Kunpeng Zhang Directory Page

Earnings conference calls are valuable venues for business communication. Empirical research has shown that the content of earnings calls contains predictive signals about future market risks, which has motivated a line of computational studies that utilize earnings transcripts for financial forecasting tasks. However, earnings call transcripts are typically very long, and the predictive signals within them are often sparsely distributed across different sections of the transcript. As a result, existing computational methods often fail to capture the essential information within the transcript that is relevant to market risks. In this work, we design a novel method to model earnings transcripts as a conversational graph where graph nodes represent discussed topics and graph edges indicate the similarity between topics. By doing so, we aim to explicitly model what is discussed (i.e., topical content), how it is discussed (e.g., cross-referencing or newly introduced topics), and in what manner it is discussed (e.g., sentiment and other linguistic features) within the transcript. We then leverage a graph neural network to learn transcript-level representations for financial risk forecasting. Experimental results show that the proposed method significantly reduces risk forecasting errors, demonstrating its capability of modeling earnings call transcripts. Moreover, this predictive power holds even after considering the firm’s fundamentals, and the performance gain over baseline models continues to grow as transcript length increases. The interpretability analysis shows that the proposed method identifies cross-referencing and newly introduced topics as influential for risk prediction. Moreover, the model tends to associate transcripts with a higher number of new topics in the Q&A section, more cross-referencing discussions, and more positive sentiment with lower predicted financial risks. This work contributes methodologically by proposing a novel predictive approach specifically tailored to the domain-specific challenge of transcript-based risk forecasting. We also discuss key design insights and implications.

Yi Yang (HKUST), Yixuan Tang (HKUST), Yangyang Fan (HK PolyU), and Kunpeng Zhang (UMD)

• AI
• FinTech

This project will focus on Baltimore City and/or Prince George’s County, with the final study site selected in consultation with researchers and community partners. The project’s central contribution will be the integration of two complementary geospatial frameworks: risk terrain and resource terrain. The risk terrain layer will identify places where firearm violence risk may be elevated based on spatial, environmental, social, and temporal indicators, such as prior shootings, vacant properties, limited lighting, transportation corridors, housing instability, alcohol outlets, commercial activity nodes, and other built-environment or neighborhood conditions (Brantingham & Brantingham, 1995; Caplan & Kennedy, 2016; Kennedy et al., 2011). The resource terrain layer will identify places where prevention capacity already exists or could be strengthened, including community violence intervention programs, schools, libraries, recreation centers, faith institutions, small businesses, health providers, youth-serving programs, neighborhood associations, and trusted informal leaders.

Principal Investigator: Kunpeng Zhang

• Resilience and risk/crisis management
• Policy innovation and strategic foresight
• Artificial intelligence (AI)

This paper evaluates existing approaches to identifying artificial intelligence (AI)-related patents and introduces a novel, scalable framework to improve classification performance. Motivated by the growing reliance on patent data in innovation research, we assess widely used methods and document their substantial performance limitations. To address these challenges, we develop a CPC-informed label-refinement framework inspired by positive-unlabeled (PU) learning to construct high-quality training data. Using this refined dataset, we train a range of machine learning, deep learning, and transformer-based models. Our models substantially outperform existing approaches, including those employed by the USPTO. We further demonstrate the empirical value of more accurate AI patent identification through two applications. First, we find that the release of ChatGPT increased the market valuation of AI patents. Second, we show that firms increased the allocation of innovative effort toward AI technologies following its release. To facilitate future research, we release our training data, source code, and a novel dataset of patent-level predictions (AIPat), which will be continuously updated to reflect the evolving nature of AI innovation.

Tianjun Wu (Nanjing University), 
Chao Min (Nanjing University), 
Waverly W. Ding (University of Maryland), 
Guolong Wang (University of International Business and Economics), 
Kunpeng Zhang (University of Maryland)

• Corporate strategy and global competitiveness
• Creativity, innovation and organizational change
• Policy innovation and strategic foresight
• Corporate and consumer activism
• Artificial intelligence (AI)
• Generative machine learning, data analytics, data fusion, and personalization