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 times have witnessed an increasing number of applications of deepneural networks towards solving tasks that require superior cognitiveabilities, e.g., playing Go, generating art, ChatGPT, etc. Such a dramaticprogress raises the question: how generalizable are neural networks in solvingproblems that demand broad skills? To answer this question, we propose SMART: aSimple Multimodal Algorithmic Reasoning Task and the associated SMART-101dataset, for evaluating the abstraction, deduction, and generalizationabilities of neural networks in solving visuo-linguistic puzzles designedspecifically for children in the 6--8 age group. Our dataset consists of 101unique puzzles; each puzzle comprises a picture and a question, and theirsolution needs a mix of several elementary skills, including arithmetic,algebra, and spatial reasoning, among others. To scale our dataset towardstraining deep neural networks, we programmatically generate entirely newinstances for each puzzle, while retaining their solution algorithm. Tobenchmark performances on SMART-101, we propose a vision and languagemeta-learning model using varied state-of-the-art backbones. Our experimentsreveal that while powerful deep models offer reasonable performances on puzzlesin a supervised setting, they are not better than random accuracy when analyzedfor generalization. We also evaluate the recent ChatGPT and other largelanguage models on a subset of SMART-101 and find that while these models showconvincing reasoning abilities, the answers are often incorrect.

Are Deep Neural Networks SMARTer than Second Graders?