Emerging trends: General fine-tuning (gft). (23rd July 2022)

• Record Type: Journal Article
• Authors: Ward Church, Kenneth; Cai, Xingyu; Ying, Yibiao; Chen, Zeyu; Xun, Guangxu; Bian, Yuchen
• Is Part Of: Natural language engineering. Volume 28:Number 4(2022)
• Page Start: 519
• Page End: 535
• Publication Date: 2022-07-23
• Subjects: Regression -- Classification -- Fine-tuning -- Inference -- Deep nets -- Fit -- Predict; Natural language processing (Computer science) -- Periodicals; Software engineering -- Periodicals; 006.35
• DOI: 10.1017/S1351324922000237
• Languages: English
• ISSNs: 1351-3249

Abstract

Abstract: This paper describes gft (general fine-tuning), a little language for deep nets, introduced at an ACL-2022 tutorial. gft makes deep nets accessible to a broad audience including non-programmers. It is standard practice in many fields to use statistics packages such as R. One should not need to know how to program in order to fit a regression or classification model and to use the model to make predictions for novel inputs. With gft, fine-tuning and inference are similar to fit and predict in regression and classification. gft demystifies deep nets; no one would suggest that regression-like methods are "intelligent."