Multi-Modal Financial Time-Series Retrieval Through Latent Space Projections
Published in Proceedings of the 4th ACM International Conference on AI in Finance (ICAIF 2023), 2023
Recommended citation: Bamford, T., Coletta, A., Fons, E., Gopalakrishnan, S., Vyetrenko, S., Balch, T., & Veloso, M. (2023). Multi-Modal Financial Time-Series Retrieval Through Latent Space Projections. In Proceedings of the 4th ACM International Conference on AI in Finance. https://dl.acm.org/doi/10.1145/3604237.3626880
Abstract
Cross-modal information retrieval between continuous time-series data and unstructured financial news reports is a crucial open challenge. We develop a shared latent embedding space using contrastive representation learning, enabling practitioners to search historical market anomalies using natural language queries and vice versa.
Recommended citation: Bamford, T., Coletta, A., Fons, E., Gopalakrishnan, S., Vyetrenko, S., Balch, T., & Veloso, M. (2023). Multi-Modal Financial Time-Series Retrieval Through Latent Space Projections. In Proceedings of the 4th ACM International Conference on AI in Finance.