The assignment aims to apply the techniques you have learned in class to investigate one of the datasets described below, build either a chatbot or question-answering system with it, and then make the system voice interactive, by connecting it to speech-to-text and text-to-speech models.

Link to assignment.

The options for the datasets for this task were

1. EDOS: Emotional Dialogues in Open Subtitles
2. SQuAD2.0: The Stanford Question Answering Dataset
3. WizardOfWikipedia: Knowledge-powered Conversational Agents

Link to tutorials The tasks listed below are only suggestions. You don’t need to perform exactly these tasks (for some datasets they may not even be appropriate). Any ideas?

Briefly describe the chosen dataset:

• First tutorial Play around with documents using some of the code from the course. You could, for example:
• First tutorial

Each of the datasets comes with a particular task that you can perform, so:

• Fourth tutorial is the goto here.

Use text-to-speech and speech-to-text models to add voice interaction to the best-performing chatbot / question-answering system found above.

• Fifth tutorial

• Created an app with text-to-speech and speech-to-text interactivity.