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

Recent work for image captioning mainly followed an extract-then-generateparadigm, pre-extracting a sequence of object-based features and thenformulating image captioning as a single sequence-to-sequence task. Althoughpromising, we observed two problems in generated captions: 1) contentinconsistency where models would generate contradicting facts; 2) notinformative enough where models would miss parts of important information. Froma causal perspective, the reason is that models have captured spuriousstatistical correlations between visual features and certain expressions (e.g.,visual features of "long hair" and "woman"). In this paper, we propose adependent multi-task learning framework with the causal intervention (DMTCI).Firstly, we involve an intermediate task, bag-of-categories generation, beforethe final task, image captioning. The intermediate task would help the modelbetter understand the visual features and thus alleviate the contentinconsistency problem. Secondly, we apply Pearl's do-calculus on the model,cutting off the link between the visual features and possible confounders andthus letting models focus on the causal visual features. Specifically, thehigh-frequency concept set is considered as the proxy confounders where thereal confounders are inferred in the continuous space. Finally, we use amulti-agent reinforcement learning (MARL) strategy to enable end-to-endtraining and reduce the inter-task error accumulations. The extensiveexperiments show that our model outperforms the baseline models and achievescompetitive performance with state-of-the-art models.

Dependent Multi-Task Learning with Causal Intervention for Image Captioning