On May 14, 2025, the Quant Talk Hong Kong 2025, co-hosted by DolphinDB and Datago, was successfully held in Central, Hong Kong. The event brought together professionals from leading financial institutions, tech firms, and academic research teams to explore the latest advancements in quantitative investing and financial technology.
AI Meets Finance: Navigating the Challenges of LLMs
The event opened with a keynote by Professor Michael Zhang, founder of Super Quantum Fund, who addressed the limitations and risks of using large language models (LLMs) in financial markets. His talk, “LLMs’s 7 Pitfalls in Finance”, highlighted key challenges such as data noise, emotional volatility, optimization complexity, and timing lag. He urged practitioners to combine AI tools with deep financial expertise to avoid over-reliance on algorithmic outputs.
Transforming Financial Text into Intelligence
Professor Yi Yang, founder of AbaciNLP, presented on building domain-specific text embedding models for finance. By comparing general-purpose models with finance-trained models, he demonstrated how fine-grained financial expressions—such as “growth” vs. “growth significantly”—can be more accurately interpreted to support better investment decisions. His team’s work on data augmentation and benchmark-driven model training is pushing the frontier of natural language processing in finance.
Building the Infrastructure for Real-Time Quant Strategies
Yangchun Chu, co-founder and COO of DolphinDB, showcased the company’s unified real-time computing platform that supports high-performance tasks in research, trading, and risk management. The platform integrates stream and batch processing, order book engine, and new solutions like Swordfish (a low-latency in-memory C++ database) and Shark (a GPU-accelerated computation engine), aiming to accelerate and scale complex financial strategies.
Decoding Retail Sentiment in Asia-Pacific Markets
Dr. Neal Long, co-founder of Datago, presented insights into individual investor behavior in the Asia-Pacific region. By analyzing sentiment, attention, and opinion divergence from local news and social media, Datago’s models reveal correlations between retail investor emotions and future market performance across Japan, South Korea, China, and Hong Kong. These insights empower institutions to better anticipate regional market shifts.
Looking Ahead
Quant Talk Hong Kong 2025 served as a high-impact platform for thought leadership and cross-disciplinary exchange. From the challenges of applying AI in finance to breakthroughs in NLP, real-time data infrastructure, and behavioral analytics, each presentation offered deep, practical insights into the evolving landscape of quantitative finance. The event underscored the importance of marrying advanced technology with financial acumen to create smarter, more adaptive investment strategies for the future.