Abstract | ||
---|---|---|
Model building in machine learning is an iterative process. The performance analysis and debugging step typically involves a disruptive cognitive switch from model building to error analysis, discouraging an informed approach to model building. We present ModelTracker, an interactive visualization that subsumes information contained in numerous traditional summary statistics and graphs while displaying example-level performance and enabling direct error examination and debugging. Usage analysis from machine learning practitioners building real models with ModelTracker over six months shows ModelTracker is used often and throughout model building. A controlled experiment focusing on ModelTracker's debugging capabilities shows participants prefer ModelTracker over traditional tools without a loss in model performance. |
Year | DOI | Venue |
---|---|---|
2015 | 10.1145/2702123.2702509 | CHI |
Keywords | Field | DocType |
user interfaces,interactive visualization,debugging,machine learning,performance analysis | Analysis tools,Graph,Usage analysis,Iterative and incremental development,Computer science,Model building,Interactive visualization,Human–computer interaction,Artificial intelligence,Machine learning,Debugging,Algorithmic program debugging | Conference |
Citations | PageRank | References |
57 | 2.10 | 15 |
Authors | ||
6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Saleema Amershi | 1 | 775 | 45.16 |
Max Chickering | 2 | 57 | 2.78 |
steven m drucker | 3 | 2399 | 286.15 |
Bongshin Lee | 4 | 2738 | 143.95 |
Patrice Y. Simard | 5 | 1112 | 155.00 |
Jina Suh | 6 | 178 | 10.04 |