ARDI Documentation - cognition

AI Information Page

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

AI Information Page The AI Information page lets you interact with your AI in a number of different ways. You can open the AI Information page from the AI List. AI Testing/Training Menu For AIs that work on a single moment in time, you'll have access to a list of the various inputs to your AI (shown above).

AI List

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

AI List You can open the AI List from the ARDI sidebar. Choose the AI option. Use the Add New AII button to create a new AI, or click on one of your existing AIs to access the AI Information Page.

AI Parameters

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

AI Parameters This section is still under construction. In the mean-time, you might find documentation about the various options for AIs in the scikit-learn documentation. AI parameters are adjusted in the settings page.

AI Types

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

AI Types There are three basic types of AI that Cognition can create for you. State AIs A State AI is one that determines the instantaneous state of an asset or assets. It can't capture things happening over time, but at any one moment it can tell you if the asset is in a particular state.

Artificial Intelligence - Ins and Outs

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Artificial Intelligence - Ins and Outs There are a variety of different types of Artificial Intelligence (AI) system, but at the most basic an AI converts a set of inputs to a set of outputs. Defining the Question To build an AI, the first thing you need to do is begin with your

AIExec

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

AIExec AIExec (AI Execute) is the most common tool used when manually creating or running Cognition-based AIs. In practise, it takes three parameters. aiexec.py --dir The dir parameter points to a folder where the JSON files that describe your AIs are stored. Each of those files should be in the

Tuning your AI

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Tuning your AI You can tune your AI settings from the AI Information Page. There are three key ways of editing an existing AI. * Adding training data, to improve accuracy, * Tuning your AI to adjust the settings that control how your AI is built, or * Editing the AI configuration to change inputs, outputs and other critical elements of your

Checking your AI Coverage

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Checking your AI Coverage The last step before we name your AI is the coverage check. This checks how well you've covered some of your input and output values with your training data. If you've got discrete inputs (values that move between 1 and 0 for example), you'll be told if you're training data doesn't cover specific numbers.

Choosing the Asset for the AI

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Choosing the Asset for the AI Next, you'll need to choose which asset this AI is going to be looking at. You'll be able to use this AI on other assets of the same type in the future (known as retargeting), but initially we focus on making your AI work on a single asset.

Choosing AI Inputs

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Choosing AI Inputs Now it's time to choose which inputs you'd like to use. Cognition uses its knowledge of your plant to determine a likely set of points that might be relevant to your AI. Unless you're trying to improve performance, you should generally leave most of these options checked,

Choose your AI Outputs

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Choose your AI Outputs This step doesn't happen for some AIs - if you're making a model or a support vector machine, you should continue to choosing your training times. While some types of AI have fixed outputs, most let you choose the various outputs you'd like to have.

Choosing your Training Data

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Choosing your Training Data When choosing your training data, you'll be taken to a page showing the chart of all of your input values. To create a new training time, simply click (or click-and-drag) and a menu will appear, allowing you to choose what outputs you expect in the selected time-frame. If you single-click, only the one moment will be added to training. If you click-and-drag to select a period of time the whole time period will be added to your training data.

Cloning an AI

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Cloning an AI Cloning an AI simply creates a new copy of the AI with a different name. From there, you can edit the configuration of the AI to make changes.

Cognition File Format

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Cognition File Format You can hand-make ARDI Cognition files to build your own custom AIs, or modify your existing files to manually adjust the AI. The core AI configuration is stored in a json file in your ARDI site folder - on a Windows system this will often be

A Complex AI: The Smart Coffee Machine

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

A Complex AI: The Smart Coffee Machine Let's add some complexity to the system, and see what starts happening. This time, we're building a smart coffee machine, and we're asking a different question - we want to know how the user wants their coffee

Machine Learning AIs in Complex Examples

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Machine Learning AIs in Complex Examples Let's use a more complex example than our simple 'is the coffee too hot' detector. We're going to take that concept much further and create a smart coffee machine. It has a much larger set of inputs and outputs, and looks like this

Create a Predictive AI

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Create a Predictive AI Please note that this feature is still under active development It's very important to note that many events can not be predicted. A predictive AI can only predict events that have some kind of repeatable pattern before the event happens. It's impossible to predict an event that is completely spontaneous or is caused by a source that we have no sensors for - such as human intervention.

Creating an AI

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Creating an AI First, make sure you're familiar with what an AI actually is. You don't need to know the technicalities, but it helps to know the basics. Next, decide on the question you want the AI to answer. What is it trying to do? The Cognition AI wizard is designed to help you build AIs that answer questions about individual assets.

Creating a New AI

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Creating a New AI First, it's important to understand what an AI actually is.

Deploying your AI

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Deploying your AI Cognition AIs aren't just for use inside ARDI. You can use them in whatever you want - we make the AI data available through a web API, you can run them from the command-line, you can embed them in Python projects and more. The Deploy AI

Editing your AI Configuration

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Editing your AI Configuration You can edit your cognition configuration file directly in the web-interface, if needed. This provides a simple text editor so you can make changes to your file. Note that it does not validate your file, so it is possible to damage your file and break your AI by saving an invalid or corrupted JSON file.

Machine Learning

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Machine Learning A Machine Learning (or ML) system is a strong AI that learns from training data rather than simply running through a set of rules.

Naming your AI

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Naming your AI You can now save your AI with a name. The system will assign an automatic name, but you can replace it with any name you like - particularly if you're planning on using retargeting toe AI so you can use it on other assets. Finally, it's time to test your AI.

Neural Networks

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Neural Networks This article is technical - you'll be able to use Cognition without it, but are welcome to read more to discover what is happening under-the-hood. Neural networks are one of the most flexible forms of AI algorithm, using a process that is based on a simplified model of the human brain.

Retargeting an AI

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Retargeting an AI You created an AI by first choosing an asset. But that doesn't mean you can only use it on that asset. Retargeting allows you to use an AI you built for one object on others - meaning your AIs can be reusable across any asset of a similar type (as long as they behave in similar ways).

Running Your AI

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Running Your AI The Cognition engine runs in Python. The Python files and libraries can be found in your addon install folder. On Windows systems, this is C:\Program Files (x86)\Optrix\ARDI\Web\Addons\Cognition\Python and on Linux system it can be found at

Cognition

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Cognition Introduction AI In General Weak AIs Strong AIsI A Complex Example Creating an AI Creation Steps Select an Asset Select Inputs Select Outputs Select Training Data Checking Coverage Name your AI Test your AI Other Topics File Format ARDI Main Menu

ARDI Cognition

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

ARDI Cognition Cognition is an ARDI addon designed to help users create Machine Learning AIs to answer questions about their process. Cognition AIs are used for a variety of purposes, including... * Detecting when assets are in unexpected states, * Detecting specific asset conditions,

Artificial Intelligence - Strong AI

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Artificial Intelligence - Strong AI Strong AI (as opposed to weak AI) is also known as machine learning, While weak AIs are hand-made by people, machine learning AIs are taught. They learn by being fed sample data. The key advantage of these types of AI is that they can make incredibly complex decisions based on huge amounts of data, arriving at insights and connections that humans would disregard.

Support Vector Machines

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Support Vector Machines Cognition normally creates neural networks (often also known as Multi-Layer Perceptrons) as the basis for all of your AIs. However, one type of AI instead uses One-Class Support Vector Machines, or SVMs. One-Class Support Vector Machines don't quite work the way that other AIs function. They take a set of inputs, but only produce a single output. They are a

Testing an AI

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Testing an AI The AI testing screen allows you to see the output of your AI over time. You can choose a new time range by clicking the text in the top-left corner, or step forwards and backwards in time by using the left and right arrows in the top-right.

Cognition Training Times File

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Cognition Training Times File The training time file for a Cognition AI is a simple text file that lists the different training times the system should learn from. The file is called trainingtimes.csv will be located inside the folder you nominate in the

Artificial Intelligence - Weak AI

anonymous@undisclosed.example.com (Anonymous) — Thu, 18 Dec 2025 22:50:17 +0000

Artificial Intelligence - Weak AI There are two fundamental types of AI - Weak AI (otherwise known as rules-based AI) and Strong AI (known as machine learning). Don't let the word weak confuse you - this sort of AI is still very useful. But it's also not what people usually think of when you discuss Artificial Intelligence.