Few-Shot &
Meta-Learning
Learning from less. Relation networks and meta-critic methods for better generalization and transfer.
LEARNING TO LEARN FLOOD SUNG / XVI ROBOTICS
Intelligence
in motion.
From learning to reasoning.
From digital agents to the physical world.
AI researcher. Robotics founder.
Building the next generation of embodied intelligence.
Explore my work ↓
Building the next generation of embodied intelligence.
01 / RESEARCHCURIOSITY → CAPABILITY
Learning is just
the beginning.
I work on systems that learn, reason, and act. My research connects few-shot learning and reinforcement learning with foundation agents — and brings that intelligence into the physical world.
Few-Shot &
Reinforcement
Learning
Learning through action. Reasoning, planning, formal proof, and long-horizon agent behavior.
REASONING THROUGH EXPERIENCEFoundation
Agents
Intelligence that gets things done. Multimodal models, computer-use agents, and tool-using systems.
FROM PERCEPTION TO ACTION02 / SELECTED PUBLICATIONSGoogle Scholar ↗
Ideas into
evidence.
11,140CITATIONS · 2026-09-08 SNAPSHOT
THE PHYSICAL FRONTIERBeyond
XVI ROBOTICS / EMBODIED INTELLIGENCE
Beyond
the screen.
General-purpose humanoid robot brains.
The next chapter of intelligence, at XVI Robotics.
03 / OPEN SOURCEGitHub ↗
Knowledge
travels further.
Research tools, implementations, and reading roadmaps.
Built to be shared. Open for everyone.
OS / 01✳ 39,556
Deep Learning Papers Reading Roadmap
Deep Learning papers reading roadmap for anyone eager to learn this field.
OS / 02✳ 2,656Meta Learning Papers
Meta learning, learning to learn, one-shot learning, and few-shot learning papers.
OS / 03✳ 1,076LearningToCompare_FSL
PyTorch code for the CVPR 2018 Relation Network few-shot learning paper.
OS / 04✳ 584DRL FlappyBird
Playing Flappy Bird using deep reinforcement learning and DQN.
OS / 05✳ 576DDPG
DDPG continuous control reimplementation based on OpenAI Gym and TensorFlow.
OS / 06✳ 281LLM with RL papers
A collection of papers on large language models with reinforcement learning.
46.1KGITHUB STARS
121REPOSITORIES
Open.BY DESIGN
STATS SNAPSHOT / 2026-09-0804 / THE PERSON BEHIND THE WORKALWAYS EXPLORING
Hi, I’m
Flood Sung.
Founder & CEO of XVI Robotics, focused on general-purpose humanoid robot brain systems. My research spans reinforcement learning, meta-learning, few-shot learning, agentic intelligence, and multimodal foundation models.
2025—Present
XVI Robotics
Founder & CEO
2023—2025
Moonshot AI
Member of Technical Staff & RL Lead
2022—2023
Inspir AI
Research Manager
2020—2022
ByteDance AI Lab
Research Scientist
Earlier
Inspir AI
Algorithm Expert
Earlier
Independent Research & Open Source
Earth textures: Solar System Scope / CC BY 4.0 · resized for the web.