Google DeepMind Turns Focus to Football with AI Technology
Google’s sister company, DeepMind, has shifted its attention from chess to football with the development of its new neural probabilistic motor primitives (NPMP) technology. This breakthrough allows AI agents to learn how to operate physical bodies using a general purpose motor control module that can translate short-term motor intentions into low-level control signals by leveraging motion capture data.
One of the challenges facing physically embodied artificial agents is the integration of motor control and long-term decision-making. DeepMind’s simulated humanoid football players learned from thousands of hours of video motion capture data to predict the trajectory of the ball and reactions of other agents. In just 50 hours of using reward functions for scoring goals, the AI agents exhibited complex behaviours at various scales while working together.
Hyundai Launches $400 Million Artificial Intelligence and Robotics Institute
Boston Dynamics is set to power Hyundai’s new robotics institute, based in Cambridge, Massachusetts. This will allow for the creation of robots used more easily, safely, and able to do a wider range of tasks. The NFI supply chain company has already invested in Boston Dynamics after signing a $10 million deal for the deployment of its Stretch robot, designed for the automation of moving and loading of boxes around warehouses.
The Stretch robot will make its official debut in a Savannah facility in Georgia in 2023. H&M and Gap have also signed up for the deployment of robots such as Stretch, allowing the enhanced efficiency, safety and optimization in the movement of goods in supply chains.
MIT’s New Machine Learning Model Can Discover Linguistic Rules on its Own
MIT has created a machine learning model that can discover linguistic rules with higher-level patterns of language that can be applied to many languages. This enables the model to produce better results by learning from examples of language, discovering how words change to communicate different grammatical functions in one language, and explaining why those forms change.
By using multiple smaller-data sets, the system is designed to discover models that are easy to understand by humans, which has implications for future research in multiple sectors. Future work could answer many questions, such as how infants acquire language, or how induction biases affect learning behavior.
Conclusion
As new AI news emerges, it’s important to stay up-to-date with the latest advancements to keep pace with the changing technological landscape. With the development of technology such as DeepMind’s NPMP and Hyundai’s robotics institute, many sectors will likely see some dramatic shifts in productivity, efficiency and safety. MIT’s newly developed machine learning model offers immense possibilities for future research in various fields, particularly in linguistics.
As we progress, it’s exciting to see the increasing integration of AI technology into multiple industries, paving the way for greater innovation and technological evolution.