We can manually adjust parameters, observe effects, and carry out comprehensive design of speech recognition and intelligent applications. Yan Ziyi, a 2023 undergraduate student majoring in Computer Science and Technology at Jiangnan University who has just completed the "Multimedia Technology" course, held an AI speech experiment box and excitedly displayed the model operation interface on the screen. At the Artificial Intelligence Practice Innovation Laboratory of Jiangnan University, such "surprise moments" happen almost every day. Entering the laboratory, a sense of technology is overwhelming: AI speech experiment boxes, AI machine vision experiment boxes, natural language processing experiment boxes and other equipment are neatly arranged. The most eye-catching figure is the focused operation of students: some are debugging AI voice assistants into microphones, some are staring at screens to observe the precise grasping of building blocks by robotic arms, and some groups are sitting together to discuss how to use natural language processing experimental boxes to play new tricks. The "Multimedia Technology" course deeply integrates AI technology. Teacher Yang Jinlong picked up a basic AI voice experiment box and introduced, "Now students can call the end-to-end generative large model in the voice experiment box to conduct intelligent voice application development experiments. Students with good foundations can even develop simple AI simultaneous interpretation projects on their own. From 'understanding principles' to' practical application ', there is only such a distance from the experimental box. ”The laboratory is equipped with 36 sets of AI voice experiment boxes, AI machine vision experiment boxes, as well as natural language processing experiment boxes containing devices such as mice, keyboards, displays, and integrated USB microphones and speakers. These devices not only enable students to "understand" and "comprehend" theoretical knowledge, but also enable them to operate machines to "speak" or "write" natural language with their own hands. This week, our group developed an emotion analysis mini program using an experimental box, which can analyze the positive or negative emotions of a paragraph in real time. It's very interesting. ”Chen Yu, a master's student majoring in Computer Science and Technology in 2024 at the university, said that the experimental box supports virtual graphics processor memory adjustment function and can simulate model performance under different computing power environments, just like "rehearsing" real development scenarios in the enterprise in advance in the laboratory. The value of a laboratory never lies in how many devices are placed, but in making the devices a bridge connecting theory and reality. ”Song Wei, Vice Dean of the School of Artificial Intelligence and Computer Science at Jiangnan University, said that every step of laboratory construction, from equipment selection to course design, revolves around "allowing students to truly hands-on and think". In order to make the new equipment truly serve teaching, the School of Artificial Intelligence and Computer Science at the school organized a teacher team six months in advance to sort out the curriculum system, bind laboratory equipment with nearly 10 core courses, and gradually promote classroom "activation". From freshmen to master's and doctoral students, students at different stages of learning can find suitable "practical projects" for themselves. The development kit in the laboratory and the technology used by enterprises are of the same origin. The skills we practice here can be directly applied in internships and workplaces. ”The student Hao Jinglong from the school said. At present, the School of Artificial Intelligence and Computer Science at the university is planning to collaborate with enterprises to establish a "practical exercise base" in the laboratory, introducing real enterprise projects into the classroom, inviting enterprise engineers to provide on-site guidance, and organizing excellent student experimental results into a "case library" to provide reusable and expandable teaching materials for subsequent courses. Our core goal is to build the innovation laboratory into a key hub connecting academia and industry, truly achieving deep integration and seamless connection between industry, academia, and research. ”Fang Wei, Dean of the School of Artificial Intelligence and Computer Science at Jiangnan University, stated that in the future, we will rely on the laboratory platform to systematically build a virtuous loop of "talent cultivation in universities - key technology research and development in enterprises - efficient transformation of industrial achievements". (New Society)