Weiqing Liu

About

Dr. Weiqing Liu is a Principal Research Manager at Microsoft Research Asia (MSRA), where he leads the MSRA’s Industry Innovation Center (MIIC), the lab’s strategic hub for industry collaboration and applied AI innovation. In this role, he oversees MSRA’s full portfolio of partnerships, driving the translation of frontier AI research into high‑impact real‑world systems.

Before directing MIIC, Dr. Liu led MSRA’s AI for Finance research team, establishing influential collaborations with major financial institutions and contributing foundational work to open‑source ecosystems such as Qlib (https://github.com/microsoft/qlib (opens in new tab)), a widely adopted platform for intelligent investment.

In recent years, his team has been actively involved in both research and practical implementation pertaining to autonomous agents supporting data-centric R&D activities. Their focus encompasses quantitative investment, a range of machine learning tasks, as well as post-training and fine-tuning of large language models (LLMs). This work has enabled the team to develop considerable expertise and a comprehensive understanding of autonomous LLM agents for general R&D applications. The results of their efforts include several published research papers—such as R&D‑Agent (opens in new tab), R&D‑Agent‑Quant (opens in new tab), FT‑Dojo (opens in new tab), Reasoning as Gradient (opens in new tab), and Agent^2 RL-Bench (opens in new tab) —in addition to the open-source project R&D-Agent (https://github.com/microsoft/RD-Agent (opens in new tab)).

In addition to agent works, the team created an innovative foundation model known as the Large Market Model. This model treats financial market orders as elemental units, enabling the identification of fine-grained market patterns. Leveraging controllable market generation and support for injected orders, the Large Market Model functions as a digital twin of real-world markets, facilitating new paradigms in various downstream tasks. Relevant publications, including MARS (opens in new tab) (ICLR 2025) and Controllable Market Generation (opens in new tab) (AAAI 2026), as well as portions of the code (https://github.com/microsoft/MarS (opens in new tab)), have been released.

Beyond agentic and market‑model research, Dr. Liu’s work spans several interconnected themes:

Sequential Decision Making

• “Universal Trading for Order Execution with Oracle Policy Distillation (opens in new tab)“, AAAI 2021.
• “Towards Applicable Reinforcement Learning: Improving the Generalization and Sample Efficiency with Policy Ensemble (opens in new tab)“, IJCAI 2022.
• “Learning Multi-Agent Intention-Aware Communication for Optimal Multi-Order Execution in Finance (opens in new tab)“, KDD 2023.
• “BPQP: A Differentiable Convex Optimization Framework for Efficient End-to-End Learning (opens in new tab)“, NeurIPS 2024.

Time-Variant Pattern

• “DDG-DA: Data Distribution Generation for Predictable Concept Drift Adaptation (opens in new tab)“, AAAI 2022.
• “Learning Multiple Stock Trading Patterns with Temporal Routing Adaptor and Optimal Transport (opens in new tab)“, KDD 2021.
• “Multi-Granularity Residual Learning with Confidence Estimation for Time Series Prediction (opens in new tab)“, WWW 2022.

Heterogenous and Hierarchical Data

• “Stock Trend Prediction with Multi-granularity Data: A Contrastive Learning Approach with Adaptive Fusion (opens in new tab)“, CIKM 2021.
• “REST: Relational Event-driven Stock Trend Forecasting (opens in new tab)“, WWW 2021.
• “Digger-Guider: High-Frequency Factor Extraction for Stock Trend Prediction (opens in new tab)“, TKDE 2024.
• “MG-TSD: Multi-granularity time series diffusion models with guided learning process (opens in new tab)“, ICLR 2024.

Complex Correlation

• “Deep Risk Model: A Deep Learning Solution for Mining Latent Risk Factors to Improve Covariance Matrix Estimation (opens in new tab)“, ICAIF 2021.
• “Temporally Correlated Task Scheduling for Sequence Learning (opens in new tab)“, ICML 2021.
• “Removing Camouflage and Revealing Collusion: Leveraging Gang-crime Pattern in Fraudster Detection (opens in new tab)“, KDD 2023.

Explainability / Interpretability

• “Learning Differential Operators for Interpretable Time Series Modeling (opens in new tab)“, KDD 2022.
• “Measuring Model Complexity of Neural Networks with Curve Activation Functions (opens in new tab)“, KDD 2020.

Dr. Liu holds B.S. and Ph.D. degrees in Computer Science from the University of Science and Technology of China (USTC), completed in 2011 and 2016. He joined MSRA immediately after graduation and continues to drive research at the intersection of agentic AI, financial intelligence, foundation models, and autonomous R&D systems.