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

We present a debiased dataset for the Person-centric Visual Grounding (PCVG)task first proposed by Cui et al. (2021) in the Who's Waldo dataset. Given animage and a caption, PCVG requires pairing up a person's name mentioned in acaption with a bounding box that points to the person in the image. We findthat the original Who's Waldo dataset compiled for this task contains a largenumber of biased samples that are solvable simply by heuristic methods; forinstance, in many cases the first name in the sentence corresponds to thelargest bounding box, or the sequence of names in the sentence corresponds toan exact left-to-right order in the image. Naturally, models trained on thesebiased data lead to over-estimation of performance on the benchmark. To enforcemodels being correct for the correct reasons, we design automated tools tofilter and debias the original dataset by ruling out all examples ofinsufficient context, such as those with no verb or with a long chain ofconjunct names in their captions. Our experiments show that our new sub-sampleddataset contains less bias with much lowered heuristic performances and widenedgaps between heuristic and supervised methods. We also demonstrate the samebenchmark model trained on our debiased training set outperforms that trainedon the original biased (and larger) training set on our debiased test set. Weargue our debiased dataset offers the PCVG task a more practical baseline forreliable benchmarking and future improvements.

To Find Waldo You Need Contextual Cues: Debiasing Who's Waldo