On January 30, 2024, iFLYTEK Xinghuo cognitive model v35. The press conference came as scheduled. At this press conference, iFLYTEK Xinghuo upgraded again, significantly improved its core capabilities such as logical reasoning, language understanding, text generation, and multimodality, and released the Xinghuo speech model and the Xinghuo open source model "Xinghuo open source-13b" for the first time. At the meeting, Liu Qingfeng, chairman of iFLYTEK, also officially launched the latest achievement of Xinghuo large model empowerment education - Xinghuo Smart Blackboard.
This time, Xinghuo Smart Blackboard released four major capabilities: multi-modal understanding and recommendation, all-natural interaction, virtual human auxiliary learning, intelligent lesson recording and sharing, and the blessing of multiple latest technologies, so that the blackboard can be upgraded from a teacher's board book tool to an intelligent teaching assistant, improve teaching efficiency, and help students learn efficiently.
1. Multimodal understanding and recommendation: make knowledge explanation more intuitive and vivid.
Interconnected OCR board book] 9 elements of handwriting recognition, standard board book quickly presented.
The board book is an important carrier of knowledge transmission, and the Xinghuo Smart Blackboard can accurately capture the chalk writing handwriting and synchronize it to the large screen in real time while retaining the teacher's chalk writing habits. At the same time, based on the multi-element OCR technology, the handwriting recognition of 9 elements such as three-dimensional graphics, functions, and chemical structural formulas is realized, and the standard board book is quickly presented, which not only respects the teacher's board writing habits, but also allows students to better see the details of key graphic formulas.
Intelligent expansion of resources] A circle of knowledge points, expand resources and accurately recommend.
In the classroom, the content written by the teacher can also be extended teaching, and the teacher can obtain high-matching knowledge cards, experiments and other extended resources with a "circle" of key knowledge points, such as the conversion between mathematics two-dimensional and three-dimensional, the operation process of complex chemical experiments, etc., so that teachers can explain abstract and complex knowledge more intuitively and vividly, and improve students' learning efficiency.
2. All-natural interaction: make it more convenient and intimate for teachers to teach.
Voiceprint Login] Let the voice be the teacher's password.
In the classroom, the teacher can log in directly through the voice, the teacher only needs to say "start class", and the large screen will automatically match the teacher's information and complete the login, so that the voice becomes the teacher's password, which is safe and efficient.
Voice control] The large screen operation is direct in one word.
Erase the blackboard", "*courseware", "random selection", "countdown to 30 seconds", "turn up the sound a little louder", ......With the blessing of core technologies such as speech recognition and semantic understanding, the voice interaction between teachers and students and the blackboard is more natural, bidding farewell to complex and cumbersome touch operations.
Mobile Lessons] Bring teachers and students closer to each other, and get closer.
Thanks to the core technology of all-natural interaction, teachers are no longer limited by the three-foot lectern because they have to touch the large screen repeatedly. The teacher can walk among the students and use voice to control the big screen, and the key knowledge points on the big screen can also be reminded remotely by flying squirrels, making the interaction between teachers and students in the classroom more efficient.
3. Virtual human auxiliary learning: let children learn more and more interested.
English Speaking Practice] Every conversation is progress.
Language proficiency is the basic element of the core literacy of the English subject, which is found in Spark 3With the blessing of 5 abilities, combined with conversational AI technology, Xinghuo Smart Blackboard can create virtual human dialogue scenes with different themes for students, and provide real-time oral expression feedback in interactive dialogues, and correct grammatical errors, helping teachers and solve problems such as students not being able to speak, dare not speak, and not speak well in oral English learning, so that students can better immerse themselves in the atmosphere of oral English practice.
Scientific Wonders] Answer brain-opening questions and protect children's curiosity.
Hello Albert Einstein, is a black hole a hole? ”
Kids, a black hole is not a hole, a black hole is.
Based on the large model of popular science jointly developed by iFLYTEK and popular science China, and the application of virtual human technology, the Xinghuo smart blackboard also allows many historical celebrities or scientists at home and abroad such as Zhang Heng and Albert Einstein to enter the classroom and answer all kinds of whimsical questions in a language that children can understand, helping teachers better stimulate children's curiosity and desire to explore.
4. Intelligent lesson recording and sharing: make classroom review more efficient and accurate.
Multi-modal transcription] non-perceptual micro-lesson recording, complete recording of knowledge points.
In addition to the content of the large screen, the writing content of the sub-board on both sides can also be recorded in real time, completely restoring the whole teaching process of the smart blackboard.
Context generation] The classroom content is presented in a structured manner, so that every knowledge point is clearly visible.
Combined with the Xinghuo cognitive model, the classroom record can automatically generate the best slice that matches the course context, click on the slice to quickly locate the lecture video of the corresponding knowledge point, and the teacher can also adjust the slice clip independently according to the actual needs. In addition, the actual record also automatically generates the best class minutes and mind map-style classroom context based on the classroom links, helping students quickly locate and review knowledge points and conduct accurate review.
Share review] Scan the code to get ** excerpts, and accurately review the weak links.
According to the student's classroom performance and after-class homework, the teacher can share the class record with parents in need, and parents can scan the code to let the student**. Students can not only review the whole process, but also directly find the links that need to be consolidated, just look at the corresponding slices, which effectively improves the efficiency of students' independent consolidation.
In the future, with the continuous improvement of AI core capabilities such as Xinghuo cognitive model, speech recognition, and OCR multi-element handwriting recognition, Wisdom Window will also carry out deeper integration of artificial intelligence and classroom teaching, and explore more possibilities of AI + education.