On the afternoon of November 19, 2025, the 11th Xueshi (Scholarly Pursuit) Doctoral Academic Forum of the School of Management and Engineering, Nanjing University was successfully held at the Xiexin Building of the School of Management and Engineering, Nanjing University. The Xueshi series of forums is co-hosted by the School of Management and Engineering, Nanjing University and the Doctoral Party Branch of the School of Management and Engineering, Nanjing University. It aims to provide a high-level academic platform for interdisciplinary integration for doctoral students, display the outstanding achievements of doctoral students in academic research and innovative practice, and highlight the school's spirit of pursuing truth through learning, and committing to serving the public.

With the theme of Building a Powerful Education Nation, Forging Ahead with Me, the 11th forum covers research in multiple cutting-edge interdisciplinary fields such as management science, financial engineering, automatic control, and machine learning. The forum specially invited Professor Li Min, Associate Professor Chen Hongqiao, Assistant Professor Huang Weihuan, Assistant Professor Liu Yan, Associate Researcher Wu Yuwei, and Associate Researcher Zhu Xiaohan from the School of Management and Engineering, Nanjing University to serve as review guests. Innovatively, this forum set up two parallel sessions, with Professor Li Min and Associate Professor Chen Hongqiao hosting the opening ceremonies respectively, and providing in-depth guidance for the doctoral students' reports together with other experts.

At the forum, 19 doctoral student reporters gave wonderful academic reports on their research topics, covering research background, core issues, model methods, empirical analysis, and innovative achievements.

## Parallel Session 1:

Doctoral student Gao Yanxu reported his research work From mitigation to aggravation: The impact of digitalization on bank misconduct. This research constructed a theoretical model of banks' expected returns to examine the impact of bank digitalization on their misconduct, and conducted empirical tests using data from 173 Chinese banks. It was found that the impact of digitalization on bank violations presents a U-shaped relationship.

Doctoral student Sun Kexin reported her research work Regime-Dependent Preference Robustness: A Dynamic Approach to Multi-Period Portfolio Selection. This research proposed two different multi-period portfolio optimization problems for investors' preference uncertainty under changes in stock market mechanisms, aiming to provide personalized investment guidance for investors.

Doctoral student Tang Xinwei reported his research work Emotion or Rationality: The Impact of Data Asset Information Disclosure on IPO Pricing Efficiency. This research linked enterprise data asset information disclosure with IPO pricing efficiency, revealing the impact of enterprise data asset information disclosure on investors' decision-making behavior.

Doctoral student Wu Yifei reported his research work Work Harder, Alpha Harder: Work Intensity and Stock Returns. The research proposed an indicator to measure corporate work intensity based on large language model technology and explored the pricing ability of this indicator in the stock market and its mechanism.

Doctoral student Xie Boyi reported his research work Robust Preference Learning for Sorting Problems with Multiple Potentially Non-Monotonic Attributes. In the management context of multi-attribute ordinal classification, this research focused on the robust preference learning problem affected by inconsistent criteria and potential non-monotonic attributes, and proposed a fast and effective solution algorithm for large-scale training samples.

Doctoral student Xu Haotian reported his research work Does Cross-Ethnic Collaboration Fuel Breakthrough Invention? Evidence from Global. The research confirmed that cross-ethnic collaboration can significantly increase the probability of breakthrough innovation, but this advantage is inhibited within multinational enterprises; technological complexity plays an intermediary role, while organizational path dependence weakens the collaboration effect.

Doctoral student Yu Naichang reported his research work Option-Implied Fees and Optimal Delta Hedging under Short-Sale Constraints: Evidence from China. This research constructed an implied fee rate from the deviation of the put-call parity to measure the short-selling cost of the underlying asset, embedded it into various minimum variance Delta hedging models, and tested the improvement effect using data from Chinese and American option markets.

Doctoral student Zhang Ke reported his research work Developing Futures Trading Strategies Based on Level-2 Information and Machine Learning Models. This research captures the short-term asset price change direction by mining information in the futures Level-2 order book, constructs a combined model based on linear and non-linear models, and conducts research around the short-term price prediction and trading strategy construction of major Chinese commodity futures.

Doctoral student Zhang Yunmiao reported her research work Following the Wind: Corporate Connections and Momentum Spillover Effects—Evidence from Financial Social Media. Based on NLP/LLM to extract the corporate connection network from financial social media, this paper studies how corporate connections based on social media affect the momentum spillover effect and its mechanism.

Doctoral student Zhu Ziyang reported his research work Volatility Forecasting of Chinese Precious Metal Futures: A Time-Model Dimension Combination of GBRT and Tree-SHAP. Its innovation is mainly at the methodological level, that is, adopting an interpretable ensemble learning framework to effectively improve the prediction and interpretation ability of the current volatility of Chinese precious metal futures, providing a further model paradigm and comparison benchmark for subsequent research in the field.

## Parallel Session 2:

Doctoral student Cao Jiashuo reported his research work Three-Way Decision-Guided Hierarchical Reinforcement Learning for High-Frequency Trading. Aiming at the blind exploration dilemma faced by reinforcement learning agents in uncertain market conditions in high-frequency trading, this research proposes a new decision-making framework integrating three-way decision (3WD) and hierarchical reinforcement learning (HRL).

Doctoral student Fu Huiqiao reported her research work Learning Diverse Human-Like Driving Styles for Closed-Loop Autonomous Driving. This research proposes a human-like behavior learning method to generate diverse driving behaviors through semi-supervised diffusion adversarial imitation learning, further improving the safety and comfort of vehicles in complex traffic scenarios and promoting the intelligent development of autonomous driving technology.

Doctoral student Guo Chaoran reported his research work How to Identify High-value Innovations? A Perspective from Construction Engineering Patents' Knowledge Graphs. This research integrates patent text and network relationship features into machine learning models to improve the accuracy of predicting high-value innovations in the construction engineering field.

Doctoral student Guo Zhengyu reported his research work The Impact of BIM Technology on Project Cooperation Efficiency. This research analyzes the impact of the application of BIM technology in construction projects on project cooperation, revealing the impact path and influencing factors.

Doctoral student Li Yuanbai reported his research work Decision-Making in Wargames: An E-CARGO Perspective. Aiming at the problems of long training time, difficult convergence, and difficult multi-agent collaboration of traditional deep reinforcement learning methods in wargames, this research proposes an enhanced wargame framework integrating role collaboration and the E-CARGO model, improving the decision-making efficiency and collaboration ability of agents.

Doctoral student Liu Jinmei reported her research work Scalable In-Context Q-Learning (S-ICQL). By integrating an innovative framework of dynamic programming and world modeling, this research solves the challenges of existing in-context reinforcement learning (ICRL) methods in learning from suboptimal trajectories and achieving accurate reasoning, improving the efficient reward maximization and task generalization ability of ICRL in decision-making domains.

Doctoral student Miao Ziyang reported his research work Can Green Low-Carbon Transition Promote Win-Win Development?—A Test of the Porter Hypothesis in China's Transportation Industry. Aiming at the transportation industry under the constraints of green low-carbon transition policies, this research constructs a factor decomposition model for the evolution of green total factor productivity and designs a test standard for the Porter hypothesis, providing effective evaluation and policy suggestions for the green low-carbon transition effect of China's and regional transportation industries.

Doctoral student Wu Xuefei reported her work Interpretable Multi-Agent Reinforcement Learning. This research aims to solve the black box problem of traditional multi-agent reinforcement learning models, making the group decision-making behavior of agents easy to be understood and trusted by humans.

Doctoral student Zhang Xiaoyu reported her research work Dual-Channel Heterogeneous Graph Neural Network for Automatic Algorithm Recommendation. Aiming at the performance bottleneck in automatic algorithm selection, this research proposes a dual-channel heterogeneous graph neural network framework. The framework integrates semantics and meta-features, captures rich interaction patterns, and improves representation consistency through dual-channel characterization and contrastive learning.

After each doctoral student's report, the participating experts gave profound and pertinent comments and highly targeted guidance. Teachers and students present actively spoke, conducting multiple rounds of high-quality Q&A interactions with the reporters. The on-site academic discussion atmosphere was enthusiastic, providing a valuable opportunity for promoting interdisciplinary integration and innovative research.

After a fair evaluation by the forum review committee, Yu Naichang, a 2022 doctoral student, and Fu Huiqiao, a 2022 doctoral student, were finally selected as the winners of the Excellent Report Award of this forum. This forum concluded successfully in a strong atmosphere of academic exchange and ideological collision.