KOGENTIX AUTOMATED MACHINE LEARNING PLATFORM (AMP)

Not just better machine learning, but better business results

AI applications can seem like a complex and convoluted world. Fortunately, that’s about to change.
Kogentix Automated Machine Learning Platform ( AMP) makes it possible to build, manage, monitor,
and deploy enterprise grade AI applications all on an effective, scalable open source software. With
AMP you don’t have to waste time scripting and managing a wide variety of ever changing tools.
Business leaders, data scientists , analysts and engineers can focus on using their data to enable
impacting business actions. For data to have an impact, the entire machine learning application process
must be automated, operationalized and governed.Kogentix AMP delivers.

The AMP Advantage

Innovate. Automate. Operate.

INNOVATE

Innovation requires market leadership, and leadership is gained through velocity. AMP offers the capability to rapidly integrate data, discover trends, and iterate on models to drive the best business outcome.

  • Kogentix AMP connects to a broad variety of data sources, enterprise or external, streaming or batch. Richer and more diverse data yields the potential for patterns and correlations that would otherwise remain hidden.
  • Kogentix AMP enable insights, recommendation to be integrated with business systems as APIs or real time. AMP also includes a rules engine that can generate triggers based on a set of domain specific conditions, making it even easier to create a business impact from a simple user interface.
  • AMP provides capability to users to integrate user defined rules with machine learning. This results in building applications which enable human knowledge intervention and action.

AUTOMATE

As the demand for AI solutions increases, enterprises will look to expand their data science capacity in 2 main ways. Firstly by enabling experienced data scientists to be more productive, and secondly, by enabling business users to leverage machine learning. Kogentix AMP creates additional capacity through automation, a business focused interface, and elimination of coding and scripting.

  • AMP delivers automated capability throughout the application life cycle. Users can immediately see correlations as causal analysis soon as the data loads. AMP can even automatically recalibrate a model once it’s in production. Of course, all of these automations are transparent and can be overwritten and controlled by the user.
  • Data wrangling, a time consuming but necessary task, is made much simpler using AMP. With a simple drag and drop interface, users can join datasets, filter data, deal with missing values that could skew results, and identify key data elements that are more important.
  • Once a workflow has been created, AMP automates the production deployment. The user simply selects deploy, enters some parameters, and deploys with a single click.

OPERATE

Over time, machine learning models lose accuracy as input data changes – maybe a competitor introduces a new product or service; a news event triggers a change in customer attitudes; usage patterns change on an IoT device. AI applications need to have enterprise security, governance, and reliability, but they also need to adapt. Kogentix AMP eliminates the tradeoff between production robustness and adaptability.

  • AMP runs natively on a distributed big data platform, taking advantage of fine grain security and access control, governance, and resiliency. All of the data leveraged by AMP remains in the underlying data platform, avoiding challenges with data consistency.
  • AMP is built for versatility of deployment. It can be installed in a customer’s data center, on a customer’s own cloud instance, or deployed as a cloud service. Customers keep control of their data, and their insights.

Kogentix AMP: End-to-End Unified platform to build Enterprise Machine Learning Applications

AMP Feature Set

Spark Support

AMP is built on top of Apache Spark, a powerful open source processing engine built around speed, ease of use, and analytics library.

Hadoop Support

AMP is certified on Cloudera Enterprise 5, the modern platform for data management and analytics. Organizations who have built their data lakes on Cloudera Enterprise, can now innovate and derive business value from their data using AMP running natively on their Cloudera enterprise platform.

Cloud Support

AMP can be installed and operated on premise or on your private /public cloud on Amazon or Azure.

Exhaustive Data Connectivity

AMP expedites access to your data assets whether enterprise or external systems, AMP provides pre-built performant connectors. Now connecting to data assets for relational, MPP systems, file, APIs, Streaming data, external data, API etc. is just a few clicks away.

Exhaustive Data Transformation

Data Wrangling is one of the most time-consuming aspects of building a machine learning application. AMP now provides a simple intuitive drag and drop interface to perform all your data wrangling. Technical data transformation, analytical or even statistical a very deep library of operators takes efficiency and speed to market to a whole new level.

Extensible

Innovation can never be kept in a box, AMP provides capability for the enterprise to extend and interoperate with various modules built in python or R. Simple click and configure interface allows enterprises to embed their existing components or external components innovated by the organization or open source community.

Supported Algorithms

AMP allows the user to leverage multiple models, and paradigms. AMP provides out of the box Clustering, Classification, Regression, Survival, Time Series, Graphs and Deep Learning capabilities all distributed and optimized for big data environment. AMP also provides governance aspects like traceability and impact analysis across various configurations and runs.

DevOps

AMP provides a click and configure option for deployment of the models, rules set or interconnected applications in streaming, batch or rest API modes. AMP also provides the governance around the process for organizations to put the required checks and balances for approvals and validations.

Rules Engine

AMP provides a human click, configure and preview interface to capture business rules and actions which can be deployed in combination or independent of other AMP components. Now organizations can act on data, define and prioritize required actions, and embed them in data flow across streaming, batch or API mode.

Scenario Analyzer

AMP provides a scenario Analyzer, you can simulate the impact of various factors on your business model and see the impact in real time, compare and visualize the results across scenarios so that you can field the what-ifs, experiment and test various scenarios and take the best possible action for your business.

Full Feature Data Governance

AMP has taken governance and control aspects of building machine learning applications as first-class citizens of the process. Instead of adding in checks and balances as an after thought, it is built from ground up at every level. So, Security, separation of duties, approvals of flows and logic are seamless to the user. Integration to security policy of the underneath data systems and integration with access management come out of the box.

Performance Tracking

In this world, the only thing which is constant is change. Data applications are no exception to the rule. Human behavior, Introduction to a new product, Changes in external factor all can suddenly start impacting the performance of models and rules. AMP provides capability to monitor and track each and every execution of the rule condition, and the performance of deployed models based on the responses coming back from the closed loop.