Percipion can be used to represent a context, understand natural language and reason, providing to businesses insights on their enterprise data
This is Percipion
Percipion™ is a Cognitive Computing software engine. It can be used and integrated by many businesses to modernize multiple challenging use cases. Please find below additional information.
Percipion™ has out of the box a set of core functions to better understand natural language and correctly assess enterprise data to provide users with accurate insights. Please find below the details.
Percipion™ is able to recognize content written in English and French.
Percipion™ is able to recognize the grammatical form of a sentence to better understand natural language.
Percipion™ is able to process misspelled words and correctly associate them to their correct form.
Percipion™ is able to identify positive and negative semantic valences in order to get a more precise data assessment.
Percipion™ is able to process written content and represent sentences in a simplified form in order to process content faster.
Percipion™ is able to identify psychological valences linked to emotions and affective states in order to get a more precise data assessment.
Percipion™ can understand the content provided in natural language (English & French).
Search after Emojis (1550 are known to Percipion™) and transcribe them into words
Split CSV files (like an export from a database) into multiple text files to assess this collection of raw data
Percipion™ has the capacity to reason the way humans do by making deductions.
Once results are available, inject them within a template to create specific files (SQL, JSON, etc…)
Once results are available, create a synthetic heat-map report or an evidence report for the end user
Transparent & Explainable
While using Percipion™, main processes that are executed are displayed on the terminal. This will provide the users with additional information about how data is processed.
In addition, words that are recognized are reported as well as their belonging semantic parent category (superordinate).
This will provide users with a full report showing how Percipion does operate. Nothing is hidden. All is displayed.
Whenever data provided in natural language are assessed, Percipion™ will write results within an XML file.
This file will show the percentage of words which have been found and associated to main categories. In addition, semantic and psychological valences will also be reported.
This report can then be used to inject data within an existing information system and also be used to generate reports for the end user (see heat-map reports below).
Results can be represent in form of a heat-map for either single files or a group of files. In addition, when multiple files are assessed, a statistical analysis will be performed in order to provide additional insights.
The heat-map itself will show which semantic fields are present. The semantic and psychological valence is also represented by a specific color in order to provide a better understanding on what has been assessed.
This report is a PDF file and can easily be used or shared within the organization.
While assessing data, Percipion™ is able to produce an additional report that is called the evidence report.
This report is very useful for forensic experts or for those who like to get a detailed report showing the sentences inside which specific words have been found. Those words will be highlighted in yellow.
Percipion™ is fully compliant with the recommendation of the European Commission to set goals for achieving climate-neutral, highly energy-efficient, and sustainable data centers. Programmed in ANSI C, it is a mono thread application with a low IT resources usage footprint (energy consumption has been measured and is equal to 2.29 Wh).
Install & Go
Percipion™ can be installed in less than 10 minutes on any Linux machine with a 64 bits architecture.
Percipion™ is fully secured and can even be used offline without any network connection.
Reference data cannot be poisoned.
Percipion™ is not linked to any third-party component or library.
Percipion™ is a knowledge-based software solution that performs consistently and meets appropriate accuracy, robustness, and cybersecurity.
Percipion™ can be hosted on your personal computer, on your premises, at the edge of the network, or on a cloud server.
Percipion™ can be executed on many cores in parallel to process enterprise data faster.
Percipion™ has real-time performances (up to 23 verbatims per second, 1 verbatim=3.5kb).
Percipion™ is empowering organizations to build their own Cognitive Computing solutions or microservices to support their digital transformation better. It can be used as a standalone application or embedded within chatbots, web, or enterprise applications. Please find below some examples.
Percipion™ can be used as a standalone application on a server, a laptop or prepared to be used on a mobile phone.
Percipion™ can be embedded within a business application and add new features to an existing software product.
Percipion™ can be used on a mail server to assess incoming and outgoing messages. This usage is very useful for customer support teams.
Percipion™ can be used on a messaging system server to assess conversations. It is an example used to detect real-time groomers.
Percipion™ can be installed on a bootable USB stick. Law enforcement can then quickly scan a computer system for inappropriate content.
Percipion™ can be linked to a no-code workflow (UIPath, Make, Zapier, …). This feature is used to capture content from SaaS applications or social networks.
If you need more information about our Cognitive Computing solution Percipion, please message us using our contact form.