Department of Computer Science, University of Oxford
Machine Learning and Global Health Network
Email: mengyan.zhang@cs.ox.ac.uk
[CV|Google Scholar|Github| Twitter|Linkedin]

I am Mengyan Zhang (张梦妍), a postdoctoral researcher at the University of Oxford, working with Prof. Seth Flaxman. I am a research member of common room in Kellogg College. Before that, I received my PhD at the Australian National University, under the supervision of Dr. Cheng Soon Ong, Prof. Lexing Xie and Prof. Eduardo Eyras. I was affiliated with Data61, CSIRO and interned at Microsoft Research Asia during my PhD. I’m also short-termly mentored by Sebastien Bubeck via the WiML-T Mentoring Program in 2021. Prior to that, I obtained my bachelor’s degree with first-class honours at the Australian National University and bachelor’s degree at Shandong University in 2018 (2 + 2 joint degree program).

My research interests are sequential decision-making in machine learning, including multi-armed bandits and active learning. I work on both theoretical and practical views of experimental design with two goals: (I) Designing robust algorithms to handle imperfect feedback and understand causal relationships in sequential decision-making. (II) Designing decision-making algorithms to solve real-world problems in various areas, for example, synthetic biology, global health, survey design, and public policy.

Jobs

• Postdoctoral Researcher (2023.05 -), Computational Statistics and Machine Learning, Department of Computer Science, University of Oxford
• NCI-ANU Associate Training Officer (2023.04-2023.07, Casual)
• Research Assistant (2023.04-2023.06) in the University of Tuebingen, working with Claire Vernade.
• Research Assistant (2022.10-2023.04) in RPTU Kaiserslautern and German Research Center for Artificial Intelligence (DFKI), Vollmer Research Group
• Research Internship ( 2021.10 - 2022.03), Social Computing Lab, Microsoft Research Asia, worked on deep contextual bandits for news recommendation.
• Academic Tutor (2019-2021) Australian National University, paid teaching position, total working hours: 300h

Research & Publications/Preprint

• PhD Thesis: Adaptive Recommendations with Bandit Feedback {ANU Open Research Library} (supervisors: Cheng Soon Ong, Lexing Xie, Eduardo Eyras) - Award: CORE Distinguished Dissertation Award Commendation

• Two-Stage Neural Contextual Bandits for Personalised News Recommendation.
  Mengyan Zhang, Thanh Nguyen-Tang, Fangzhao Wu, Zhenyu He, Xing Xie, Cheng Soon Ong. Under Review. {pre-print}

• Gaussian Process Bandits with Aggregated Feedback.
  Mengyan Zhang, Russell Tsuchida, Cheng Soon Ong. AAAI 2022. {pre-print; code; poster; one-page abstract}

• Machine learning guided batched design of a bacterial Ribosome Binding Site.
  Mengyan Zhang, Maciej Bartosz Holowko, Huw Hayman Zumpe, Cheng Soon Ong. ACS Synthetic Biology Journal 2022. {paper; C3DIS 2020 Talk; SEED 2021 Talk}

• Quantile Bandits for Best Arms Identification.
  Mengyan Zhang, Cheng Soon Ong. International Conference on Machine Learning 2021. {paper; code; poster; talk}

• Opportunities and Challenges in Designing Genomic Sequences.
  Mengyan Zhang, Cheng Soon Ong. ICML Workshop on Computational Biology 2021. {paper; poster; talk}

• Active Learning on Knowledge Graph. {software; flowchart; design}

• Honours project: Classification of historical death and occupation coding {thesis} (supervisors: Peter Christen, Timothy Graham)

Teaching & Supervision

• Served as College Advisor in Kellogg College (MT23), providing mentorship to 7 postgraduate students.
• Guest lecturer at RMIT Bioinformatics and Multi-omics data analysis (BIOL 2524) : introduction to ML and applications in biology (remotely, 3 lectures, May 2023) – course materials
• Tutor Statistical Machine Learning (S1 2019, S1 2020, S1 2021)
• Tutor Introduction to Machine Learning (S2 2020)
• (Mar.-Jun. 2021) co-supervision on Nathan Hu for applying DNABERT to yeast promotor. See details here!

Talks & Presentations

• Feb. 2024: LAS Group, ETH Zurich, Switzerland
  slides: Sequential Decision-Making: Theory and Applications in Public Health
• Dec. 2023: Google DeepMind, London
  Design Choices in Sequential Decision-Making with Bandit Feedback 
• Nov. 2023: AIMS seminar, Oxford
  Sequential decision making in public health
• Nov. 2023: Bayes@CIRM Workshop, Marseille, France
  Bayesian optimisation with aggregated feedback
• Jul. 2023: University of Adelaide ADSC Seminar
  Sequential Decision-making: Theory and Applications
• Jun. 2022: ANU AI+ML+Friends seminar (PhD Completion Talk)
  slides: Adaptive Recommendations with Bandit Feedback
• Feb. 2022: Microsoft Research Asian Social Computing Group Seminar
  slides: Bandits in Recommendation System
• Jan. 2022: Microsoft News and Feeds Group
  slides: Best arm identifications: classical settings and methods
• Dec. 2021: WiML workshop in NeurIPS
  Poster presentation: Gaussian Process Bandits with Aggregated Feedback
• Jul. 2021: ICML Workshop on Computational Biology
  Spotlight Talk: Opportunities and Challenges in Designing Genomic Sequences
• Jul. 2021: Thirty-ninth International Conference on Machine Learning
  Poster presentation: Quantile Bandits for Best Arms Identification
• Jul. 2020: Machine Learning Summer School (acceptance rate: 13.84%)
  Poster presentation: Quantile Bandits for Best Arms Identification
• Dec. 2019: Collaborative Conference on Computational and Data Intensive Science
  Talk: Optimized Experimental Design for Translation Initiation using Machine Learning

Visit & Conferences

• Research visit Andreas Krause at Learning & Adaptive Systems (LAS) Group in ETH Zurich, Switzerland, via NCCR Automation Fellowship (Funded, up to CHF 20,000), Feb - April 2024.

• Research visit Silvia Chiappa at Causal Intelligence Team, Google DeepMind, London, 6th Dec 2023.

• Bayesian Statistic autumn school held at CIRM, Marseille, France, from 30 October to 3 November 2023.

• Research visit Prof. Dino Sejdinovic in the School of Computer and Mathematical Sciences at The University of Adelaide, July 2023.
• Reinforcement Learning Summer School (RLSS) 2023, June 26th to July 5th, 2023, Barcelona
• BioInference 2023, 8th-9th June 2023, University of Oxford
• Machine Learning Summer School (MLSS) 2020 at the Max Planck Institute for Intelligent Systems, Tübingen, Germany! (Acceptance rate 13.84%.)

Service

• Reviewer for NeurIPS 2023, AAAI2024, ICLR2024.

Events & News

• Dec. 2023 I visited Silvia Chiappa at Google DeepMind (London) and met the causal intelligence team!
• Nov. 2023 I attend Bayes@CIRM in Marseille, France!
• Jul. 2023 I visit Prof. Dino Sejdinovic in The University of Adelaide!
• Jul. 2023 I attend my PhD Graduation ceremony in ANU!
• Jun. 2023 I attend RLSS 2023 in Barcelona!
• Apr. 2023 I officially join the University of Oxford as a postdoc!
• Feb. 2023 My PhD thesis is officially accepted by ANU!
• Oct. 2022 I start my research assistant internship in Vollmer group remotely!
• Aug. 2022 I submit my PhD thesis: Adaptive Recommendations with Bandit Feedback!
• Jun. 2022 I give my PhD completion talk at ANU AI+ML+Friends seminar!
• Jun. 2022 Our work Machine learning guided batched design of a bacterial Ribosome Binding Site is accepted by the ACS Synthetic Biology Journal!
• Dec. 2021 Our work Gaussian Process Bandits with Aggregated Feedback (abstract) is accepted to be presented at WiML workshop in NeurPIS2021! Come and talk to us!
• Dec. 2021 Our work Gaussian Process Bandits with Aggregated Feedback is accepted by AAAI2022! (Acceptance rate 15%)
• Step. 2021 I start my internship at Microsoft Research Asian (MSRA) in Social Computing team, mentored by Fangzhao Wu.
• June. 2021 Our work Opportunities and Challenges in Designing Genomic Sequences is accepted in the 2021 ICML Workshop on Computational Biology as spotlight talk! Come to talk to us in our poster session!
• May. 2021 Our work Machine Learning guided workflow for Ribosome Binding Site engineering got accepted in Synthetic Biology: Engineering, Evolution & Design (SEED) 2021 Conference for oral abstract! Maciej Holowko will present, see you there!
• May. 2021 Our paper Quantile Bandits for Best Arms Identification got accepted for ICML 2021! See you online!
• Update Mar. 2021 I am collecting good resources for machine learning study, see here
• Mar.-Jun. 2021 I am co-supervising Nathan Hu on applying DNABERT to yeast promotor. See details here!
• Jan.-Jun. 2021: I am participating in the WiML-T Mentoring Program. My paired mentor is Sebastien Bubeck!
• Apr. 2020: I am accepted to MLSS 2020 at the Max Planck Institute for Intelligent Systems, Tübingen, Germany! (Acceptance rate 13.84%.) Lecture videos and slides can be found here.