Experimental Design for Bathymetry Editing
arXiv:2007.07495v1
Originality Synthesis-oriented
AI Analysis
This addresses a fundamental issue in machine learning for real-world labeling tasks, but it is incremental as it highlights a known limitation without proposing a new solution.
The paper tackled the problem of poor performance in machine learning due to the IID assumption in data splitting for a bathymetry editing task, finding that random splits often lead to significant deviations and suboptimal results.
We describe an application of machine learning to a real-world computer assisted labeling task. Our experimental results expose significant deviations from the IID assumption commonly used in machine learning. These results suggest that the common random split of all data into training and testing can often lead to poor performance.