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

Visual Reasoning (AVR) comprises a wide selection of variousproblems similar to those used in human IQ tests. Recent years have broughtdynamic progress in solving particular AVR tasks, however, in the contemporaryliterature AVR problems are largely dealt with in isolation, leading to highlyspecialized task-specific methods. With the aim of developing universallearning systems in the AVR domain, we propose the unified model for solvingSingle-Choice Abstract visual Reasoning tasks (SCAR), capable of solvingvarious single-choice AVR tasks, without making any a priori assumptions aboutthe task structure, in particular the number and location of panels. Theproposed model relies on a novel Structure-Aware dynamic Layer (SAL), whichadapts its weights to the structure of the considered AVR problem. Experimentsconducted on Raven's Progressive Matrices, Visual Analogy Problems, and Odd OneOut problems show that SCAR (SAL-based models, in general) effectively solvesdiverse AVR tasks, and its performance is on par with the state-of-the-arttask-specific baselines. What is more, SCAR demonstrates effective knowledgereuse in multi-task and transfer learning settings. To our knowledge, this workis the first successful attempt to construct a general single-choice AVR solverrelying on self-configurable architecture and unified solving method. With thiswork we aim to stimulate and foster progress on task-independent research pathsin the AVR domain, with the long-term goal of development of a general AVRsolver.

One Self-Configurable Model to Solve Many Abstract Visual Reasoning Problems