Deep learning has been key to remarkable progress in artificial intelligence. Recently, I came across some intriguing applications of deep learning. Salvator Mundi is the most expensive painting ever sold in the world; however, it is still in debate whether the painting should be attributed to Leonardo da Vinci in whole, in part, or not at all. To solve this puzzle, an attorney developed a convolutional neural network that can analyze the likelihood of a painting being drawn by the supposed artist. The results showed that most of the crucial parts of Salvator Mundi were likely done by da Vinci. Paleontology is also benefiting from deep learning. Paleontologists commonly use CT scans to identify and compare fossils, but the segmentation of CT images is often subjective and time-consuming. Researchers thus came up with the idea of constructing neural networks to enable objective and efficient segmentation of dinosaur fossil scans. It is astonishing how digging deeper into the stories through deep learning can lead to such improvement to research.

中文翻译：

---

深入挖掘故事[编者注]

深度学习一直是人工智能取得显着进步的关键。最近，我遇到了一些有趣的深度学习应用。Salvator Mundi 是世界上最昂贵的画作；然而，这幅画是否应该全部、部分或完全归于达芬奇，仍然存在争议。为了解决这个难题，一位律师开发了一种卷积神经网络，可以分析一幅画是由所谓的艺术家绘制的可能性。结果表明，大多数关键部分Salvator Mundi 很可能是由达芬奇完成的。古生物学也受益于深度学习。古生物学家通常使用 CT 扫描来识别和比较化石，但 CT 图像的分割往往是主观的且耗时的。因此，研究人员提出了构建神经网络以实现对恐龙化石扫描进行客观有效分割的想法。令人惊讶的是，通过深度学习深入挖掘这些故事可以导致研究的这种改进。