Post by Pedro Fialho, INESC-ID
Natural language is inherent to many games, for example in story guided games where players are provided with sets of spoken and/or written sentences denoting ways of continuing (and conducting) the game. However, the user experience of such games would be improved if natural language processing (NLP) is involved, for instance with speech and/or text inputs that allow the player to engage in bidirectional natural language conversations with the game.
For the game developer, NLP provides valuable tools too, that: a) support game development, such as by reducing the need to script all valid user inputs; b) enable more variability in games, by automatically conducting the game based on linguistically implicit user intentions and/or emotional clues, and c) may also reduce costs, as is the case when using Text To Speech (TTS) technology to avoid pre recording utterances.
During gameplay, the need for multimodal inputs arises when the player is able to do more actions for better gaming performance, but only hand based controllers are available. SpeechIO enables games with speech and text input modalities, by providing seamless access to technologies such as Automatic Speech Recognition (ASR), for speech to text transcription, and Natural Language Understanding (NLU), to match the natural language in a user’s input with the game script.
Other uses of the SpeechIO asset include detecting emotions in text and speech, or recognizing intention in dialogue acts. More information on how and where to use SpeechIO coming soon.
Posts are written and signed by its authors. RAGE Project does not responsabilize for the opinions and comments published in them.