Rasa announces to release Open Source version of its machine learning framework that can be used to build automated text- and voice-based assistants. The company releases Rasa AI Assistant Framework 2.0 as an open source project on the Github repository. The open source version of this framework can be downloaded from the Github repository and can be used by Data Scientists to develop text and voice based assistants for their clients.
The Rasa Open Source AI Assistant Framework 2.0 brings many improvements, including dialogue management, training data format, and interactive documentation. This version of the framework is also easy to learn and reduces the learning curve significantly. Developers will be able to get started fast with the help of tutorial, articles and examples. This version is advanced version which brings many advanced configuration that can be used by the experienced Data Scientists.
The Rasa open source machine learning framework is very powerful tool for development of applications that can interact with the humans using text-and voice-based conversations. This framework can be used to build contextual assistants on:
- Facebook Messenger
- Slack
- Google Hangouts
- Webex Teams
- Microsoft Bot Framework
- Chat
- Mattermost
- Telegram
- Twilio
This framework can be used to develop custom conversational channels or voice assistants as:
- Alexa Skills
- Google Home Actions
Rasa AI Assistant Framework 2.0 helps developers in building contextual assistants capable application that can handle log of back-and-forth conversations involving layered conversations. The framework is able to help developers in building applications that can understand the previously discussed information and then uses this information for generating the meaningful conversation. The Rasa enables the developers in developing assistants application that can scale to meet business needs.
Rasa Open Source 2.0 brings policy configuration changes that draw a clear distinction between the policy rules and the machine learning logic rules. This makes much easier in enforcement of business logic rules. The RulePolicy is used in the application for forms, maps actions to intents and other logics. All these changes simplify the newly released framework.
There is a significant change to the forms implementation in the Rasa Open Source library. Previously it was in the Rasa SDK, but now it has moved into the Rasa Open Source library. This change will further lighten Python Rasa SDK and will ease implement on of server action in new programming languages.
The new Rasa Open Source framework comes with the support for YAML Training Data, which helps in using the custom metadata in the training examples, helping in dividing big files into smaller files and the ability to support rich media in the response selector. All these enhance the usability of the Rasa Open Source framework in high quality application development.
The suggested configuration features introduced in Rasa Open Source further simplifies installation by recommending a default NLU pipeline when a new project is initialized. The suggested configuration can be further overridden by advanced users for high level customization of the pipeline.
Other features included in Rasa Open Source are:
- Retrieval Intents
- Rasa Open Source 2.0 is fully compatible with Rasa X, a free, closed source UI tool for development of AI assistant
- Documentation and Rasa Playground
- New examples
- Support for continuous development
- New theme and many others
Rasa Open Source 2.0 is available on GitHub.
