Forma transformer model forecasts financial statements 20 quarters ahead

Researchers have developed Forma, a novel transformer-based model capable of forecasting complete financial statements up to 20 quarters into the future. Forma significantly outperforms various classical machine learning techniques, including gradient boosting and large language models, on the newly released ProForma-20Q benchmark. This benchmark requires forecasting 78 financial statement line items and is scored by change-space R^2, with Forma showing particular strength at longer forecast horizons crucial for valuation. AI

IMPACT This research demonstrates specialized models can outperform generalist LLMs for complex forecasting tasks, potentially impacting financial analysis tools.

RANK_REASON The cluster contains an academic paper detailing a new model and benchmark for financial forecasting. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.LG →

• arXiv
• gradient boosting
• large-language models
• machine learning
• ProForma-20Q
• Transformer++

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →