Re: Facebook's new Graph Search: An endorsement of the RDF approach to healthcare data?

On 1/17/13 1:06 PM, Lin MD, Simon wrote:
> Kingsley,
>
> I tried the link you send. With 55 Billion+ live triples, the response time is very much impressive.
>
> However, I must say that I am lost on viewing the results -- the presentation is not designed to anybody outside of the RDF community!

One step at a time. We've built a deeply nested faceted browser into the 
DBMS engine. At this point in time, the UI works for anyone that's 
familiar with entity relationship model based data representation and 
entity relationship semantics.

>
> By the way, how you compare your approach to neo4j? I am planning for a research prototype that needs ontology-driven query of millions of nodes for treatment comparisons.

We produce a massively scalable DBMS with native support for Linked Data 
creation, ingestion, integration and overall management. I don't know 
the yardsticks you seek with regards to the requested comparison. How 
about looking around for some live neo4J hosted instances of Linked 
Data? That would at least enable you compare from your particular 
perspective etc..

Kingsley
>   Your inputs are appreciated!
>
> Simon
>
> ==================================================
> Simon Lin, MD | Director, Biomedical Informatics Research Center | Marshfield Clinic Research Foundation | 1000 N Oak Ave, Marshfield, WI 54449 | 715-221-7299 | Lin.Simon@mcrf.mfldclin.edu | www.marshfieldclinic.org/birc
>
>
> -----Original Message-----
> From: Kingsley Idehen [mailto:kidehen@openlinksw.com]
> Sent: Thursday, January 17, 2013 8:38 AM
> To: public-semweb-lifesci@w3.org
> Subject: Re: Facebook's new Graph Search: An endorsement of the RDF approach to healthcare data?
>
> On 1/16/13 8:12 PM, Rafael Richards wrote:
>> Yesterday Facebook announced a new feature called Graph Search.  This
>> Graph includes 1 billion people,  240 billion photos, and over 1
>> trillion connections.
>>
>> Graph search is privacy aware:  every piece of content has its own
>> specified audience. Most content is not public; you can only search
>> for content that has been shared with you.
>>
>> http://readwrite.com/2013/01/15/facebook-graph-search
>> https://newsroom.fb.com/News/562/Introducing-Graph-Search-Beta
>> https://newsroom.fb.com/Photos-and-B-Roll/4321/Graph-Search-Announcment
>>
>>
>> RDF data stores are also currently capable of  loading a trillion
>> triples ("connections"), and we have hardware such as the Cray
>> purpose built for graph analytics.
>>
>> http://www.franz.com/about/press_room/trillion-triples.lhtml
>> http://www.cray.com/Products/BigData/uRiKA.aspx
>> http://investors.cray.com/phoenix.zhtml?ID=1766098&c=98390&p=irol-newsArticle
>>
>>
>> There is also work done on a natural language query interface for RDF
>> using Cyc as the foundation ontology.
>> http://www.w3.org/2001/sw/sweo/public/UseCases/ClevelandClinic/#figure3
>>
>> While not a "success story" of RDF per se,  Graph Search (if anyone
>> knows what its actual technology is)  may be at least an endorsement
>> of the RDF-like approach to managing, repurposing, and securing data.
>>
>> Is there any reason to believe that an RDF-based system could not also:
>> 1. Enable similar storage and query as Graph Search?
>> 2. Provide similar data-atomic granular control of privacy of
>> (personal or healthcare) data similar to that of Graph Search?
>>
>>
>> Rafael
>>
>>
>> _____________________
>> Rafael Richards MD MS
> Rafael,
>
> Also note that there is a live 55 Billion+ triples instance of the LOD
> Cloud cache that showcases what's always been possible re. Linked Data,
> RDF, and machine readable Entity Relationship Semantics.
>
> Links:
>
> 1. http://lod2.openlinksw.com -- live instance
> 2. http://bit.ly/11ezuCz -- detailed instance report covering loading
> and some simple analytics
> 3. http://bit.ly/W8MYMj -- Uniprot citations overview
> 4. http://bit.ly/13HGvwX -- Bio2RDF entities associated with Genetic
> Dissorders
> 5. http://bit.ly/101tE55 -- Biopax data sources by collective attributes .
>
>
>


-- 

Regards,

Kingsley Idehen	
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen