Designing conversational AI is not as tough as it pronounces. Tools such as Lex developed by Amazon, Bot Framework by Microsoft, and Google’s Dialogflow, supports conversational AI development. Uber expects to pull off the remaining barriers with its customized platform. The ridesharing-service providing organization titled it as Plato Research Dialog System, which has released in open source this week.
In a blog post, Uber spokespersons specified that the company’s AI research division, Plato, is projected to establish, train, and deploy conversational AI mediators, who would allow data scientists to gather information from prototypes and demonstration systems. It features a clear and explicable design. Moreover, it assimilates model-tuning optimization frameworks and existing deep learning technology to avoid re-writing codes.
This first update rolled out for Plato supports interactions via text, speech, or structured information. In addition, each conversational agent can be able to interact with other agents, human users, or even data. Additionally, Plato can store and incorporate all the information gathered during training periods into every conversational agent component.
Plato supports a wide variety of conversational AI architectures, and several well-known machine-learning libraries such as TensorFlow developed by Google, Uber’s Ludwig, and PyTorch by Facebook, can be used to train each element.
On a similar note, Uber’s flying taxi services providing subsidiary, Uber Air, has been seeking for vehicle partners to launch its services across the different regions. Last year, Uber teamed with an American aerospace manufacturer company, Karem Aircraft. This week, the aerospace manufacturer has raised $25 Million in Series A funding round in order to emerge its new air taxi spin-off. Hanwha Systems, a Korean industrial conglomerate led the funding round.
Uber is planning to begin testing of these aircraft next year and expects to commercially launch Uber Air in both the US and Australia from 2023. The company is looking to start providing services in California, Los Angeles, Texas, some regions of San Frisco, and Melbourne.