Exploring CONTENT CLARIFIER for AUTISM Use Cases
IBM AbilityLab™ Content Clarifier is designed to help simplify, summarize, and enhance content in order to increase comprehension for people with cognitive disabilities, the aging population, or those learning English as a second language.
I am exploring a number of IBM Watson services and IBM Technologies as components to help with Autism Use Cases / Demo's I'm building and open sourcing. I am exploring how this tool may be used along side Tone and Emotion services to help kids with Autism navigate the world.
http://contentclarifier.mybluemix.net/#/landing - as noted "Accessing to the demo:
1. If you don't have one already, create a Bluemix ID here.
2. Send an email to firstname.lastname@example.org with your Bluemix ID, requesting access."
3 minute video from IBM Accessibility:
Conversational - Simplified
Observation - good tool for "de-fancyfing" complex langauge - but did note that it added confusion if used en-masse to document - i.e. helped in 3 of 4 cases I tried, but added degree of confusion in 1 in 4. My take away is this could be a good "Help me" tool - to have an autistic user ask for a simplified or alternate view on demand - of a snippet.
this news article was condensed. did a pretty good job. two areas to watch though: 1) Attribution - in one case a Teresa May comment about Mark Carney was attributed to Teresa, not Mark. and 2) Condensing often loses speaker/attribution - so for he-said; she said articles, this can be important info
Did a nice job of the technical nuts and bolts - condensing
Did a nice job of distilling key concepts from the longer corporate message
Ultra-Mode: A second pass of simplification is performed on summarized content.
I only used the Web / UX for this tire kick - going to poke around for APIs and see about integrating on back end. Watch this space :)
USEFUL TECHNOLOGY : Content Clarifier tool has an API that is accessible for automated cognitive assistance systems - and that there are several situations where “smart simplification and summarization” would be very useful. Some good stuff here. So far I only used web/UX to test - to come: integrate with APIs for 'real time' testing.
SUMMARY IS SUPERB: For the limited tests I performed, the CONDENSED feature was most useful. There are some use cases – especially with multi-speaker or multiple sources of information exist, where it loses too much signal – but for corporate statements or technical articles, looks like a great way to distill.
FOCUSED CLARIFICATION IS BEST: Content Clarifier looks like best-fit will be on targeted utterances, rather than applying globally, or ‘always on’; The simplification feature for fancy words, when applied globally, often created more confusion. Deployed in a targeted manner, and in a way that can learn over time, is probably a sensible place to start. For example – “Help me understand that last statement” – could provide a few new ways to interpret and understand.
About this blog
This is an informal blog that explores tools, code and tricks that group members have developed to engage IBM Watson cognitive computing services - from the R Programming Language. Packages include RCURL to access Watson APIs - for services that include Natural Language Classifier and Speech to Text. THIS IS MY PERSONAL BLOG - it does not represent the views of my employer. Code is presented as 'use at your own risk' (it has lots of bugs)
Created: September 13, 2015English