Zhiwei Liu

Zhiwei Liu

Auto-generated profile, based on publicly available information

Zhiwei Liu

The University of Manchester, Manchester, United Kingdom

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
24 August 2024
Verified
Conference Paper
Open access
ZLZhiwei Liu
Zhiwei Liu
KYKailai Yang
Kailai Yang
QXQianqian Xie
Qianqian Xie
TZTianlin Zhang
Tianlin Zhang
SASophia Ananiadou
Sophia Ananiadou
Sentiment analysis and emotion detection are important research topics in natural language processing (NLP) and benefit many downstream tasks. With the widespread application of large language models (LLMs), researchers have started exploring the application of LLMs based on instruction-tuning in the field of sentiment analysis. However, these models only focus on single aspects of affective classification tasks (e.g. sentimental polarity or categorical emotions), and overlook the regression tasks (e.g. sentiment strength or emotion intensity), which leads to poor performance in downstream tasks. The main reason is the lack of comprehensive affective instruction tuning datasets and evaluation benchmarks, which cover various affective classification and regression tasks. Moreover, although emotional information is useful for downstream tasks, existing downstream datasets lack high-quality and comprehensive affective annotations. In this paper, we propose EmoLLMs, the first series of open-sourced instruction-following LLMs for comprehensive affective analysis based on fine-tuning various LLMs with instruction data, the first multi-task affective analysis instruction dataset (AAID) with 234K data samples based on 3 classification tasks and 2 regression tasks to support LLM instruction tuning, and a comprehensive affective evaluation benchmark (AEB) with 8 regression tasks and 6 classification tasks from various sources and domains to test the generalization ability of LLMs. We propose a series of EmoLLMs by fine-tuning LLMs with AAID to solve various affective instruction tasks. We compare our models with a variety of LLMs and sentiment analysis tools on AEB, where our models outperform all other open-sourced LLMs and sentiment analysis tools, and surpass ChatGPT and GPT-4 in most tasks, which shows that the series of EmoLLMs achieve the ChatGPT-level and GPT-4-level generalization capabilities on affective analysis tasks, and demonstrates our models can be used as affective annotation tools. This project is available at https://github.com/lzw108/EmoLLMs/.

12 February 2024 in Information Fusion
Verified
Journal Article
Open access
ZLZhiwei Liu
Zhiwei Liu
TZTianlin Zhang
Tianlin Zhang
KYKailai Yang
Kailai Yang
PTPaul Thompson
Paul Thompson
ZYZeping Yu
Zeping Yu
SASophia Ananiadou
Sophia Ananiadou

8 May 2025
Verified
Conference Paper
Open access
ZLZhiwei Liu
Zhiwei Liu
XZXin Zhang
Xin Zhang
KYKailai Yang
Kailai Yang
QXQianqian Xie
Qianqian Xie
JHJimin Huang
Jimin Huang
SASophia Ananiadou
Sophia Ananiadou


17 August 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
ZLZhiwei Liu
Zhiwei Liu
QQQiqi Qiang
Qiqi Qiang
JGJie Gong
Jie Gong
RGRunteng Guo
Runteng Guo
HLHaoyang Liu
Haoyang Liu
SASophia Ananiadou
Sophia Ananiadou
Simulation has long been an important tool in finance, from Monte Carlo pricing and risk analysis to artificial markets and agent-based models. Recent advances in high-frequency data, reinforcement learning, deep generative models, and large language model (LLM) agents have substantially expanded the scope of financial simulation. Contemporary studies no longer use simulation only to approximate price paths or evaluate portfolios; they increasingly construct interactive market environments, synthetic limit order books, heterogeneous trading populations, user simulators, and multi-agent financial decision systems. This survey reviews the emerging literature on simulation in finance through four dimensions: the financial problems being simulated, the datasets and environments used to support simulation, the methodological paradigms adopted by existing studies, and the evaluation protocols used to assess realism and performance. Based on a structured review table covering market simulators, agent-based models, LLM-based financial agents, reinforcement-learning environments, and generative market models, we propose a taxonomy of financial simulation research and identify major open challenges, including benchmark fragmentation, limited data availability, simulation-to-reality gaps, and the difficulty of evaluating agent behavior in complex financial systems.