Alibaba DAMO Academy Releases Remote Sensing AI Large Model
-
Alibaba DAMO Academy has released the industry's first remote sensing AI large model. According to reports, the model can identify farmlands, crops, buildings, and other features, enhancing the analysis efficiency of remote sensing applications such as disaster prevention, natural resource management, and agricultural yield estimation. Currently, the model is available for use on the AI Earth Earth Science Cloud Platform.
The remote sensing AI interpretation general segmentation model (AIE-SEG) is the first in the field to achieve task unification in image segmentation. A single model can achieve "zero-shot extraction of all objects," identifying nearly 100 types of remote sensing features such as farmlands, water bodies, and buildings. It maintains high accuracy in multiple tasks and can automatically optimize recognition results based on user feedback. In certain scenarios, compared to traditional remote sensing models, the accuracy of instance extraction can be improved by 25%, and the accuracy of change detection can be improved by 30%.
Based on these foundational capabilities, the remote sensing AI large model provides "out-of-the-box" API services. Users can customize different remote sensing AI interpretation functions according to their needs, such as water body extraction, farmland change monitoring, and photovoltaic identification.
Previously, Alibaba DAMO Academy also announced the launch of its self-developed open-domain text understanding large model on the ModelScope community. SeqGPT is a text understanding large model without domain limitations. It can perform tasks such as entity recognition, text classification, and reading comprehension without training. The model is obtained by fine-tuning Bloomz on hundreds of task datasets and can be used for free on graphics cards with as little as 16GB of memory.