This paper introduces "Abstract Syntax Networks," a model designed to convert natural language descriptions into executable code (like Python or SQL) by predicting the structure of the code directly. Source: ACL Anthology P17-1105
The code often refers to specific academic papers in the ACL Anthology (a digital archive of conference papers on Natural Language Processing). Depending on which conference year you are looking for, here are three high-quality "1105" papers you can explore: 1. Abstract Syntax Networks for Code Generation (2017) Topic: Artificial Intelligence / Programming 1105mp4
It focuses on how computers can understand "gapped" sentences—where words are omitted but understood (e.g., "Paul likes coffee and Mary tea"). The authors propose methods to help AI fill in these missing pieces. Source: ACL Anthology N18-1105 This paper introduces "Abstract Syntax Networks," a model
3. Text Categorization by Learning Predominant Sense of Words (2019) Machine Learning / NLP Abstract Syntax Networks for Code Generation (2017) Topic:
This paper uses a Transformer-based model to categorize documents more accurately by figuring out the specific meaning of a word based on the domain it's used in (e.g., "bank" in finance vs. "bank" in geography). Source: ACL Anthology P19-1105