Twitter | Search | |
Aaron Bradley
Knowledge Graph Strategist . Linked data enthusiast, search veteran, compulsive categorizer. I like to cook!
22,609
Tweets
1,518
Following
7,980
Followers
Tweets
Aaron Bradley 6h
Map services USPs over time. MapQuest, 1996 - Show me 100 Elm St. Google Maps, 2005 - Show me how to get from my house to 100 Elm St. Amazon, 2019 - Show me how to get from my house to the mail room at 100 Elm St.
Reply Retweet Like
Aaron Bradley 6h
Replying to @aaranged
A two-way data flow is also in scope here (emphasis mine): "Provide a rich interface and services for application [sic] to consume *and contribute to* the graph."
Reply Retweet Like
Aaron Bradley 6h
This job posting from Microsoft provides some clues about where Satori is going - and by extension what we might expect to see in Bing. The extraction and curation of information "from data sources in public domains" is especially interesting.
Reply Retweet Like
Aaron Bradley 7h
Replying to @EmekaOkoye
AFAIK yes, not as detailed - but you tell me :) Wine: / Pizza: (1) (2)
Reply Retweet Like
Aaron Bradley retweeted
ML Papers Explained | A.I. Socratic Circles 11h
KGAT: Knowledge Graph Attention Network for Recommendation
Reply Retweet Like
Aaron Bradley 9h
Replying to @juansequeda
Juan that did, indeed, remind me of a representation :)
Reply Retweet Like
Aaron Bradley 9h
Replying to @ChungFeiWu
Ha, not precisely - but you may find this of interest/use
Reply Retweet Like
Aaron Bradley 9h
Replying to @sf433
Ha ... or graph paper.
Reply Retweet Like
Aaron Bradley 9h
An introduction to linked data in three acts, featuring , and >
Reply Retweet Like
Aaron Bradley 9h
Replying to @sf433
Ghosting would be kind of a blank node....
Reply Retweet Like
Aaron Bradley 9h
Replying to @octodude
I myself like a nice glass (or two) of wine with my pizza...
Reply Retweet Like
Aaron Bradley 9h
Replying to @octodude
Reply Retweet Like
Aaron Bradley 9h
Replying to @ChungFeiWu
Finally, note that search engines aren't the only .org consumers in town. Pinterest, for example, uses .org/Recipe - but (because recipe rich pins aren't the same as recipe rich results) their constraints are going to be different from Google's.
Reply Retweet Like
Aaron Bradley 9h
Replying to @ChungFeiWu
But aligning with Bing, Yahoo! and Yandex on what those constraints should be isn't realistic, given that they don't have such a product.
Reply Retweet Like
Aaron Bradley 9h
Replying to @ChungFeiWu
.org/Dataset and usage constraints on that schema were necessary, but not sufficient, conditions support Dataset Search: Google had to build the actual interface, search results, etc.
Reply Retweet Like
Aaron Bradley 9h
Replying to @ChungFeiWu
WRT to "final scope" alignment is nice in theory, but practically is driven by product development. E.g. Google built a fair amount of infrastructure to facilitate Dataset Search...
Reply Retweet Like
Aaron Bradley 9h
Replying to @ChungFeiWu
2) See (1). :) Where the search engines align are on the vocabulary (which is why .org is a collaboration between the main engines), where they diverge is on: a) which schemas they use and; b) for schemas they do use, usage constraints.
Reply Retweet Like
Aaron Bradley 9h
Replying to @ChungFeiWu
1) cont'd.... For example, one engine might want to display the condition of a product in the SERPs, and so would require Product.itemCondition; one that didn't wouldn't require a value for that property - or, of course, might not make any use of the schema at all.
Reply Retweet Like
Aaron Bradley 9h
Replying to @ChungFeiWu
YW! 1) Yes. Keep in mind that the purpose for which each search engine uses a particular schema is going to vary. For Google these are leveraged largely for rich results, so the requirements are related to what data they want to be able to display in those rich results.
Reply Retweet Like
Aaron Bradley 9h
About modelling styles in ontologies - C. Maria Keet
Reply Retweet Like