The Fin AI

Developing holistic benchmarks incorporating instruction-based, multi-turn, multi-modal, multi-lingual, Retrieval-Augmented Generation (RAG) based, and agent-based evaluations to accurately assess financial LLM performance.

Papers:

1. The FinBen: An Holistic Financial Benchmark for Large Language Models

2. PIXIU: A Large Language Model, Instruction Data and Evaluation Benchmark for Finance

3. Dólares or Dollars? Unraveling the Bilingual Prowess of Financial LLMs Between Spanish and English 

4. No Language is an Island: Unifying Chinese and English in Financial Large Language Models, Instruction Data, and Benchmarks

5. Empowering Many, Biasing a Few: Generalist Credit Scoring through Large Language Models

6. The Wall Street Neophyte: A Zero-Shot Analysis of ChatGPT Over MultiModal Stock Movement Prediction Challenges

7. AuditWen:An Open-Source Large Language Model for Audit

Project: https://github.com/The-FinAI/PIXIU 

Releasing a series of open-source financial LLMs including 1.5B, 8B, and 70B parameter models along with their checkpoints and training scripts, facilitating accessibility and innovation in model development.

Papers:
1. Open-FinLLMs: Open Multimodal Large Language Models for Financial Applications

Building a multi-agent collaborative system equipped with human-like cognitive capabilities and quantified dynamic risk profiles, designed to optimize the quality of sequential, multi-turn financial decision-making processes. This innovative system leverages multi-modal financial data inputs, forging a robust link between intelligent system design and behavioral finance principles. It is versatile enough to be employed in a wide range of financial tasks.

Papers:

1. FinCon: A Synthesized LLM Multi-Agent System with Conceptual Verbal Reinforcement for Enhanced Financial Decision Making

Creating a federated learning-based application framework for reliable and trustworthy financial LLM applications, emphasizing privacy and security in financial AI deployments.

Papers:

1. [2408.02927] HARMONIC: Harnessing LLMs for Tabular Data Synthesis and Privacy Protection (arxiv.org)