Feature engineering is transforming raw data into features that represent the physical industrial process

Space-Time dependencies

Build an abstract representation of your process through collaboration & smart recommendations.

SQUIDBOT.io is a deep learning hub for manufacturers where engineers build and train industrial agents in minutes. The Space-Time module will guide you easily in building how your agent perceives information and proposes decisions accordingly.

Guide your agent through Pathemata Mathemata

There will be no better approach than teaching agents than Pathemata Mathemata: learning through pain. In other words, it is much better to teach agents on every piece of information, including bad information.

User-friendly space-time integration module

The Space-Time module provides engineers with user-friendly and intuitive commands that can make spatial & temporal parameter setting simpler.

Brainstorm Space-Time Train Scaling Selection Overview Feedback Target Space Domain Overview Tuning Evaluation Deploy Feature Extraction Data Processing Action Tensor Architecture (2D view) Details T-1 Feature ID FF3 T-2 T-3 T-4 Acme Power Corp. Turbines_KPIs RPM1 RPM2 RPM3 RPM4 WF1 WF2 WF3 WF4 P1 P2 P3 P4 T1 T2 T3 T4 AF1 AF2 AF3 AF4 FF1 FF2 FF3 FF4 TIN1 TIN2 TIN3 TIN4 AP1 AP2 AP3 AP4 Hyperparameters Description Fuel Flow Turbine 3 Status Feedback/Action PCA Score 0.93 Settings Type Numeric Limits Target Min Target Set Detect Entropy
  • User-friendly guide on Space-Time dependencies

    Users are guided in simple steps to understand the logic behind the agent learning process: interaction with environments & decision spaces.

  • Exploration/Exploitation Tradeoff

    Should industrial agents explore more or just get along with exploited targets? It is easier to tune this parameter from now on.

  • How to design state, action & target vectors

    Your agents targets are just the best states resulting from great actions.

Make your production more convex

Designing the right space-time environment makes your agents more convex to volatility

  • Risk mitigation

    Convexity allows your agents to be more prepared for eventual volatility

  • Reality is probability

    The better your space-time medium, the closer you get to reality

Check out our mission guidelines

Request your trial

request your trial