Five university teams are selected to participate in the last live interactions phase of Alexa Prize Simbot Challenge. The teams were chosen based on, among other things, customer feedback and scientific profits for the technical papers produced by each team.
The five university teams selected to participate in the live -interaction phase challenge are:
Alexa customers can interact with the finalists’ simbots by saying “Alexa, playing with Robot” at Echo Show or four TV devices. These ratings and feedback help the student teams improve their bots to compete to see who wins the best prices, a total of $ 650,000 in cash prizes.
The Alexa Award is a unique partnership program for the Industry Academy that provides an agile real-world experimental framework and tools to accelerate scientific discovery. University students have the opportunity to launch innovations online and quickly adapt to feedback from Alexa customers.
It is wild to see the enormous talent, creativity, passion and contributions from all Alexa award Simbot -Teams during this dedication challenge to accelerate the science of robust human AI interactions, and conversation embodied AI.
Simbot Challenge is focused on helping to promote the science of embodied AI agents who can effectively engage in people to understand, learn and collaborate to reach their given missions.
“The next generation Auto -ait AI assistants must be robust and versatile who are able to learn and solve challenging tasks that require multimodal interactions with people, other agents and the surroundings,” said Reza Ghanadan, a senior main scientist in Alexa AI and head of the Alexa AI award. “It is wonderful to see the enormous talent, creativity, passion and contributions from all Alexa Prize Simbot -Teams during this consecration challenge to accelerate the science of robust human AI interactions, and interviewed AI.”
As part of the Alexa Award Simbot program, Amazon provided data, software tools, ML models, visual and conversation AI tubing wires and the embodied AI framework, Alexa Arena, to help teams innovate, launch and experiment with their new AI ideas online and improve their research throughout the competition.
In the final phase, university teams come to develop a bot that best best for commands and multimodal sensor inputs from a virtual world. Like previous Alexa prize challenges, Alexa customers also participate in this phase.
In this case, customers interact with virtual robots driven by the universities’ AI models on their ECHO show or four TV devices and try to solve more difficult tasks within the virtual environment. After the interaction, they can provide feedback and ratings for university bots. This feedback is shared with university teams to help promote their research.
The winning teams are determined under the Simbot Challenge Finals event scheduled for the first week of May. Publications from all ten semi -finalist team appear on the Amazon Science site later this year.
Learn datasets
In connection with the Simbot Challenge, Amazon released public Teach, a new data set with more than 3,000 human to human dialogues between a simulated user and simulated robot communicating with each other to perform household tasks.
In Teach, the simulated user cannot interact with objects in the environment, and the simulated robot does not know that the task to be performed requires that they communicate and collaborate to successfully perform tasks. The public benchmark phase of Simbot Challenge, which ended in June 2022, was based on the Teach Dataset performance from Dialogue History (EDH) Benchmark, which evaluates a model’s ability to predict subsequent simulated robot actions, provide the dialogue history between the user and robot, robot and observations.