EP19: AI in Finance and Symbolic AI with Atlas Wang

In this episode of The Information Bottleneck, Ravid Shwartz-Ziv and Allen Roush host Atlas Wang to discuss NeurIPS 2024, including the increasing presence of VCs, the quality of workshops versus main conferences, and the conference web app. The conversation shifts to low dimensionality in neural networks, symbolic equations, and the potential for neural networks to learn symbolic formations. They explore the practical applications of converting neural networks to symbolic forms for efficiency and interpretability, touching on the importance of unsupervised learning, dimensionality reduction, and synthetic data. The discussion further delves into the use of generative AI in high-frequency trading, the challenges of predicting market behavior, and the evolving profiles of AI researchers in finance, concluding with thoughts on work-life balance and the future of AI in the finance industry.