Under the guidance of the Publicity Department of the Central Committee of the Communist Youth League and the Lunar Exploration and Space Engineering Center of the China National Space Administration, China Youth Daily and Baidu jointly launched an online interactive activity called “The World And the Lantern Festival” on February 15. With the first full moon image prototype returned by “Chang ‘e I”, interactive uses AI to generate personalized paintings, so that tens of millions of netizens can intuitively experience the AIGC (AI Generated Content) ability of Baidu Wenxin model for the first time. The event attracted Juan Antonio Samaranch Jr., president of the International Olympic Committee’s Coordination Commission for the Beijing Winter Olympics, Cao Xue, head of the “Bing Dwen Dwen” design team, and han Tianyu, a short-track speed skater, to share their beautiful pictures of the moon with Baidu AI.
The “World Together Lantern Festival” event allows everyone to create an illustrated art work with the help of AI technology. Participants only need to choose the place where they want to view the moon, and wenxin Big Model can quickly understand and associate ideas, and use the techniques and style of well-known artists to create paintings that fit the theme. For example, when we type in “Beijing,” wenxin’s big model can draw the Great Wall and the moon perched on the mountains. If you type “Egypt”, you can see the pyramids, the symbols of Egypt, and even poetically the buildings in the moonlight. When you type in “Mars,” an orange-red surface and the hard-working Zhurong rover come into view in the corner.
AI’s learning and creation ability in the field of art is breaking through our cognition, and also gives us more space for imagination in the integration and innovation of technology and art. For example, through the “World Shared Yuanxiao” activity, the works completed in a few seconds in the hands of participants break through the limitations of physical region, language and other restrictions, and can enjoy the moon scenery created by “AI painters” around the world, and then add a poem corresponding to the scenery, it will become a good work to convey wishes to relatives and friends during the Lantern Festival.
Wen Xin Grand model is the world’s leading “knowledge enhancement grand model”, which continuously integrates learning from large-scale knowledge and massive data, and is knowledgeable and knowledgeable, with higher learning efficiency and better model effect, and has the world’s leading language, vision, cross-modal understanding and generation ability. It can understand language, see pictures and videos, understand the real world, and communicate fluently with people. It is also an “AI creator”, drawing, composing lyrics, writing scripts, novels and so on.
For the first time, netizens around the world can experience the creativity of China’s leading AI technology — AIGC capability based on Wenxin Grand Model. AIGC can stimulate creativity, increase content diversity and interactivity, improve content production efficiency, while reducing production costs, and promote innovation and creation in a variety of areas, including virtual digital people, media content production, and art creation.
According to introducing, wen xin big model currently include NLP (natural language processing (NLP) model, the CV (computer vision) “, “cross modal model, model, basic universal model, also contains a large model in the key areas, key task, and plenty of tools and platform, and become the first base of one thousand AI any profession development model. At present, Wenxin Grand Model has been widely used in Internet, industry, energy, finance, communication, media, education and many other fields through flying paddle and Baidu intelligent cloud platform.
Baidu Research Institute predicts that in 2022, the large model will become practical, and the effect, universality, generalization, interpretability and operation efficiency of the large model will continue to improve, and the application threshold will continue to reduce, so as to achieve a wide range of implementation in more scenarios.