2023/01/31 by Harshit Joshi, Joshi, Harshit, Abishai Ebenezer +13 · 9 voices · 5 citations
Computer Science · Decision Sciences · Mathematics · #Algorithm #Artificial intelligence #Computer science #Data Quality and Management #Data deduplication #Data mining #Database #Domain (mathematical analysis) #Information Retrieval and Search Behavior #Information retrieval #Language model #Machine learning #Mathematics #Natural language processing #Similarity (geometry) #Sketch #Spreadsheets and End-User Computing #Transformer
paper · pdf · doi:10.48550/arxiv.2301.13779
published in arXiv (Cornell University) (Cornell University)
openalex publication_date 2023/01/31 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Spreadsheets are a vital tool for end-user data management. Using large language models for formula authoring assistance in these environments can be difficult, as these models are expensive to train and challenging to deploy due to their size (up to billions of parameters). We present FLAME, a transformer-based model trained exclusively on Excel formulas that leverages domain insights to achieve competitive performance while being substantially smaller (60M parameters) and training on two orders of magnitude less data. We curate a training dataset using sketch deduplication, introduce an Excel-specific formula tokenizer, and use domain-specific versions of masked span prediction and noisy auto-encoding as pre-training objectives. We evaluate FLAME on formula repair, formula completion, and similarity-based formula retrieval. FLAME can outperform much larger models, such as the Davinci (175B) and Cushman (12B) variants of Codex and CodeT5 (220M), in 10 of 14 evaluation settings for the repair and completion tasks. For formula retrieval, FLAME outperforms CodeT5, CodeBERT, and GraphCodeBERT.