To drive the disruptive innovation in the current time, the technology of Natural Language Processing (NLP) is up surging to a whole new level. The power to decode a large sum of information into the linguistic/multilingual genre is playing a protagonist at the business platform. With advancement in AI-powered NLP entity, OEMs are delving deep into the employment of the technology for benefitting other organizations and businesses. Rosoka Software is one such OEM NLP software company.
The company specializes in the multilingual entity and relationship extraction. Rosoka automatically finds the PEOPLE, PLACES, ORGANIZATIONS, and other highly relevant things from narrative text and connects these entities together through Predicate-Subject-Object (PSO) relationships. It also provides multivector sentiment analysis, salience determination, and geotagging of PLACE and FACILITY entities. The software is 100% pure Java that provides XML or JSON output through the Java API or REST web service endpoints.
From Inception to Innovation
Rosoka’s CEO has a background in linguistics. He studied as a computational linguist and has degrees from Illinois Central College, University of Illinois Urbana-Champaign, and Georgetown University. He has been actively working with Natural Language Processing (NLP) technologies for over 25 years. He started as a linguistic intern for a small NLP startup and have gradually moved up the food chain to founding his own NLP company: Rosoka Software. With Rosoka, the company is able to visualize the shortfalls in what was then the state-of-the-art in NLP. The company saw that there was a set of guiding principles that needed to be adhered to take the technology forward. The software needed to be purely multilingual. It had to process any language in a document without having to a priori know what language it was processing. The software had to be portable with the same results regardless of the operating system. The software needed to be able to run just as easily on Linux as it would on Windows or Mac. The software had to be scalable, not just up, but out and down. Rosoka has taken a chocolate/peanut butter approach merging CEO’s thoughts and understanding with Linguistic Theory, with his CTO’s understanding of Information Theory.
The Creator and CEO: Gregory Roberts
Gregory has always had that entrepreneurial spirit. For as long as he can remember, he wanted to run his own company. His first go at being an entrepreneur was in grade school. When he was in 5th grade, he got a Commodore 64 computer. After playing with it for a couple of months, he had an idea for a game. He didn’t quite have the coding skills to create the game, but he was able to convince a couple of his friends that it was a good idea and they could make some money selling the game. They never did make any real money with the game. However, that experience taught him that one doesn’t need to know how to do everything, he/she just needs people to believe in the vision. One of the things he does best is getting those around him to understand his vision and to take on that vision as if it was their own. He is truly grateful for his team and what they have allowed him to accomplish.
Putting NLP at the Front Table
Until the company developed Rosoka, NLP was pretty much the domain of supercomputers or massively parallel processing. Rosoka takes NLP to the edge and beyond with what is called ubiquitous NLP (uNLP). Rosoka can be loaded on a Raspberry Pi and get the same extraction results as when Rosoka is used in a multi-node Enterprise cluster. “This allows a lot of processing to be done at the endpoint or the “tip of the spear”. Technologies like “OK Google”, Siri, and Alexa require that everything is sent back to the mothership for processing. Very little, if any, processing is done at the end point. The privacy implications alone are tremendous. It’s your data, do you really want to ship it to a third party and back for processing? said Gregory”
Driving Disruption through AI
Many would answer this question with Artificial Intelligence will replace people. Gregory thinks that is a very naive answer. What AI really provides is to remove the tedious work to enable people to do what they are good at, making more intelligent decisions. This, in turn, is creating new opportunities. AI, combined with Big Data Analytics, has the ability to smooth out statistical bias by analyzing over larger and larger datasets to produce more accurate results.
Maintaining Innovative Company Culture
Rosoka Software is an OEM company. This means that other companies embed Rosoka’s software into their solutions. Gregory is genuinely amazed at how some of these companies have used Rosoka’s software in unique ways. A good example is a British company embedded Rosoka’s software in a system used to route Royal Marine helicopters through hostile environments. The mobile system collected social media geotagged within a geofenced area where the helicopter would fly from Point A to Point B. They used Rosoka to calculate the sentiment of social media to provide a likely safer route to fly.
Evaluation of Hard Work
Rosoka Software has been called, “The best NLP software that no one has heard about” by one of its customers. Rosoka has traditionally done very little marketing. The company was founded to support the US Intelligence Community and DoD markets. These are discerning customers with critical needs and their word of mouth means more to Rosoka’s business growth than traditional marketing methods. Rosoka has a 92% customer retention rate. The customers Rosoka have, they keep. Retaining customers and working with them over the long haul to be a part of their complex systems is very important. Rosoka’s customers know they are getting the best software along with the best customer service of any NLP software company. This is reflected in a Dun and Bradstreet Open Ratings Performance Evaluation in 2018 where Rosoka Software scored 94 out of 100.
Extracting Opportunities from Challenges
There are two major challenges that stand out in Gregory’s mind. First, the company is often considered “Beyond State of the Art”. Rosoka competes with companies that focus on tools that force the user to provide large amounts of hand annotated training data before their algorithms can even run. Government agencies and commercial customers write requirements for the system that locks in this level of technology without even considering the possibility of what NLP can do to solve their real business problems. Second, “Humans are always more accurate than a machine”. Regardless of whether that statement is true or not (it is not), high-value hand annotations combined with high volume automated extraction will always provide a better analytical picture in terms of breadth, depth, and scope of coverage that is just not possible with a hand annotated, proprietary system. The real value of Rosoka is the ability for software to automatically analyze large volumes of data, allowing the wetwear (people) to focus on making decisions, not doing monotonous grunt work. Rosoka provides a force multiplier to what an analyst or group of analysts can do on their own.
Big Data Industry of the Future
Commenting on the future head, Gregory said, “I see the industry changing from using solutions provided by large systems integrators on “Big Iron” that only Fortune 100 or Governments could afford to moderately priced NLP solutions like ours. Organizations that use our technology in a nimble fashion develop a clear competitive edge. Big Data Analytics is no longer just for the big guys with big budgets.”