Predictive Objects improves the outcome of business and industrial processes through actionable customer and asset
intelligence. The platform leverages machine learning and big data to automate the creation and deployment of predictive models,
augmenting human experts for faster and more accurate predictions that run where the data is produced and where
decisions need to be made.
Predictive Objects can be deployed and run anywhere: inside connected objects or machines, at the edge of the network,
in industrial IoT platforms, in public or private clouds - bringing the decision model to the object and to the data.
Predictive Objects removes the data selection bias by providing access to all
data available to the organization - internal and external. This enables Predictive Objects
to process the entire Gaussian distribution of data, and detect weak signals that would
otherwise be ignored. The process is greatly accelerated through massive automation,
shortening projects from months to hours.
Ingest any data of any type, internal or external, as batches and/or streams.
Check and cleanse all data, homogenize and normalize sources, verify integrity.
Turn raw information into temporal sequences based on the data type and density.
Anonymize data before processing through record transposition for maximum protection.
Predictive Objects’ Brute-Force approach to algorithm selection automates the search for the best combination of algorithms and data sets, by exploring all the possibilities in the realm defined by the goals.
Predictive Objects is technology - and business - agnostic, and systematically tries all possible combinations through its Brute-Force approach, making sure the best combination of algorithm and data is always selected, without bias.
The Meta Active Machine Learning, Predictive Objects’ wisdom that is shared across projects and enriched at each model computation, accelerates dramatically the search for the best model, shortening to hours a process that used to take months of exploration by data science teams.
The model computed by Predictive Objects is enriched and operationalized by a
domain specialist who applies actual human expertise to determine which
variables are actionable and have the potential of influencing the outcome.
Predictive Objects’ prescriptions are generated in natural language using the domain-specific terminology and phraseology. They are then directly usable and actionable by machine operators, customer service reps and other non-experts, without specific training.
The involvement of the domain expert in the validation and/or preparation of prescriptions makes them immediately acceptable by their consumers who feel more involved when presented with actions and recommendations that are sustained and explained.
Deployment to production of the models is immediate
to any target platform: object, device, platform, cloud.
There is no need for coding, dramatically shortening
the time it takes for projects to bring value.
Once the model is ready, deployment to production is immediate, regardless of the target environment. There is no need to "code" an operational version, and even less to maintain and manage distinct versions for multiple targets.
Find quickly if a model work, test hypothesis, and ensure ongoing efficiency: the performance of models is continuously evaluated to test their accuracy, precision and recall. Production and deployment of new and revised models is ongoing.
Predictive Objects can be deployed and run anywhere: inside connected objects or machines, at the edge of the network, in industrial IoT platforms, in public or private clouds - placing the predictive model as close as possible to the system it impacts.
Through the use of standards and also native integrations, Predictive Objects
delivers actionable prescriptions, as well as the timeline, the data points and
the individualized decision tree that generated the recommendation.
Predictive Objects’ prescriptions can be multi-step, with sequences of
recommendations or next best actions.
All predictions and recommendations from Predictive Objects are available in push or pull mode through standard RESTful APIs, making it easy to integrate the platform with any backend system or enterprise workflow. Third party systems can easily reap the benefits of artificial intelligence, with minimum integration efforts.
Predictive Objects is natively integrated with modern business intelligence and dataviz systems as well as packaged apps, including the most used Computerized Maintenance Management System (CMMS), Customer Relationship Management (CRM), Manufacturing Execution System (MES), etc.
At the core of Predictive Objects, Meta Active Machine Learning automates the search for the right combination of algorithm and data set in the very broad space of all possibles. How? By nudging the Brute-Force algorithm selection toward the most efficient subrealm based on goals for precision and recall that have been positioned upstream.
Even though they may seem totally unrelated, use cases in different industries and business sectors actually share common algorithm spaces and benefit from each other’s experience and feedback. Meta Active Machine Learning acts as a centralized repository of all models and is constantly fed by feedback on their efficiency.
Being able to nudge the Brute-Force algorithm selection in the right direction dramatically reduces the time to convergence of the model, especially in use cases where this convergence can be difficult to achieve.
By automating the search for algorithm and data sets, Predictive Objects removes the preconceived bias and subjectivity of the human and helps the organization obtain the best possible outcome.
Meta Active Machine Learning does not ignore weak signals in the source data, it is actually the only technique that ensures they don’t get lost in massive but less-significant data volumes.
Interested in our Predictive Objects features, set-up, model, etc?
Fill in the following form to request a meeting.