We are excited to launch LinqAlpha AI Lab, a research initiative dedicated to advancing Alpha Intelligence—the ability of AI systems to understand financial information, identify meaningful market signals, and translate them into reliable investment insights and decisions. We collaborate with researchers across leading financial institutions, AI organizations, and universities worldwide, and validate our findings through rigorous peer review at leading academic conferences and journals. At LinqAlpha AI Lab, we pursue this mission through three closely connected research directions.

Understanding Alpha Intelligence examines how foundation models interpret financial information and reason about markets. We study their investment biases, uncertainty, reasoning processes, and internal representations to understand when their conclusions can be trusted, how their behavior can be controlled, and whether their decisions are aligned with investor objectives.

• Your AI, Not Your View: The Bias of LLMs in Investment Analysis (ICAIF’25, https://dl.acm.org/doi/full/10.1145/3768292.3770375)

• Evaluating LLMs in Finance Requires Explicit Bias Consideration (ICML’26, https://arxiv.org/abs/2602.14233)

• When Summaries Distort Decisions: Information Fidelity in LLM-Compressed Financial Analysis (https://arxiv.org/abs/2606.29251)

Measuring Alpha Intelligence goes beyond general-purpose benchmarks, which cannot fully determine whether an AI system is effective in a company's specific business context. We develop evaluation frameworks tailored to each organization's data, workflows, objectives, and risk constraints, assessing both final outputs and the evidence, reasoning, and decisions behind them.

• Linq-Embed-Mistral Technical Report (https://arxiv.org/abs/2412.03223)

• FinDER: Financial Dataset for Question Answering and Evaluating Retrieval-Augmented Generation (ICAIF’25, https://dl.acm.org/doi/full/10.1145/3768292.3770361)

• FinAgentBench: A Benchmark Dataset for Agentic Retrieval in Financial Question Answering (ICAIF’25, https://dl.acm.org/doi/full/10.1145/3768292.3770362)

Building Alpha Intelligence translates these insights into models and systems for real-world investment workflows. Our research spans financial forecasting, alternative data, semantic asset representations, multi-agent collaboration, portfolio decision-making, and robust agentic investment systems.

• From Text to Alpha: Can LLMs Track Evolving Signals in Corporate Disclosures? (ICLR’26 FinAI Workshop, https://arxiv.org/abs/2510.03195)

• Forecasting Future Language: Context Design for Mention Markets (ICLR’26 FinAI Workshop, https://arxiv.org/abs/2602.21229)

• LLM as a Risk Manager: LLM Semantic Filtering for Lead–Lag Trading in Prediction Markets (ACL’26, https://aclanthology.org/2026.acl-industry.68/)

Beyond our research, LinqAlpha AI Lab researchers actively contribute to building the academic community around financial AI by organizing major workshops and forums that bring together researchers and practitioners from finance and AI. These include the ACM ICAIF AI for Finance Symposium (2024–2026), the ICLR Workshop on Advances in Financial AI (2025–2026), the NeurIPS Workshop on Generative AI in Finance (2025), the EMNLP Workshop on Financial Technology and Natural Language Processing (FinNLP, 2026), and KDD Finance Day (2026). Through these initiatives, we aim to continue helping shape the academic conversation around financial AI, foster collaboration across disciplines and institutions, and identify important research challenges for the field.