English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

As deep learning predictive models become an integral part of a largespectrum of precision agricultural systems, a barrier to the adoption of suchautomated solutions is the lack of user trust in these highly complex, opaqueand uncertain models. Indeed, deep neural networks are not equipped with anyexplicit guarantees that can be used to certify the system's performance,especially in highly varying uncontrolled environments such as the onestypically faced in computer vision for agriculture.Fortunately, certain methodsdeveloped in other communities can prove to be important for agriculturalapplications. This article presents the conformal prediction framework thatprovides valid statistical guarantees on the predictive performance of anyblack box prediction machine, with almost no assumptions, applied to theproblem of deep visual classification of weeds and crops in real-worldconditions. The framework is exposed with a focus on its practical aspects andspecial attention accorded to the Adaptive Prediction Sets (APS) approach thatdelivers marginal guarantees on the model's coverage. Marginal results are thenshown to be insufficient to guarantee performance on all groups of individualsin the population as characterized by their environmental and pedo-climaticauxiliary data gathered during image acquisition.To tackle this shortcoming,group-conditional conformal approaches are presented: the ''classical'' methodthat consists of iteratively applying the APS procedure on all groups, and aproposed elegant reformulation and implementation of the procedure usingquantile regression on group membership indicators. Empirical results showingthe validity of the proposed approach are presented and compared to themarginal APS then discussed.

Group-Conditional Conformal Prediction via Quantile Regression Calibration for Crop and Weed Classification