OpenText™ AI and Analytics offers a comprehensive set of tools that help enterprises overcome data challenges at their pace. It ensures a smooth path towards AI technologies focused on improved decision making, document transformation for accessibility, and sensitive data discovery for risk identification. Check out the latest OpenText of Cloud Editions (CE) announcement to learn more about the most recent release.
May 2023: What’s new in OpenText AI and Analytics CE 23.2
OpenText™ Magellan™ Text Mining
Regrouped classification profiles
In Magellan Text Mining CE 23.2, there are several UI/UX improvements and performance enhancements designed to obtain insights faster. One of those features allows you to configure text mining tasks faster by making it easier and faster to navigate classification profiles. In text mining terms, profiles are custom or out-of-the-box configurations aimed at a variety of text mining tasks. Over time, the number of these profiles can grow in quantity making navigating them difficult. In this release, the Magellan Text Mining Studio user interface has been enhanced to support regrouped profiles allowing users to easily search and navigate through multiple profiles.
Improved model statistic navigation
This version of Magellan Text Mining also comes with an improved experience to better navigate through a model’s statistics. This allows less experienced users to interpret classification results and make better decisions regarding the next course of action like focus on precision, recall, a specific classification, or what to do with incorrectly classified documents.
Machine learning model training performance
There have also been significant improvements to machine learning model training performance allowing users to train classification models in a much shorter time to derive high-quality insights faster. Users can train complex classification models with 10-20 thousand annotated sample files in under 1 hour.
New pre-processing rules
Pre-processing rules help to better prepare raw input for text mining tasks and thus guarantee more accurate results. In 23.2, new rules for upper case sequence pre-processing and contract file cleaning are introduced in Magellan Text Mining Studio. These rules can be cloned and modified according to your specific content scenarios.
OpenText™ Magellan™ Risk Guard
Headless sensitive data discovery services
OpenText Magellan Risk Guard is the AI-driven sensitive data discovery software that allows businesses to uncover risky content. In CE 23.2, new services allow developers to interact with the backend of Magellan Risk Guard through headless API calls, that is the communication path between an application and OpenText Magellan Risk Guard. This expands the integration capability with any desired workflow or external application requirement.
BETA Risk Guard for SaaS
Also in 23.2, the work on OpenText Magellan Risk Guard for Google Cloud Platform is progressing ahead of a SaaS offering. Risk Guard is providing customizable APIs for easy integration within any application. This first SaaS offering allows you to annotate files to validate if there are any PII and classify image and text according to sensitive information taxonomies. This version will also provide users with modifiable profiles to select the models and postprocess the results of annotation and classification according to their SaaS business scenarios.
OpenText™ Magellan™ Integration Center
Cloud deployment support
OpenText Magellan Integration Center is used to integrate and migrate data seamlessly across the enterprise as well as assist in automating processes. This release will support containerized deployment on Azure Kubernetes Service (AKS) and Google Kubernetes Engine (GKE) for self-managed hosting as well as OpenText Cloud managed service providing single tenant deployments.
Web Administration Console enhancements
A web administration console introduced in a prior release received a few new enhancements including grants administration where you can get and modify permissions directly in the web console and support for exporting your repository.
Expanded connectivity and platform support
Enhancement to the OpenText Content Server REST connector include scalability and performance improvements and also includes new support for moving objects and delete operations using the REST API. There is new platform support for releases of Oracle Linux (9.0, 9.1) and RedHat Enterprise Linux (9.0) and newly updated connectivity for various releases of OpenText Content Server (23.1), OpenText InfoArchive (23.1), Microsoft SQLServer 2022, PostgreSQL (15.1, 15.2) and OpenText Directory Services OTDS (23.1).
OpenText™ Magellan™ Data Discovery
Multitenant deployment support
Magellan Data Discovery allows business users and analysts to quickly access, blend, explore and analyze data with no complex data modeling or coding required. In this release, Magellan Data Discovery supports multitenancy allowing customers to have different groups of users sharing resources isolated from the other groups in the same instance of Magellan Data Discovery. This means different groups can share information and analysis inside a tenant but users will never see other tenant users or their information, or even know that other tenants exist. This is a step forward towards a full SaaS cloud offering for Magellan Data Discovery.
Improved exporting/sharing of analysis
Exporting from one server to another has been improved allowing Magellan Data Discovery users to export analysis from one instance, like a development server, and import into another, like a production server. Previously, analysis could only be shared between users of the same instance. Export import functionality has been modified to allow a password when creating the export file, that will need to be provided when importing the file. This will provide Magellan Data Discovery users more flexibility and capabilities when using and sharing analysis.
OpenText™ Magellan™ BI & Reporting
WCAG 2.1 AA accessibility certification
OpenText Magellan BI & Reporting is the enterprise-grade BI platform that generates, manages, and securely delivers data-driven information assets at enterprise scale. To promote inclusivity, several projects have been ongoing to deliver an accessible report viewing experience, empowering users living with a wide variety of visual, motor, cognitive, and other disabilities to consume reports effectively. For this release, we have officially received Web Content Accessibility Guidelines (WCAG) 2.1 AA accessibility certification.
The certification applies to both web-based report viewers — the Standard Viewer and the Interactive Viewer — including cases where they are embedded into third-party applications using the JavaScript API. The certification applies to keyboard navigations and shortcuts, screen reader support, chart patterns, marker contrast support, descriptive captions, and alternate text. The certification was conducted on version 22.3 so that versions of the Analytics Designer and higher are certified to author WCAG 2.1 accessible reports.
For more information about OpenText AI and Analytics 23.2, check out the release notes for Magellan Text Mining, Magellan Risk Guard, Magellan Integration Center, Magellan Data Discovery, and Magellan BI & Reporting
February 2023: What’s new in OpenText AI and Analytics CE 23.1
OpenText™ Magellan™ Text Mining
Integrate externally created classification models
The shift to utilize external classification models, like open-source transformer-based BERT models, within Magellan Text Mining allows you to leverage the treasure trove of preexisting models through both the API and now the UI. This innovative new approach allows you to combine existing OpenText composite AI models with BERT models providing an increased level of content classification.
Sentiment analysis extended to include empathy detection
A new machine learning model to detect empathy, like feeling anxious or growing impatient, combined with the existing models to detect emotion, tonality, intention, and hate speech provides a complete understanding or your customer, citizen, or employee generated content. This natural language understanding model is useful for profiling user generated content such as comments, blogs, chats, etc. The new empathy model includes the following classifications:
| Empathetic concerns | Personal distress |
|---|---|
| Bad Mood Bad Shape Demotivated Feeling Anxious Feeling Bad Feeling Depressed Feeling Fear Feeling Guilty Feeling Lost Feeling Miserable Feeling Pain Feeling Sad Feeling Worried Feeling Sorrow | Deserve Diagnosed Feeling Sick Frustrated Growing Impatient Hard Time Inaccurate Insane Insults Irritated Not Working Pressure Something Crazy Something Wrong Stressed Trauma Uncertainty Unjust Waste of Time |
Granular benchmarking service for text classification models
Data Scientists and Computational Linguists need tools to fine-tune classification models. Detailed metrics showing the accuracy of a model for each document and for each classification allow them to focus on the area of the taxonomy where precision or recall issues must be addressed. By looking at the problematic areas in the model, the users can make informed decisions on how to improve accuracy. For example, they can quickly identify 4-5 classifications with a lower recall and decide to add more training exemplars to the model or build the rules to trigger classifications based on key phrases found within text. With Granular Benchmarking, not only the modelling, but also the presentation of the classification results is made clearer and comprehensible.
New scriptable connectivity platform
The Scriptable connectivity platform allows Magellan Text Mining Studio users to build Python scripts to access the content that was previously not available for analysis without building new connectors. The option to build Python scripts directly within the studio is not only a time-saver but the users now have control over connectivity scripts allowing them to easily re-adapt those scripts for new content repositories.
Automatic rule creation based on classification name
Building deterministic rules in a model is a time-consuming exercise, but it always begins with the creation of a basic rules layer. An automated, one click, procedure to generate the rules based on the taxonomy labels helps users to create that basic layer without manipulating large XML or JSON files. Once the rules are created, the users can immediately start classifying the content and with the automated process the users can quickly tag the content to build training sets.
OOTB Pre-processing rules for email segmentation
Pre-processing rules help users to remove the portions of the input text that have no value for text mining. For example, the user can automatically remove email signatures or any text that is not directly related to the body of the email. These new pre-processing rules for email segmentation can be cloned and modified with additional key words and key phrases to further refine this pre-processing step.
OpenText™ Magellan™ Integration Center
New document transformation capabilities
Version 23.1 of Magellan Integration Center includes a few new document transformation capabilities including the ability to merge multiple files in PDF, TIFF formats into a single document and a new system function that can converts .doc and .docx and TIFF files to PDF.
New Integration Center Web Admin Console
A new web administration console has also been released in version 23.1 and includes many features previously only available in the desktop client. Some of these features include the ability to create repository entries, add, delete, or update server hosts and client hosts, declare services, define applications that a client host can use, add or update users and more… all the from the browser.
Content Server REST API connector enhancements
With this release there are several improvements to the Content Server REST API connector including URL creation support for Content Server or xECM repositories, the ability to perform record management classification, and several enhancements with the artifact copy operations.
OpenText Core Content connector
A new two-way connector to OpenText Core Content is now bundled with Magellan Integration Services providing access to this content store.
OpenText™ Magellan™ Risk Guard
Support for bulk classification model import tasks
Users can browse existing classification models and select all or several classifications to expose at once within Risk Guard user interface. This feature helps to customize the models available within Risk Guard much faster and makes that task more convenient.
Metrics to validate the usage of Magellan Risk Guard
Users can rely on the size of the files that were processed in MRG project and browse those from metrics. This feature allows users to better plan Risk Guard tasks.
Processing of German PII/PSI content greatly expanded
The availability of German PII/PSI models in Risk Guard along with the expansion of the dynamic linguistic profile for processing German content opens up Risk Guard to organizations needing to evaluate German content for sensitive or risky information.
For more information about OpenText AI and Analytics 23.1, check out the release notes for Magellan Risk Guard, Magellan Text Mining and Magellan Integration Center.
August 2022: What’s new in OpenText AI and Analytics CE 22.3
OpenText™ Magellan™ Risk Guard
Discovery of files with sensitive data related to specific data subjects
Business users can support compliance with consumer and employee privacy regulations by discovering and taking action to protect potentially risky files related to specific customers or employees. This enables users to easily create a Magellan Risk Guard risk assessment dashboard for specified data subjects.
New UI capability to customize the discovered risks in built-in reporting
Information security, privacy management and compliance teams can optimize their sensitive data risk assessments by including only the most relevant risk types for their project, company or industry. With this update, it is simple to use the Magellan Risk Guard UI to select or de-select risk types. The risk-type customizations will automatically be applied to the built-in dashboards and reports.
OpenText™ Magellan™ Text Mining
New UI capability to add dictionaries to use with Magellan’s composite AI models
Data scientists and business analysts can more easily support text analytics use cases that require their own controlled vocabularies. With this Magellan Text Mining update, users can self-service add or customize dictionaries to use with AI models.
OpenText™ Magellan™ Data Discovery
New supported drivers for ingesting data
Create insights and conduct advanced analysis on 13 new data sources including MySQL, Oracle, SQL Server, Salesforce, and even the files and sensitive data risks identified by Magellan Risk Guard.
OpenText™ Magellan™ BI & Reporting
Security updates, including an upgrade from Apache Axis 1 to Axis 2
Information security and IT teams can support data security by upgrading to Magellan BI & Reporting version 22.3. This update includes thousands of security enhancements as part of Magellan’s commitment to proactively upgrade the underlying third-party components.
New deployment option for Google Kubernetes Engine (GKE)
IT teams can now select from more deployment options to address a broader set of cloud deployment requirements: Google, Amazon and Microsoft Kubernetes services.
For more information about OpenText AI and Analytics 22.3, check out the release notes for Magellan Risk Guard, Magellan Text Mining, Magellan BI & Reporting, Magellan Integration Center and Magellan Data Discovery.
June 2022: What’s new in OpenText AI and Analytics CE 22.2
Magellan Risk Guard
AI-powered risk analysis reporting for OpenText™ Extended ECM and Documentum: Business users can accelerate compliance by quickly creating custom interactive reports without needing AI, data science or analytics expertise. This update enables users to customize their analysis of content in Extended ECM or Documentum by creating interactive PII and PSI reports that are integrated into the Magellan Risk Guard user interface (UI). To learn more, read the blog post about the new reporting capabilities.
User delegation management: Information security, privacy management and compliance teams can secure and limit access to the content risk assessment results available in the Magellan Risk Guard UI. With this update, end-user roles can be set up and managed in compliance with each organization’s information governance policies.
Magellan Text Mining
Model design recommendation for dictionary and text classification: Data scientists and business analysts can create and customize natural language processing (NLP) more quickly with the simplified NLP modeling experience in the 22.2 release. The updated intuitive experience includes recommendations based on an assessment of the model’s classifications, rules and features so that the user can focus on needed optimizations.
For more information about OpenText AI and Analytics 22.2, check out the release notes for Magellan Risk Guard and Magellan Text Mining.
March 2022: What’s new in OpenText AI and Analytics CE 22.1
The latest updates include democratized risk analysis, new text mining models, expanded language support, and external ML model integration.
OpenText™ Cloud Editions (CE) 22.1 features exciting updates across the OpenText enterprise applications ecosystem, including significant enhancements to our AI and Analytics products below:
- OpenText™ Magellan™ Risk Guard
- OpenText™ Magellan™ Text Mining
- OpenText™ Magellan™ Data Discovery, Cloud Edition
- OpenText™ Magellan™ ML Model Management
- OpenText™ Magellan™ Integration Center
OpenText Magellan Risk Guard
Organizations need a way to find out what risky or sensitive information may be within their enterprise content repositories. However, each organization has unique requirements driven by their operations.
In the 22.1 release, OpenText Magellan Risk Guard has new innovations that make it easier to tailor the application to business operations. By addressing many common requirements, these built-in capabilities reduce the time, effort, and services needed to customize Risk Guard.
Configurable user roles and permissions
To democratize content ownership and risk analysis, OpenText Magellan Risk Guard 22.1 provides configurable roles and permissions allowing individual users to analyze their content and manage their own risk assessment projects.
Remediation actions for OpenText Documentum
In addition to using OpenText™ Documentum™ as a content source for analysis, you can now take remediation action as well. In this release, you can copy, move, or delete content across broader content stores, including Documentum.
French UI support
In our efforts to make content risk analysis available to more users around the globe, the Magellan Risk Guard user interface is now available in the French language.
Magellan Text Mining
NLP and NLU models for French
Natural Language Process (NLP) and Natural Language Understanding (NLU) models are now available to analyze French language content.
OCR for non-searchable PDF
Some PDF files are constructed in such a way that the text embedded within the document is not searchable. This release of Magellan Text Mining provides OCR for these documents turning this previously unusable content into valuable insights.
OpenText Magellan Data Discovery
Scikit-learn model evaluation
New features in Magellan Data Discovery allow for the evaluation of scikit-learn models. End users can easily determine the state of a model with a graphical representation. Different measures are considered depending on the model type and can return results of valid, invalid, or degraded.
OpenText Magellan ML Model Management
Publish TensorFlow models in Magellan
Organizations need to be able to get insights from small datasets which requires different frameworks. In the latest release, models developed in TensorFlow can be easily integrated into Magellan, expanding the availability of models within Magellan.
OpenText Magellan Integration Center
Expanded enterprise connectivity
Enhancing usability of Magellan Risk Guard and other Magellan products, this release allows for broader connectivity to leading content sources, including SharePoint connector that includes token-based authentication, JIRA support, and Linux file share support.
Improved crawler performance and scalability
Enhancements improve the performance of the crawler enabling processing of large content and data sets.
Additional details on the latest updates can be found in the product release notes. To learn more about OpenText AI and Analytics, visit our website.
