Christoph Sandrock
Papers
1
Total Citations
4
H-Index
1
About
Christoph Sandrock is a researcher focused on the intersection of machine learning and human computation, with a particular emphasis on improving data quality through uncertainty quantification. His key research area centers on leveraging annotator confidence—not just their labels—to enhance the reliability of training data. In his most-cited work, "Combining Self-reported Confidences from Uncertain Annotators to Improve Label Quality" (2019), Sandrock introduced a novel framework that treats annotators' self-assessed confidence as a valuable signal, rather than discarding it. This approach addresses a critical challenge in crowdsourced labeling: varying levels of annotator expertise and reliability. By modeling these confidences, his method demonstrably improves label accuracy over traditional majority-vote or single-annotator baselines. While his citation count (4) reflects a nascent stage of impact, the work’s conceptual contribution has been recognized for its practical relevance in domains like medical imaging and natural language processing, where noisy labels are pervasive. Sandrock’s research offers a pragmatic path toward more robust supervised learning systems, making him a notable voice in the growing field of human-in-the-loop machine learning.
Research Focus
Key Achievements
Top Papers
- 1