperial College London
  Merit achieved (69.15%)
• 2018–2021: BEng in Electrical and Electronic Engineering, Swansea University
  First Class Honours (82.25%)

💼 Professional Experience

• March 2023 – June 2023: Data Scientist at SearchPilot, London, UK
  • Outlier Detection Pipeline: Developed outlier detection pipeline using clustering algorithms (K-means, GMM, DBSCAN), Isolation Forest, and Variational Autoencoder.
  • Model Performance Improvement: Improved company’s model performance by 75.75%, including a decrease of 38% in prediction error, an increase of 0.5% of R², and a narrower confident prediction band of 2.81%.
  • Combinatorial Optimization: Solved combinatorial optimization to make control and variant groups as similar as possible using meta-heuristic algorithms (GA, PSO, etc.).
  • Technical Skills: Applied advanced machine learning techniques for real-world business problems.