Borderless debuts ChatIoT: AI's big model moves from the digital world to the physical world

Bringing the power of big models to the physical world is not only Microsoft, Google, and a Chinese startup. As a technology company focused on the development of IoT development platforms, Edgeless follows the productivity changes brought by AI big models and debuts ChatIoT, the first to apply the technology of big models to the IoT space.

After Microsoft, Google and other companies announced the use of Big Model capabilities for robotics development, Edgeless has built a new AIoT ecology by deeply integrating Big Model with Shifu, an IoT development framework, which means that device access and use in IoT development will change from traditional programming mode to natural language based interaction. This innovation enables people to build AIoT scenarios without coding, thus improving development efficiency and lowering the technical threshold.

In the wave of production changes triggered by ChatGPT, big model technologies are flourishing and impacting many fields. Today, a key question is before us: how to enable big models to empower the real economy, unleash greater potential, and drive the IoT field to achieve intelligence and efficient collaboration?

Edgeless introduces ChatIoT to bring new possibilities to the IoT space. Edgeless further extends the capabilities of the Big Model to the broader IoT development space by simplifying the process of device access and application development. This technology innovation will reduce the difficulty of IoT projects and provide the industry with more intelligent and flexible ways to interact with each other. Looking ahead, AIoT technology will bring profound changes to smart manufacturing, smart cities, smart homes and other fields, further improving the digitization of the physical world and advancing the intelligent development of production and life.

Big Language Model: A New Way of Thinking for IoT Development

Natural language control of the physical world is making breakthroughs. Microsoft announced in March this year that it was testing the use of OpenAI's language model ChatGPT to enable remote control of home and industrial assisted robots. The goal of the research is to verify whether ChatGPT can go beyond text and think about the physical world to help robots accomplish tasks. The Microsoft team says they were able to use this to show that developers can guide ChatGPT's understanding of its surroundings and tasks, which in turn enables linguistic intuition to control the physical world.

Google likewise introduced PaLM-E, a new general-purpose robot model, by transferring knowledge from different vision and language domains into robotic systems. This is a key breakthrough in introducing large scale language models into robotics by directly ingesting raw robot sensor data streams to enable the training of language models. The resulting models enable efficient robot learning and maintain task completion capabilities based on vision models.

In China, Edgeless has plunged into combining large models with the physical world in the same technical path. The beginning of software-defined hardware in IoT development emerged in the process of device access. There are a vast number of IoT devices on the market, each using a different protocol with different drivers, resulting in developers always being inefficient when designing and developing IoT scenarios. However, the Big Language Model has the potential to help developers improve this efficiency. The Big Language Model can design and develop its own drivers that can manipulate new devices based on the characteristics of the new devices based on learning the existing drivers and protocols.

After access is complete, developers can use the Big Language Model for application development. Big Language Models can be used to create, complete, or combine code, whether it is sourced from code snippets or natural language descriptions, and can greatly improve the efficiency of IoT development. With such capabilities, these models can help professional and amateur developers build innovative applications.

A large language model based on GPT can be very helpful in the complete chain from access to application development. Each device uses a different protocol with different rules, making it difficult to have direct interaction with IoT devices using big model applications such as ChatGPT. This requires a highly compatible IoT development framework to provide a standardized device development interface, which in turn enables natural language based data collection, data processing and device counter control.

Edgeless ChatIoT: AI large model interaction model with devices

Edgeless launched ChatIoT based on the Shifu IoT development framework, which enables natural language device-oriented programming and opens up standard APIs so that applications such as ChatGPT can take full advantage of the device's capabilities. On the basis of Edgeless' technology, the AI Big Model can be used to distribute and deploy applications based on device interoperability. Shifu, independently developed by Edgeless, acts as a middleware that translates natural language commands into device-specific commands. chatIoT means achieving a standardized model for LLM in IoT applications, allowing users to use natural language to interact with different devices.

By integrating Shifu's ability to integrate AI big models, Edgeless gives tools such as ChatGPT the ability to extend its reach into the physical world, revolutionizing the IoT paradigm by introducing natural language as a new user interface. The IoT development framework Shifu serves as the underlying infrastructure that enables different AI big models to communicate with devices, such as ChatGPT to complete the cycle of "using natural language to issue commands to IoT devices and make them do the corresponding actions," thus extending ChatGPT's reach into the physical world.

In a typical IoT multi-device scenario application, the ChatIoT development model integrates robotic arms, AGVs (Automated Guided Vehicles), cameras, and other devices to achieve efficient collaboration. Developers use tools such as Edgeless Shifu and ChatGPT to work together to complete the process from device connection and driving to device task collaboration.

First, ChatGPT is allowed to learn Shifu's development framework and understand how to connect different devices such as robotic arms, AGVs and cameras to the system. Developers need to provide metadata about the devices, such as their capabilities and connection addresses. Then, let ChatGPT generate the digital twin microservices for each device through the driver and write the corresponding configuration files. Edgeless will provide rich and complete cases to the big model through the source code repository to accelerate the device adaptation process.

Next, describe the device task collaboration requirements, including movement range, timing and grasping actions for robotic arms, path planning and transportation tasks for AGVs, and monitoring and image processing for cameras. ChatGPT is requested to write applications for controlling and coordinating these devices. During this process, Edgeless will provide generated templates that describe how users can operate the devices through drives from the user's perspective, allowing the larger model to understand the scenario requirements more deeply.

After these steps, ChatGPT's natural language dialog will enable AIoT-based applications for multi-device IoT scenarios. Devices such as robotic arms, AGVs and cameras can work together efficiently to achieve a seamless connection between smart manufacturing, smart logistics and other application scenarios. This will greatly improve production efficiency and reduce costs, bringing revolutionary changes to the field of IoT.

In the future, Borderless will continue to update the device model library, further iterate ChatIoT's development and interaction model, and deeply combine the big model technology with IoT scenario development. chatIoT, as an innovative AIoT solution, lowers the technical threshold of IoT projects by simplifying the device access and application development process. chatIoT will help drive the IoT industry to Application scenarios are more rich and diverse, thus meeting the changing market needs.