Meet Zenti – An innovative startup using machine learning to prevent suicide.

Zenti is competing with some large companies like Palantir, IBM International Business Machines Corp. (NYSE:IBM) Watson and Digital Reasoning. They have all been incredibly successful in providing large scale solutions to companies who can afford teams of consultants and engineers to come in and help them solve their data processing challenges.

zenti

 We met up with Steven Cracknell, CEO & CPO of Zenti, a revolutionary machine learning platform that enables the end-user to leverage the power of big data analysis.

Story continues below

 Q: Can you please tell us about Zenti and the specific problem or challenge that you are addressing?

Zenti has developed a machine learning algorithm which can process unstructured text data to identify human emotion and intent and understand communication in context. Our system does this in real-time, on any volume of data and in any language.

Zenti’s simple to use software allows for subject matter experts to easily build their own algorithm/classifier to identify content of interest, and then apply this algorithm/classifier to multiple sources of unstructured data.

The problem we believe we have overcome, is making machine learning and big data analysis, accessible to companies in a way that is affordable, scalable and that supports subject matter experts finding the data they want, without requiring a data science or engineering degree.

Q: Please tell us a little about yourself. 

steven cracknellI would describe myself as business innovator and product development specialist who has an ability to transcend the boundaries that traditionally lie between technology and business. I like to keep things simple with a goal of giving end users what they want.

Throughout my career I have had the opportunity to work on some amazing initiatives, ranging from designing a unique machine learning mechanism for the identification of human intent and suicidal behavior, to developing a sophisticated volatility screening tool at a tier 1 investment bank, to devising a customized trading system for a disabled trader.

I am a strong believer in making a difference by ‘giving back’ when the opportunity arises and someone who constantly looks to share my knowledge and expertise. This perspective has allowed me to share my skills on social development initiatives like reaching and mobilizing consumers in developing countries, to designing and running business workshops to women entrepreneurs in emerging economies (Goldman Sachs 10,000 Women Initiative).

Q: Please tell us a little about your technology that drives your platform?

Zenti takes an approach that leverages statistical methods to analyze and understand text data and removes the need for developing and maintaining complex ontologies or enormous Boolean expressions. Zenti captures human pattern recognition by having a non-technical, subject matter expert make a simple binary decision. Each decision becomes a “training event” for a particular classifier.  This accumulation of “training events” captures the human pattern recognition for recognizing target data. Finally, Zenti’s scalable platform amplifies that pattern recognition so that enormous amounts of text can be classified (and ranked based on a score).

SYRXQ: What markets will you focus on in the near future and what is your plan to conquer them?

We have proven our technology through our suicide prevention work with Dr. Joe Franklin (with input from Prof. Matthew Nock of Harvard and Dr. William Sledge of Yale). Together with these leading researchers and clinicians we have developed a means to accurately identify suicidal risk factors and accurately profile individuals at risk of suicide. Working in alliance with the National Veterans Foundation, we are now applying the technology to helping reduce veteran suicide.

“With Zenti’s technology, we could revolutionize the ability to predict suicidal behaviors and immediately translate this ability into real world progress” – Dr. Joe Franklin

“This work with Zenti promises to be groundbreaking. We, like all hotlines, normally have to wait until a Veteran calls us for help. When they make that call because they are suicidal, it’s at ‘stage 4’, the final stage of suicide ideation, which is sometimes too late. Zenti’s technology will allow us to find them and reach out to them at earlier stages.” – NVF President & Founder Shad Meshad.

Our next step is to expanded the capability of our algorithm to apply the same proven methodology to target the deviant who demonstrate signals of violence and destruction and radicalization with murder as intent. This is to enable early identification with a view to intervene and/or de-radicalize. We intend to partner and work with thought leaders and organizations specializing in this space.

Q: What are the key user benefits and features of your platform?

Simplicity and ease of use. Zenti actually makes machine learning and big data analytics accessible and easy (and dare I say fun), by enabling an end user to develop their own algorithm based on their insights and what they want to do with their data.

Q: Who are your competitors?

Companies like Palantir, IBM International Business Machines Corp. (NYSE:IBM) Watson and Digital Reasoning are certainly the biggest players in this market. They have all been incredibly successful in providing large scale solutions to companies who can afford teams of consultants and engineers to come in and help them solve their data processing challenges.

Q: What advantage does your product offer in contrast to your competitors?

We differentiate ourselves by empowering the end user or subject matter expert to develop their own algorithm using our software. By removing that ‘middle layer’ we provide a solution that allows an end user to take control of their own data and apply their expertise to get the results they want.

Q: What makes your platform stand out?

Scalability and affordability.

Big Data
Big Data

Zenti Demo App

To demonstrate the capability of Zenti, we have created a simple demo app which shows how once a user has created a classifier of interest, it can be applied to a data stream to organize and make sense of the content.

 

Using the 1% Twitter Inc (NYSE:TWTR) feed, we show how a user can concentrate on the content they care about most (e.g. Finance, Politics, Sport) and filter out content like Trolling, Obscenity and Pornography (NOTE: the content displayed with these classifiers can be disturbing and offensive)

What it also demonstrates in the case of Twitter, is that instead of a user having to Follow individual accounts, they could create and follow their own dedicated topics of interest.

Click here to take the app for a test run: Zenti Demo App

(requires user to sign in with Twitter and is best viewed on a desktop)

Q: Tell us about your team?

Our team consists of proven leaders in scalability, reliability, creativity and developer tools who all have a strong desire to create a product that help people and make a difference.

Q: How will you succeed in such a saturated market?

Perseverance, having a thick skin and proving the significance of what we are doing by applying our technology to helping deal with the epidemic of veteran suicide first, and using this as a platform to deal with other large scale social issues.

Q: What is your focus for the next 6-12 months

Zenti will continue to grow our partnerships related to suicide prevention and will be publishing a number of papers for peer reviewed journals that will be co-authored by leading clinical and research psychologists, who have validated the technology.

We have also begun a program of commercialization efforts in the US, Europe and Asia, where we will make Zenti’s technology available to be licensed by companies who need the ability to process large volumes of unstructured text data, quickly and easily.

Q: Can you give us some real world use cases for the application?

In general terms, Zenti can be applied to any free text typed by humans, and where the sheer volume of data makes it impossible for a human to read through all of it to makes sense of it.

Specifically:

  1. Doctors or therapists at large medical institutions have a great deal of notes they take, taken by other medical professionals and entered by the patient. No doctor can read all the available text. Zenti would allow the doctor to get to just the small fraction they need instead of looking through all that text.
  1. Similarly, many customer service organizations take in enormous amounts of free text related to emails, product reviews, chat sessions and phone call transcripts. Each has distinct sources and distinct destinations within the company. Zenti allows each individual to get that fraction of the interactions they need to understand and serve the customer, regardless of how the customer interacted with the company, but does not require that customer service person to look through all content from all sources.
  1. For people countering violent extremism, no person or team can look through all the materials generated that may contain relevant content. So for instance, if there is a group who wants to help young people that might radicalize, they can search using keywords to find at risk youth. They will quickly be overwhelmed by the number of items to look through based on a key word search. With Zenti they can create a classifier that, run at scale, can look through billions of messages a day and that can rank the highest risk messages or message creators.

 

suicide prevention

Q: Who are your target clients?

Our target clients include anyone trying to help others and make the world a better place, that needs to search through more free text than they can possibly search through by themselves. This includes medical professionals, researchers, people countering violent extremism, people looking to help those at risk of self-harm and people looking to identify those intending to harm others.

Thanks for joining us. It is very exciting and inspiring to see how technology is being applied to make the world a better place.

 

An ad to help with our costs