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

Influence estimation analyzes how changes to the training data can lead todifferent model predictions; this analysis can help us better understand thesepredictions, the models making those predictions, and the data sets they'retrained on. However, most influence-estimation techniques are designed for deeplearning models with continuous parameters. Gradient-boosted decision trees(GBDTs) are a powerful and widely-used class of models; however, these modelsare black boxes with opaque decision-making processes. In the pursuit of betterunderstanding GBDT predictions and generally improving these models, we adaptrecent and popular influence-estimation methods designed for deep learningmodels to GBDTs. Specifically, we adapt representer-point methods and TracIn,denoting our new methods TREX and BoostIn, respectively; source code isavailable at https://github.com/jjbrophy47/tree_influence. We compare thesemethods to LeafInfluence and other baselines using 5 different evaluationmeasures on 22 real-world data sets with 4 popular GBDT implementations. Theseexperiments give us a comprehensive overview of how different approaches toinfluence estimation work in GBDT models. We find BoostIn is an efficientinfluence-estimation method for GBDTs that performs equally well or better thanexisting work while being four orders of magnitude faster. Our evaluation alsosuggests the gold-standard approach of leave-one-out (LOO) retrainingconsistently identifies the single-most influential training example butperforms poorly at finding the most influential set of training examples for agiven target prediction.

Adapting and Evaluating Influence-Estimation Methods for Gradient-Boosted Decision Trees