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

Human capability to anticipate near future from visual observations andnon-verbal cues is essential for developing intelligent systems that need tointeract with people. Several research areas, such as human-robot interaction(HRI), assisted living or autonomous driving need to foresee future events toavoid crashes or help people. Egocentric scenarios are classic examples whereaction anticipation is applied due to their numerous applications. Suchchallenging task demands to capture and model domain's hidden structure toreduce prediction uncertainty. Since multiple actions may equally occur in thefuture, we treat action anticipation as a multi-label problem with missinglabels extending the concept of label smoothing. This idea resembles theknowledge distillation process since useful information is injected into themodel during training. We implement a multi-modal framework based on longshort-term memory (LSTM) networks to summarize past observations and makepredictions at different time steps. We perform extensive experiments onEPIC-Kitchens and EGTEA Gaze+ datasets including more than 2500 and 100 actionclasses, respectively. The experiments show that label smoothing systematicallyimproves performance of state-of-the-art models for action anticipation.

Knowledge Distillation for Action Anticipation via Label Smoothing