麻省理工学院的人工智能 利用3D视频数据 预测水果苍蝇胚胎的发育 准确度达到90% MIT's AI predicts fruit fly embryo development with 90% accuracy using 3D video data.

麻省理工学院的研究人员开发出一种人工智能模型, 能够以90%的准确度预测果胚胎的细胞水平发展, 预测细胞的分裂,折叠和运动行为, MIT researchers have developed an AI model that predicts fruit fly embryo development at the cellular level with 90% accuracy, forecasting cell behaviors like division, folding, and movement using high-resolution 3D video data. 它接受了三个胚胎录像培训,通过跟踪细胞几何和通过双绘图系统进行互动,成功地预测了第四个未知胚胎中发生的事件。 Trained on three embryo videos, it successfully predicted events in a fourth unseen embryo by tracking cell geometry and interactions through a dual-graph system. 该模型有朝一日可以帮助了解人类发展和疾病,但其更广泛使用取决于能否获得详细的实时成像数据。 The model could one day aid in understanding human development and disease, but its broader use depends on access to detailed, real-time imaging data. 由NIH资助的这项研究标志着在将AI与发展生物学相结合方面迈出了一大步。 The research, funded by the NIH, marks a major step in combining AI with developmental biology.