Defense and homeland security agencies are inundated with an endless stream of random data, some of it structured (numerical) and some of it unstructured (text, sound and images). The challenge is finding the critical pieces of information amidst all the “noise” – and responding quickly and effectively, even as huge volumes of data and events continue to unfold. Data collection mechanisms and visualization technologies have traditionally been deployed to reason with the data, but they cannot handle the volume and variety of random data – or provide the real-time monitoring required to identify threats and take action immediately.

Cognika’s automated monitoring and prediction solution — Cognika ESP — is ideal for defense and homeland security applications. The product’s underlying framework combines advanced machine learning (AI) technology with unique algorithms that emulate human learning, memory and predictive reasoning to generate hierarchical Bayesian models from structured and unstructured information. From these data sources, the system automatically:

  • Identifies patterns and sequences in the data
  • Reduces “noise” and focuses on critical data points
  • Enables real-time monitoring and alert systems

What’s more, the Cognika ESP smart analytics engine continues to “learn” over time, making higher confidence predictions (inferences) as new data arrives.

Potential applications in the defense and homeland security area include:

  • Border/Port Security: Data fusion from multiple sensors (e.g. X-ray, radiation) into common operating picture
  • Command and Control: Automated, dynamic, real-time data processing with analytic capabilities for structured, unstructured and streaming data
  • Infrastructure Protection: Analytics tools to quantify interdependencies and cascading consequences to manage disruptions for critical infrastructure
  • Non-invasive Monitoring: Identifying and tracking unknown or potential threats from individual at key checkpoints. Real-time detection of deception or hostile intent through integrated system of machine-human methods.
  • Anti-Terrorism Surveillance: Systematic collection and analysis of information related to understanding terrorist group intent
  • First Responder Systems: Providing all relevant information, predictions and recommendations in a dynamic real-time environment

 

Frequently asked questions:

1. What are the differentiating technologies?

Cognika's ESP technology platform enabled advanced and customized analytics interfaces to complex data. The platform is capable of working with both structured (e.g. Spreadsheets, relational databases etc.) and unstructured information (such as text documents, emails etc.). The ESP platform has also been extended to process information from other media such as audio, images and video - it is important to note that this is achieved by processing the content of the source in addition to any attached metadata.

The powerful technology provides a system for very sophisticated analytics in order to make reasoning/inferencing more intuitive, accessible, manageable and meaningful.

2. What is "latent semantics"?

As the phrase indicates, the ESP platform is capable of gleaning hidden or hitherto unknown relationships between concepts using machine-learning and statistical methods. These relationships can be verified by a human analyst by examining the chain of relationships & predicates the system uses to achieve its inference. For example, a scientist may learn which genes have the highest correlation with a certain medical condition, which biomarkers have shown most promise in relation to detecting a chronic disease. Since knowledge is dynamic and evolving and the sheer scale of it overwhelming. For example, there were almost 700,000 articles published in PubMed in 2008 alone, it has therefore become impossible for researchers and scientists to keep up with the sheer volume of information. Cognika's ESP platform makes this information deluge consumable.

3. What is Cognitive Computing?

The phrase cognitive computing has come to describe Artificial Intelligence technologies that mimic the human brain for learning and reasoning. Just as human brains are capable of processing information along multiple dimensions or senses (such as vision, hearing, touch, smell etc) Cognika's ESP platform employs a similar multi-dimensional approach to reasoning and inferencing. The paradigm can best be described as:

a. "Recognition" i.e. find the similar cases in past dataset (for the current case)

b. Prediction-by-analogy: use the similar cases to extrapolate and predict outcomes of interest.

This approach is analogous to Case-based reasoning in Artificial Intelligence.

4. What current data sources has Cognika indexed?

Cognika provides a full range of data-services; it has currently indexes several public domain data-sources such as PubMed, ClinicalTrials.gov, FDA datasets, NIH related grants databases and multiple other sources. Cumulatively these run into several terabytes of data, that are queried and analyzed in a few milliseconds. We expose our data using REST-ful web-services allowing for powerful querying and usage of our technology in a Software-as-a-service (SaaS) model.

5. What is the process and timeline to include new data sources - both internal to our company and external?

The typical process to index new datasources involves a step of mapping the legacy schema and designing an index model appropriately. This is usually a one-time activity and depending on the complexity of the underlying could last anywhere from a few days to a few weeks. Once the design is complete, the systems is "hands-free" - it can detect updates to the datasets and update the index automatically.

6. What advantages and additional insights does Cognika provide over and above investments already made in analytics products such as SAS, etc?

Cognika applications allow knowledge workers to make sense of complex information in ways that traditional interfaces can't. Data that may seem too overwhelming, too multi-faceted, or too complex. Cognika makes information more intuitive and understandable, often leading to discoveries of relationships and nuance not otherwise possible.

  • Cognika's ESP platform is very flexible and customized for specific applications.
  • Multiple instances can run simultaneously and together to depict different aspects of underlying data.
  • Cognika is designed for real-time views of data, and real-time interaction.
  • Cognika supports nearly infinite "dimensions" of data.
  • Cognika supports multiple data sources: Flat-files, most relational databases, XML, Web Services, and custom data sources
  • Cognika is very efficient with large numbers of nodes and huge databases and works with commodity hardware
  • Cognika's REST-based API helps developers to integrate the system with minimal effort and overheads
  • Cognika offers both predictive and diagnostic views of the underlying model allowing for deep exploration of data

7. Do you leverage other media sources such as imagery and audio/video?

Cognika's algorithms are capable of processing information in other media formats such as images, audio and video etc. The algorithms look at the content of the image through various image feature extraction techniques (as opposed to just querying on metadata - which is what most image or video search engines do).

8. Is your platform "self-learning" and can you give me an example?

By self-learning we mean the system's knowledge is constantly evolving and updating as fresh information is integrated. For example, between 2000 to 2500 articles are published on PubMed every single day, and with them come a vast range of novel research and findings. Hence our model is refreshed every 24-hours to integrate this information and its inferences reflect these insights instantly. For some other datasets the refreshing could be instantaneous to reflect the needs of the application.