# Partial Or Complete, That's The Question

## TL;DR

This paper challenges the belief that complete annotations are necessary for structured learning, demonstrating that partial annotations can sometimes be more effective by leveraging interdependent variables within structures.

## Contribution

It introduces an information theoretic framework to evaluate partial versus complete annotations and shows that partial annotations can outperform complete ones in certain structured learning tasks.

## Key findings

- Partial annotations can sometimes outperform complete annotations.
- Interdependent variables within structures enable effective learning from partial data.
- The proposed framework guides annotation strategies for structured learning.

## Abstract

For many structured learning tasks, the data annotation process is complex and costly. Existing annotation schemes usually aim at acquiring completely annotated structures, under the common perception that partial structures are of low quality and could hurt the learning process. This paper questions this common perception, motivated by the fact that structures consist of interdependent sets of variables. Thus, given a fixed budget, partly annotating each structure may provide the same level of supervision, while allowing for more structures to be annotated. We provide an information theoretic formulation for this perspective and use it, in the context of three diverse structured learning tasks, to show that learning from partial structures can sometimes outperform learning from complete ones. Our findings may provide important insights into structured data annotation schemes and could support progress in learning protocols for structured tasks.