Enables organizations to embrace “the Intelligence of Every ‘Thing’,” delivers enhanced speed and volumes at which data can be processed while critically improving the accuracy of results
Los Altos, CA, October 16, 2014 – Saffron Technology, the first cognitive computing platform that learns, reasons and anticipates like you and for you, today unveiled Saffron 10, the newest version of its groundbreaking Natural Intelligence Platform. Saffron 10 fuses the power of computing with brain-like intelligence, enabling organizations to more quickly focus on the knowledge that matters to them most (including things you may not think to look for), anticipate what will happen next, and optimize decisions based on that criteria.
“Cognitive systems are beginning to have a dramatic impact on enterprises providing answers, recommendations and courses of action to difficult problems,’ explains David Schubmehl, Research Director, IDC. “Saffron 10’s new machine learning algorithms, ‘Thought Processes’, and the ability to interact seamlessly with other big data systems such as Hadoop will provide organizations with the tools and capabilities that they need to implement high ROI advisory and recommendation solutions,” continued Schubmehl.
Like the human brain, Saffron 10 connects information at the entity level to learn how a particular fact, event, person, product, outcome or “thing” connects to everything else. It understands the context behind these connections, providing businesses with the essential tools to see specific events or patterns hidden in large, complex data sets.
Saffron deploys a patentedassociative memory, or “natural learning” approach, that finds connections among data across diverse sources, without the need for rules or modeling, while learning incrementally and anticipating outcomes based on patterns it finds in the data. Designed to be easily integrated with existing investments and initiatives, Saffron 10 ingests data natively from all legacy file systems, making it seamless for businesses to apply the power of Saffron to other applications.
“Saffron 10 is hastening our vision of powering the Intelligence of Every ‘Thing’. The explosion of devices and sensors require a next generation of data sense making tools squarely focused on separating the signal from the noise, allowing users to not just see important patterns, but anticipate and adapt on-the-fly as new information arrives,” said Saffron CEO Gayle Sheppard. “By combining these composite memories with a deep understanding of time and sequence, Saffron is enabling far more accurate risk analysis, personalization, fraud detection, and other solutions that were previously limited by static data modeling and non-temporal understanding,” continued Sheppard.
Saffron is commercially deployed for real-time operational risk intelligence and decision support in smart device manufacturing, financial services, energy, healthcare and national security industries. Saffron applications are numerous including model-free fraud detection, personalized anticipatory maintenance, country risk assessment and personalized customer experience.
About Saffron Technology
Saffron Technology is the first cognitive computing platform that learns, reasons and anticipates like you, for you. The platform learns incrementally and adapts in real time, ingesting data from disparate sources and automatically connecting the dots to illuminate the knowledge that really matters and anticipate what will happen next. Saffron enhances the speed and volumes at which data can be processed but also critically improves the accuracy of results. Businesses using Saffron can anticipate market trends, optimize processes, mitigate risk, personalize customer experiences or find new revenue streams. Founded in 1999, Saffron Technology is headquartered in Los Altos, California. For more information, please visit www.saffrontech.com.
Saffron Technology, a company building technology able to process information like the human brain to better help Fortune 1000 companies understand their data and make more informed decisions, today announced that Chris McGugan has joined its Board of Directors to help the company continue its growth trajectory.
“I first invested in the company in 1999. I’ve known since then that Saffron’s approach was way ahead of anything else – even before Big Data became a household term,” McGugan stated. “Fast forward to today and Saffron has taken a quantum leap forward in associative/machine learning with the build-out of its Natural Intelligence Platform which is truly changing the way businesses make informed decisions.”
In development since 1999, the company’s patented associative memory technology finds connections among data across diverse sources, without the need for rules or modeling, while learning incrementally and anticipating outcomes based on patterns it finds in the data. Saffron’s Natural Intelligence Platform goes beyond finding patterns in the data: it helps clients identify similar events and understand root causes to rapidly inform decisions and reduce unnecessary risk and waste. Saffron is ideal for real-time operational risk intelligence and decision support; for industries including manufacturing, retail, defense, and healthcare.
“We are fortunate to have Chris McGugan lend his talents to our Board. As we continue to broaden the market for our Natural Intelligence Platform, Chris’s insights and experience will be a significant asset to us,” said Saffron CEO Gayle Sheppard. “In particular, Chris will help us accelerate our market penetration in key vertical markets such as manufacturing, healthcare and finance.”
McGugan has been vice president of Vice President and General Manager of Emerging Products & Technology at Avaya since 2008. He previously held senior positions with Symbol Technologies, Cisco Systems and Belkin, among others.
Saffron Technology combines the power of computing with the genius of the human brain to help Fortune 1000 businesses understand what is in their data, analyze risk, predict future outcomes and make more informed decisions. Like the human brain, Saffron’s Natural Intelligence Platform unifies hybrid data and associates new data inputs with existing ones. It identifies similarities, converging and repeating patterns and learns continuously in the process. Unlike the human brain, Saffron is not limited in size and can process tremendous amounts of data, including prior data inputs, to better anticipate trends and identify anomalies. Founded in 1999, Saffron is headquartered in Cary, North Carolina. For more information, please visit www.saffrontech.com.
Predictive Analytics is red hot! Last week, Saffron attended Predictive Analytics World, collocated with Text Analytics World. Chris Sailer from the Gates Foundation (a Saffron customer) presented “Predicting Threats For The Gates Foundation – Protecting Our People, Investment, Reputation and Infrastructure”. Our CTO, Dr. Paul Hofmann, presented novel ways of thinking about machine learning and prediction, including a more universal, non-assumptive approach to statistics based on Kolmogorov Complexity, which is different and more powerful than data modeling in traditional statistics (more on this soon!).
The world of predictive analytics for Big Data is seeing a plethora of new approaches adding to its rich history as depicted in The Analytics Big Bang. More recently, Eric Seigel, founder of Predictive Analytics World and Text Analytics World, defines the field in the title of his book, Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die. Given its long history and current revolution, predictive analytics covers both new school and old school methods. But new school thinking such as more brain-like Cognitive Computing will differentiate the old ways from the new ways of thinking about prediction.
For example, Google’s Prediction API is essentially a classifier. Classifiers are fundamental to predictive analytics, but the semantics of “prediction” can be a bit tricky. Suppose a classifier is used to identify a thing as animal, mineral, or vegetable. The class is unknown, so in this sense we are “predicting” what the thing is, but only in current tense. If classifying email spam, then we are “predicting” that the user would likely also classify an email as spam, again in present tense. On the other hand, predicting customer churn is more clearly predictive of a potential future event – assuming the customer is still a customer at the current time. However, in regard to human behavioral prediction, the single criticism on Wikipedia for Predictive Analytics is that environments and people are always changing. Static, generalizing statistics are too simplistic and sterile.
Instead, “anticipation” is gaining currency as something very much like prediction but with more nuanced and advanced meaning. Emerging examples include:
Anticipatory Theory. Cross-disciplinary efforts are forming to capture the many understandings of anticipation from philosophy, biological control systems, social psychology, and AI such as in UNESCO’s Project Anticipation, which provides many readings on the subject. Quoting the best-known definition from one of these papers, “An anticipatory system is a system containing a predictive model of itself and/or its environment, which allows it to change state at an instant in accord with the model’s predictions pertaining to a later instant’’. Predictive models are assumed but also assumed to always be incomplete and able to adjust on the fly. Anticipatory systems are cognitive in this sense of knowing what they know and do not know, where novelty and surprise are taken as an opportunity to learn more – if the predictive method can learn on the fly. In the Wikipedia definition of anticipation, curiosity drives opportunities to learn more. As a cognitive system, our human brains seek information.
Anticipatory Sense Making. Anticipation Theory is being put into practice. For one, a team of active practitioners has shared their Anticipatory Thinking, stemming in part from national security experiences such as in Singapore. The problem is centered on dealing with low probability events, not the prediction of likely events. Anticipatory sense making tries to understand all the complexities and potential interactions that could unfold in the future, not just making sense of the past. To get a handle on this difficult problem, the computation of “trajectory” and “convergence” play a central role. As the authors argue, trajectory and convergence are much more that the cue-based mapping of stimulus-response when making a one-point prediction. We need to track where things are headed in information space. US National Security interest also know the need for such a new direction. In private presentations, one leader has challenged, “Even my dog has the intelligence and curiousity to learn and choose its options. Why can’t we do the same?” In another major new analytic lab to explore new solutions, trajectories are seen as the way to move past historical data mining and current streaming data alerts. Predicting low probability events is a tough challenge for traditional methods, so new ways are being sought.
Anticipatory Mobile Computing. Toward what we will all enjoy in the future, smart phones are becoming “cognitive phones” as described in Anticipatory Mobile Computing: A Survey of the State of the Art and Research Challenges. The prediction of future events regarding locations, social interactions, and health hazards are but a few of the applications. Obviously, the data for mobile computing is extending to a swath of new sensors for location, social, and health inputs to name a few. However, machine learning of the anticipatory kind will be critical to future human-machine interaction. “Cognitive phones” must overcome the mismatch of human thinking with current machine technology. Currently, we must now adapt to machines rather than the other way around. To realize this vision, we need autonomous models that continuously learn over time, as in the definition of anticipation. The authors further address “the exploitation versus exploration tradeoff” between accuracy and curiosity. Noting the importance of “latent learning” as one answer, constantly learning from the environment without explicit training and feedback is required.
Saffron provides both anticipatory sense making and highly accurate prediction, but in a different way than traditional data models, more like our brains as an anticipatory cognitive system. A summary of anticipatory principles and how Saffron addresses both prediction and anticipation will be in future posts. Watch this space!
When we think about how doctors and engineers work, it is reasoning by experience and by similarity. Now the arrival of ‘big data’ is driving the need for big individualized memories… so we can leverage all that experience.
The human capacity to remember the complete history of something and to bring all that knowledge and experience to bear, for any one vehicle or one patient is very limited.
Nevertheless, doctors and engineers reason by experience and by similarity.
Even if this is not the same vehicle or patient, they ask themselves what has been seen before.
What did they learn over the years that is like this?
Have they seen this problem on a different machine? How about a similar type of problem or a similar factual background?
This was an approach that worked while the sum of experience was learnable, or at least manageable via books, records and computers.
Yet healthcare is a great example of an industry where big data is now prevalent everywhere you look.
There are lots and lots of patients, and lots of different, complex, fast changing situations for the practitioners, physicians and others to make sense of.
Similarly manufacturing, and more specifically maintenance repairs, is becoming increasingly complex.
Each vehicle, facility, engine and other machinery has its own history. These consist of millions of parts, all the individuals involved in manufacture and maintenance, plus all the use since then. A single vehicle can last decades, which goes beyond human capacity.
So how do we capture all that data and experience and bring it to bear in a new situation at high speed and with intelligence – so that the new / current problem can be diagnosed and dealt with.
We need a way to understand each vehicle and each person, with its own history and own specific and individual problems.
This complete history, is in fact a memory – a computer based memory containing everything single thing we have learned.
Every person, place and thing.
Such memories are now feasible – we call them a memory base.
And just like a human memory, we can ‘recall’ what is relevant and what makes sense in our current situation.