01
Active Learning Markets
Market mechanisms and learning algorithms for cost-effective label acquisition under heterogeneous data prices.
PhD Researcher · Imperial College London
Xiwen Huang
Data Markets, Online Learning & Quantitative AI
I study how learning agents should sequentially acquire costly data under budget and market constraints, with applications in active learning, bandits, forecasting, and multi-agent data markets.
Research focus
Sequential decision-making under costly information
About
Learning, markets, and intelligent data acquisition
I am a PhD researcher in Design Engineering at Imperial College London, supervised by Professor Pierre Pinson.
My research lies at the intersection of machine learning, optimisation, algorithmic economics, and market design. I develop algorithms for acquiring and allocating data when information is costly, dynamic, and strategically supplied.
I am particularly interested in active learning markets, cost-aware online learning, Thompson Sampling, combinatorial bandits, forecasting, and multi-agent decision-making.
Research
Current research themes
Methods for learning efficiently when data, labels, and information must be actively purchased.
02
Online & Sequential Learning
Decision-making over time under uncertainty, limited budgets, shifting distributions, and incomplete feedback.
03
Multi-Agent Data Markets
Allocation and learning in environments with multiple buyers, sellers, exclusivity constraints, and strategic interactions.
Publications
Selected work
Research on data acquisition, active learning, and market-based machine learning.
2026
2026
Under revision
QueryMarket: Cost-Aware Online Active Learning in Data Markets
Manuscript under revision for resubmission to IEEE Transactions on Neural Networks and Learning Systems
Online active learning and sequential data acquisition under budget constraints, with applications to solar power forecasting.
Current
Ongoing research
Sequential Allocation in Multi-Buyer–Multi-Seller Data Markets
Ongoing research
Price-Sensitive Feature Selection for Machine Learning with L0 Regularization
2026
Under major revision
Recovering Efficient Supply Function Equilibrium with a Tax-Subsidy Mechanism
Manuscript under major revision at Production and Operations Management
Studies strategic electricity market modelling, Nash equilibrium computation, and tax-subsidy mechanism design.
Talks & Recognition
Selected presentations
2026
Markets for Data and Analytics Applications
IFORS 2026 · Vienna, Austria
2026
Generalised Data Markets: Learning, Agents, and Applications
Invited seminar · Johns Hopkins University
2025
Active Learning Markets for Forecasting
International Symposium on Forecasting · Best Student Presentation Award
2025
Strategic Data Acquisition & Active Learning Markets
Technical University of Munich and ENBIS Young Statisticians Session
Contact
Interested in online learning, data markets, or quantitative AI?
I am open to research collaborations and opportunities related to machine learning, bandits, data acquisition, forecasting, and applied AI systems.