Opening the smart cockpit AI era, NIO’s NOMI GPT device-cloud multi-modal large model is officially launched

On April 12, NIO announced that NOMI GPT has officially launched push service. The update pushed this time is based on a new technical architecture, including NOMI intelligently creating a NOMI GPT device-cloud multi-modal large model. NOMI GPT includes self-developed multi-modal perception, self-developed cognitive center, emotion engine, and multi-expert Agent to achieve global connectivity of NIO products, services, and communities, and provide more efficient and enjoyable AI services.

After the NOMI GPT upgrade, users can experience a number of new interactive experiences including large model encyclopedia, unlimited fun chats, magical atmosphere, fun emoticons, car Q&A, and AI scene generation, and enjoy outsmarting AI A new kind of travel with empowerment. The launch of NOMI GPT will be available simultaneously for models equipped with Banyan·Rong intelligent system. Since its release, NOMI has gone through more than 130 version iterations, with more than 2,000 new features and updates.

NOMI GPT cognitive center, industry-unique emotion engine and industry-leading end-side multi-modal perception are NOMI’s uniquely upgraded three core technology architectures. Together with billions of parameter types, The millisecond-level response speed gives NOMI powerful complex logical reasoning capabilities in perception, cognition, decision-making, etc., and it has the interactive ability to conduct open Q&A with users.

NOMI GPT’s cognitive center—— Understand the intention and have a tacit understanding, and execute it Smart enough, flexible and not intrusive

NOMI GPT cognitive center has extremely efficient information cognitive capabilities, can accurately judge user intentions and needs in massive information sources, and seamlessly schedule components within the large model architecture , to achieve a high-quality user experience that “understands user intentions with tacit understanding, executes smart enough, and is flexible and non-intrusive”.

NIO’s service recognition center integrates the most complete knowledge base in the industry, including central control distribution, multi-mode rejection, cross-domain inheritance & jump, and multi-agent decision-making, allowing NOMI to have It has the ability to process complex tasks, can conduct reasoning and analysis of perceived information, calmly schedule and execute various tasks, and answer various questions. Through the cognitive center, NOMI will intelligently determine the type of interaction, Q&A chat, car assistant, etc. based on what you say, combined with the conversation context and sensory information. It can also support semantic creation of complete shortcut scenarios.

The industry’s unique emotional engine——NOMI与你“Like each other”

The realization of NOMI is the key to realizing interpersonal interaction through the industry’s unique emotional engine. Through long-term communication and interaction, memory, and learning with users, NOMI can achieve personalized growth, making NOMI He is just a cold "assistant", but an emotional partner who understands you.

NOMI's memory ability includes short-term memory and long-term memory. Short-term memory can be used for topics that the user is interested in recently, such as people and objects, and long-term memory can remember the preferences of the user, family and friends.

With the help of the emotional engine, NOMI can learn and grow faster. Through real-time feedback, post-event reflection, human training, etc., it can continuously become smarter and master new skills and knowledge.

The emotional engine also allows each NOMI to develop a "personality" that matches the user's personality, including personality, three views, dreams, etc., making every conversation and expression with NOMI more natural , more warm, and as the companionship time increases, NOMI will subtly become more consistent with the user's personality.

Industry-leading end-side multi-modal sensing——No network, NOMI is also available “See it and recognize it”

With the industry-leading end-side multi-modal sensing capabilities, the car-machine Internet access is no longer a necessary condition for users to use NOMI. NOMI will also have the ability to "see and recognize", "hear and understand", "know hot and cold, and have touch" to achieve account isolation and sensitive information. When leaving the car, in-vehicle communications are encrypted end-to-end, which not only protects user privacy but also greatly expands the perception capabilities of in-vehicle artificial intelligence.

In addition to enabling independent operation on both ends, NOMI, which integrates multi-modal perception such as images (visual), audio (hearing), and body sensors (tactile), can achieve a more comprehensive perception like humans. The ability to detect people and things inside and outside the cockpit, and describe billions of parameters obtained through multi-dimensional perception in natural language.

Based on the ability of multi-modal perception, NOMI can now see your items, clothing, status, and even vehicles, buildings, animals and plants outside the cabin like a friend, and provide services that understand you better and provide better services. Interesting communication.

The device-side multi-modal sensing is being tested internally and is expected to be rolled out to NIO models within this year.

The above is the detailed content of Opening the smart cockpit AI era, NIO’s NOMI GPT device-cloud multi-modal large model is officially launched. For more information, please follow other related articles on the PHP Chinese website!