#### Xuedong Shang, PhD (商雪岽 字雨山)

Machine Learner, Quant

### Short Bio

I obtained my PhD degree at SequeL team (now SCOOL team) -- a research unit of Inria Lille-Nord Europe excelled in sequential decision making problems (bandits and reinforcement learning). I was fortunate to be advised by Emilie Kaufmann (CNRS) and Michal Valko (DeepMind), with whom I also wrote my MSc thesis. My current research focuses on sequential decision making with or without side information. In particular, I'm interested in online learning with partial feedback, a.k.a. multi-armed bandits and its applications (hyper-parameter optimization for machine learning, online algorithm selection, etc). I am also interested in theoretical reinforcement learning. From March 2020 to August 2020, I joined Noah's Ark lab for an internship on AutoML, working with Balázs Kégl (Noah's Ark). Before I was an ancien élève de l'Ecole normale supérieure de Cachan (Ker Lann)1, from which I received a BSc and a MSc of Computer Science. Prior to that I also obtained a BSc of Mathematics from Université Pierre et Marie Curie (Paris VI)2. I also visited University of Kyoto during my MSc for an internship under the supervision of Marco Cuturi (Google Brain). My detailed experiences can be found in my Curriculum Vitae.

1 - Now ENS Cachan has been separated and renamed to ENS Paris Saclay and ENS Rennes.

2 - Now Sorbonne Université.

### Interests

My main research interest lies in:

• Bandit Theory
• (Deep) Reinforcement Learning
• Optimization
• Automated Machine Learning

I'm also interested in:

• AI for Finance
• Blockchain and Web3
• AI for Social Good
• Online Learning
• Natural Language Processing
• Time Series
• Optimal Transport
• Recommender System

### Contact

• List.map (fun x -> "shang.xuedong14@" ^ x ^ ".com") ["gmail"]
• SequeL
Inria Lille – Nord Europe
40 Avenue du Halley
59650 Villeneuve-d'Ascq, France
• (+33) 6 51 88 xx xx
• I'm quite active on different social networks, however please refrain from following me on Instagram, Tumblr and Pinterest, don't ask why ;)