Miao is currently a postdoctoral researcher at Institute for Language, Cognition and Computation (ILCC), University of Edinburgh. He works with Prof. Mirella Lapata (ACL Fellow; University of Edinburgh and Google DeepMind) as part of EdinburghNLP, and previously spent a year there as a visiting PhD student. His postdoctoral research is funded through Turing AI World-Leading Researcher Fellowship, awarded to Prof. Mirella Lapata. Miao completed his PhD in NLP/AI at School of Computing and Information Systems, University of Melbourne, where he was jointly advised by Prof. Eduard Hovy (ACL Fellow, AAAI Fellow; University of Melbourne and Carnegie Mellon University) and Prof. Jey Han Lau (University of Melbourne).

Research Interests

Miao's research lies in natural language processing (NLP) and artificial intelligence (AI). His long-term research vision is to develop AI models capable of expert-level reasoning, particularly in integrating complex information from multiple, heterogeneous, and distributed sources. Such models could augment human cognition by complementing our limitations in processing speed and working memory. His research currently focuses on: (1) understanding how humans integrate dispersed information and reason over it to solve complex problems (ACL'24, TMLR'26), with the goal of uncovering computational principles that can inform AI development, including in tasks such as making investment decisions, writing scientific meta-reviews, and synthesising knowledge from hundreds of scientific articles, (2) developing foundation models capable of such complex information integration and reasoning, with a current emphasis on long-context language models and autonomous agents, by translating insights from human cognition into learning paradigms that enable effective learning under under limited data, computation, and experience (AAAI'23, EMNLP'23, ACL'26), and (3) establishing rigorous automated evaluation methodologies for both model outputs and their behaviours, particularly in reasoning language models, to ensure their usefulness, reliability, and verifiability in real-world applications (ACL'24, ACL'25, TMLR'26). His PhD thesis, "Multi-Document Summarization: From Ideational to Opinionated Sources," explored how models can understand and integrate different types of information across multiple documents, reflecting a longstanding research interest that continues to drive his work on complex information integration and reasoning.

Selected Publications

• SciTrek: Evaluating and Improving Long-Context Numerical Reasoning over Scientific Articles
  Miao Li, Alexander Gurung, Irina Saparina, and Mirella Lapata.
  TMLR (September 2026).
  [paper] [openreview] [data] [code]

• Long-Context Reasoning Through Proxy-Based Chain-of-Thought Tuning
  Miao Li, Irina Saparina, Alexander Gurung, and Mirella Lapata.
  ACL 2026. Oral.
  [abstract] [paper] [code] [slides] [poster]

• Decomposed Opinion Summarization with Verified Aspect-Aware Modules
  Miao Li, Jey Han Lau, Eduard Hovy, and Mirella Lapata.
  ACL 2025 Findings.
  [abstract] [paper] [code] [poster]

• A Sentiment Consolidation Framework for Meta-Review Generation
  Miao Li, Jey Han Lau, and Eduard Hovy.
  ACL 2024. Oral.
  [abstract] [paper] [code] [dataset] [slides] [poster]

• NewsBench: A Systematic Evaluation Framework for Assessing Editorial Capabilities of Large Language Models in Chinese Journalism
  Miao Li, Ming-Bin Chen, Bo Tang, Shengbin Hou, Pengyu Wang, Haiying Deng, Zhiyu Li, Feiyu Xiong, Keming Mao, Peng Cheng, and Yi Luo.
  ACL 2024.
  [abstract] [paper] [code] [dataset] [slides] [poster]

• Summarizing Multiple Documents with Conversational Structure for Meta-Review Generation
  Miao Li, Eduard Hovy, and Jey Han Lau.
  EMNLP 2023 Findings.
  [abstract] [paper] [code] [dataset] [poster]

• DeltaScore: Fine-grained Story Evaluation with Perturbations
  Zhuohan Xie, Miao Li, Trevor Cohn, and Jey Han Lau.
  EMNLP 2023 Findings.
  [abstract] [paper] [code]

• Compressed Heterogeneous Graph for Abstractive Multi-Document Summarization
  Miao Li, Jianzhong Qi, and Jey Han Lau.
  AAAI 2023. Oral.
  [abstract] [paper] [code] [slides] [poster]

• A Benchmark and Comprehensive Survey on Knowledge Graph Entity Alignment via Representation Learning
  Rui Zhang, Bayu Distiawan Trisedy, Miao Li, Yong Jiang, Jianzhong Qi.
  VLDB Journal (May 2022).
  [abstract] [paper] [code]

• A Topic Augmented Text Generation Model: Joint Learning of Semantics and Structural Features
  Hongyin Tang, Miao Li, Beihong Jin.
  EMNLP 2019. Oral.
  [abstract] [paper] [appendix]