December 13, 2010 – Digital Reasoning Systems Inc. has released their flagship solution for cloud-scale analytics in government and commercial markets – Synthesys v3.0.
BusinessWire brought this news to our attention in their article, “Digital Reasoning Announces Synthesys™ v3.0.” This recent advancement in achieving automated understanding of structured and unstructured data is very timely with the growth of data and the increasing demand for timely solutions. This need is met by Synthesys’ conversion of massive amounts of unstructured data.
“Synthesys 3.0 is the culmination of a decade of solving large scale data analytics problems for the intelligence community,” said Tim Estes, CEO and founder of Digital Reasoning Systems. “With this release we are bringing to the market unprecedented understanding of unstructured data in cloud-scale architectures with our integration of Hadoop and Cassandra.”
Interestingly, this is accomplished without the assistance of a taxonomy or ontology. They call it an “unbiased” approach. It seems like without the strong foundation of a taxonomy, the data would still have some of the disadvantages of unstructured data.
Melody K. Smith
Sponsored by Data Harmony, a unit of Access Innovations, the world leader in indexing and making content findable.
Hello Ms. Smith! I got forwarded your web site by Dave who I see has also responded. You have a great site. I’ll be adding the feed to my reader and I’ll pass the link on.
You are right,without a taxonomy the unstructured data would remain difficult to leverage. Synthesys uses an association network to produce a model from the data. The usual method of imposing a model on data works, but with limitation. Those limitations get magnified when you are working with large stores of data. For me a large store of data is something greater than 100 million messages or documents. The limitations exist at any level though.
The association network provides the same sort of information you get from a standard human generated taxonomy. It will tell you how things relate. The benefit of a mapping algorithm like an association network is that you never run into issues of which exact category to place something in.
Taxonomy has been vital to information science and will continue to be vital. The nature of it will change as information systems evolve to be more detailed, powerful and faster.
Continue the great work on your site!
Pete Mancini
Senior Consultant
Nectarine Imp LLC
Melody,
Thanks for picking up on our Synthesys press release. We agree that this is a timely topic as we are getting an increasing amount of requests to try to make sense out of “big data” – sometimes up to hundreds of millions of documents. However, I would want to clarify that we are not “anti Taxonomy” but we believe that there are times that taxonomy and ontology can introduce bias as to what words mean. Only by letting the model come from the data, it is possible to discover code-words, intentional or unintentional misspellings, hidden connections, etc. We recently posted an excellent guest blog by Pete Mancini (nectarineimpllc.com) on this very subject. I would encourage your readers to see what Pete has to say here: http://www.digitalreasoning.com/2010/blog/ontology-and-identity-in-the-digital-reasoning-model/
thanks for the discussion opportunity –
Dave Danielson
VP Marketing
Digital Reasoning Systems