Multilingual Symbolic Support for Low Levels of Literacy on the Web
E.A. Draffan, Mike Wald, Chaohai Ding and Russell Newman
The Problem
-
Content on the web can be:
-
Complex ≈16% over 60s have mild cognitive impairment (WHO)
-
Hard to simplify
-
Incomprehensible for those with poor literacy levels.
2

What are low levels of literacy?
“750 million adults – two-thirds of whom are women – still lack basic reading and writing skills”. The benchmark for the 86% of those from age 15 and over who “can both read and write with understanding” is based on “a short simple statement on his/her everyday life” UNESCO

Initial Strategies
-
Text simplification
-
Symbol labels require text cleaning, removal of special characters, handling of ambiguous meaning, spelling correction and extraction of parts of speech (PoS) – Natural Language Processing
Sample thanks to Dundee University AAC project consent form

Symbol to Concept Mapping
-
Concept based on the label to symbol linking ≈70% accurate.

Semantic Word Embedding
Google Spreadsheet with similarity rating

Results and Future Work
-
Image recognition needed, but the results may not always help with topic classification!
Semantic reasoning to provide related symbols based on common knowledge base improved to 85.47%
Car?

Thank You
-
This work has been part of an Alan Turing Pilot project about AI and Inclusion (https://www.turing.ac.uk/research/research-projects/ai-and-inclusion) coordinated by the Web Science Institute.

