Click screenshot thumbnail below to view Flash demo (4.7 MB).

Today's computers don't behave very much like human beings. DataSea changes that.

DataSea's technology understands meaning, context, and synonyms and puts it all together to do what you really want in one simple step. That means you can communicate with DataSea the way you would with a human being. DataSea extracts knowledge from all the information you give it and delivers and actions and answers you really want. It doesn't look up answers in a table, it derives them from the connections it finds in the entire body of data. It's a different architecture of information storage and problem solving. It "connects the dots."

DataSea's proprietary Inferencing Engine incorporates innovative reasoning capability based on neural-like networks. Its unique methods map to natural human behaviour, harnessing the power of natural language to give people an intuitive, powerful way to interact with computers and devices. No need to memorize command sets or navigate through menus and icons.

Inferencing is the ability to use human-like reasoning to "connect the dots" between different data points. Most computing today is based on lookup, statistical modeling and decision trees. This is why so much of today's computing involves navigating through menus, icons, or the frustrations of keyword-based hit lists.

Natural language is plain English. It's that simple. You can interface with DataSea just by speaking or texting the way you would with a human being, and it understands you.

DataSea fuses both structured and unstructured data sources into one network and extracts knowledge from all the information you give it, including input, and delivers and actions and answers you really want based on the connections it finds between your input and its fused knowledgebase.

Besides acquiring new information, its "neural-like" architecture allows DataSea to learn from your feedback: while any imprecise query can give a number of correct answers, DataSea allows you to select what you mean through disambiguation. The system learns from this and gives better results next time.

If you Google   "Bob, phone"   you'll get hits with those two words in them, but not necessarily his phone number. The same input to DataSea will give you his phone number, and also call logs about Bob. It shows things related to those terms, and explains the relationships. And if you ask DataSea   "Bob, phone, Berkeley"   it would give you his Berkeley phone number,   510-845-1111. Further, if you say   "call Bob and say Hi, I will be late"   it would place a call and synthesize the message for you. We refer to this as natural control.

So DataSea is like Google in that you use words that you come up with (content addressable), instead of pulldown menus. But it is more than that: DataSea gives single-step access to fused information sources.

It answers complex questions and makes mini-reports. It lets you do things like make calls and send email, all of these things it does in one step.

For example, if you want to find out what Bob's sister's phone number is, you ask
            "what is Bob's sister's phone number"

instead of this common sequence of pull-down steps from an Address Book application;
        Contacts → Search"Bob" → PersonalInfo → Notes → "Mary Jones is sister" ↵
        → Cancel → Contacts → Search"Mary" → Telephone

        or, worse, this SQL query:
                select Home_Phone, Cell_Phone from AddressBook where Name = (select Name from AddressBook ↵
                where SiblingName like 'Bob %')