Introduction
Artificial intelligence (AI) and information retrieval (IR) systems and techniques have been widely adopted in financial services to tackle various tasks, such as information retrieval from business documents, retrieval from non-textual content like tables and graphs, recommending financial products and services to customers, providing decision support for investment practices, automating of due diligence protocols, detecting fraudulent transactions, financial sentiment analysis on social media, and understanding Environmental, Social and Governance (ESG) impact on investment practices.
Knowledge from IR systems can help augment human intelligence. However, discovering and extracting the knowledge conveyed inside unstructured financial data, like SEC filings, prospectuses, business reports, and other enterprise documents are extremely challenging due to the massive volume of data, large variation in the data format, low signal-to-noise ratio, scarcity of expert annotated datasets, task ambiguity, hurdles regarding data integrity and privacy, robustness against domain shift, and high-performance requirements set by industry and regulatory standards. Manual extraction of knowledge is usually inefficient, error-prone, and inconsistent, so it is one of the key technical bottlenecks for financial services companies to accelerate their operating productivity. These challenges and issues call for robust artificial intelligence, information retrieval, and machine learning algorithms and systems to help. The automated processing of unstructured data to discover knowledge from complex financial documents requires bringing together a suite of techniques such as natural language processing, information retrieval, semantic analysis, and complex reasoning. In addition, how knowledge is captured and represented, synthesized across diverse sources, and used within AI systems, is crucial to developing effective solutions in financial services.
Furthermore, based on the reflections and feedback from our past KDF workshops, the 2023 workshop is particularly interested in multi-modal understanding of financial documents, retrieving and reasoning over tabular data within financial documents, and financial domain-specific representation learning. The workshop will be composed of three components: invited talks, paper presentations, along with a shared task competition. We cordially welcome researchers, practitioners, and students from academic and industrial communities who are interested in the topics to participate and/or submit their original work. The topics and submission requirements can be found in call for papers.
Registration
KDF ‘23 has been closed. Please stay tuned for KDF ‘24.
Important Dates (Anywhere on Earth)
| Abstract Submission Deadline (Optional) | Monday | May 1, 2023 (AOE) |
| Paper Submission Deadline (extended) | Sunday | May 21, 2023 (AOE) |
| Submission Notification Date | Friday | May 31, 2023 |
| Shared Task End Date | Monday | June 12, 2023 |
| Shared Task Paper Submission Deadline | Monday | June 30, 2023 |
| Workshop | Thursday | July 27, 2023 |
Contact Information
We look forward to seeing you in KDF 2023! For general inquiries about KDF, please write to kdf.workshop@gmail.com.