Victoria's Blog

The SpanRuler component of spaCy allows you to create rules to recognize spans or entities within your data. Lj and I created a spaCy project to showcase the functionality of the SpanRuler within a NER pipeline, but when we didn't see the improvement we were looking for in the initial pipeline evaluation, I looked into the data and found some inconsistencies in the annotations. This led me to go back and create a Prodigy workflow to relabel data to get more consistent annotations. Machine learning is rarely a linear process that magically produces results, and iterating between your models and your data will ensure a solid foundation to build your custom ML solutions on. The combination of machine learning with rule-based approaches is a synergy that is often overlooked. However, there are a lot of benefits to creating patterns to recognize your data of interest. It can help speed up the annotation process, allow you to better understand your data, and even improve your pipeline.