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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Modern graph-based syntactic dependency parsers operate by predicting, foreach token within a sentence, a probability distribution over its possiblesyntactic heads (i.e., all other tokens) and then extracting a maximum spanningtree from the resulting log-probabilities. Nowadays, virtually all such parsersutilize deep neural networks and may thus be susceptible to miscalibration (inparticular, overconfident predictions). In this paper, we prove thattemperature scaling, a popular technique for post-hoc calibration of neuralnetworks, cannot change the output of the aforementioned procedure. We concludethat other techniques are needed to tackle miscalibration in graph-baseddependency parsers in a way that improves parsing accuracy.

Maximum Spanning Trees Are Invariant to Temperature Scaling in Graph-based Dependency Parsing