Figure, a renowned robotics developer, has made waves with its latest breakthrough in the field of robotics. The company recently showcased a video demonstration featuring its first humanoid robot engaging in real-time conversations, all thanks to the integration of generative AI from OpenAI.
With the collaboration between Figure and OpenAI, the humanoid robot, known as Figure 01, is now capable of having full-fledged conversations with humans. This remarkable feat highlights the robot’s ability to comprehend and respond to human interactions instantaneously, opening up new possibilities in human-robot communication.
The partnership with OpenAI has equipped Figure 01 with high-level visual and language intelligence, enabling the robot to perform fast, low-level, and dexterous actions. This combination of cutting-edge technologies empowers the robot to carry out various tasks with precision and efficiency.
A video demonstration showcased Figure 01’s capabilities in a makeshift kitchen, where the robot interacted with its creator, Senior AI Engineer Corey Lynch. The robot effortlessly identified objects such as an apple, dishes, and cups when prompted by Lynch. Notably, Figure 01 recognized the apple as food and proceeded to collect trash into a basket, showcasing its multitasking abilities.
Lynch further elaborated on the Figure 01 project, emphasizing the robot’s comprehensive capabilities. He mentioned that the robot can describe its visual experience, plan future actions, reflect on its memory, and explain its reasoning through verbal communication. This extensive range of skills is made possible by leveraging a large multimodal model trained by OpenAI, which can process images from the robot’s cameras and transcribe speech captured by onboard microphones.
The term “multimodal AI” refers to the ability of artificial intelligence to comprehend and generate various types of data, including text and images. By harnessing multimodal AI, Figure 01 can seamlessly integrate visual and language information to achieve a more holistic understanding of its surroundings.
Importantly, Lynch clarified that Figure 01’s behavior is learned, operates at normal speed, and is not controlled remotely. The model employed by the robot considers the entire history of the ongoing conversation, including past images, to generate appropriate language responses, which are then conveyed to the human via text-to-speech technology. Moreover, the same model is responsible for selecting the most suitable learned behavior to fulfill a given command and executing it using neural network weights.
Figure 01 is specifically designed to provide concise descriptions of its surroundings, utilizing “common sense” to make informed decisions. For example, it can infer that dishes will be placed in a rack. Additionally, the robot can interpret vague statements, such as a mention of hunger, and take appropriate actions like offering an apple, all while explaining its rationale behind each action.
The introduction of Figure 01 has sparked considerable excitement and praise on social media platforms. Many individuals have expressed their astonishment at the robot’s advanced capabilities, placing it as a significant milestone in the journey towards future technological advancements.
In response to the enthusiastic reception, Lynch humorously engaged with social media users, sharing their concerns and references to science-fiction movies. However, he assured the public that the development of Figure 01 aligns with practical objectives, aiming to enable robots to perform useful tasks and contribute to various domains, including space exploration.
As the integration of AI technology with physical humanoid robotics continues to evolve, Figure joins the ranks of other notable companies seeking to merge these fields. Hanson Robotics, with its Desdemona AI robot, has also been at the forefront of pushing the boundaries of human-robot interaction.
Figure AI and OpenAI have not provided an immediate response to Decrypt’s request for comment. Nonetheless, the unveiling of Figure 01 underscores the ongoing effort to explore the potential of AI-powered robots in more comprehensive and meaningful ways than ever before.
In conclusion, Figure’s breakthrough in developing a conversational humanoid robot is a testament to the rapid advancement of robotics and AI. With Figure 01’s remarkable capabilities, the boundaries of human-robot communication are being pushed, paving the way for more intelligent and interactive robots in the future.
FAQs
1. What is generative AI?
Generative AI refers to artificial intelligence technologies that have the ability to generate new and original content, such as images, videos, or text, based on existing data and patterns.
2. What is multimodal AI?
Multimodal AI is a type of artificial intelligence that can comprehend and generate different types of data, including text and images. It allows AI systems to integrate information from various modalities to gain a more comprehensive understanding of the world.
3. How does Figure 01 process conversations with humans?
Figure 01 processes conversations by utilizing a multimodal model trained by OpenAI. It analyzes images captured by its cameras and transcribes speech recorded by onboard microphones. The model considers the entire history of the conversation, including past images, to generate appropriate language responses.
FAQs
1. What is generative AI?
Generative AI refers to artificial intelligence technologies that have the ability to generate new and original content, such as images, videos, or text, based on existing data and patterns.
2. What is multimodal AI?
Multimodal AI is a type of artificial intelligence that can comprehend and generate different types of data, including text and images. It allows AI systems to integrate information from various modalities to gain a more comprehensive understanding of the world.
3. How does Figure 01 process conversations with humans?
Figure 01 processes conversations by utilizing a multimodal model trained by OpenAI. It analyzes images captured by its cameras and transcribes speech recorded by onboard microphones. The model considers the entire history of the conversation, including past images, to generate appropriate language responses.
Sources:
– OpenAI (Open AI’s official website)
– Hanson Robotics (Hanson Robotics’ official website)
The source of the article is from the blog shakirabrasil.info