PaperWithCode最新一期Newsletter公布了2月最受关注的十篇论文：

1️⃣  How do Vision Transformers Work?

2️⃣  Competition-Level Code Generation with AlphaCode

3️⃣  Slicing Aided Hyper Inference and Fine-tuning for Small Object Detection

4️⃣  FILM: Frame Interpolation for Large Motion

5️⃣  Visual Attention Network

6️⃣  On Neural Differential Equations

7️⃣  EvoJAX: Hardware-Accelerated Neuroevolution

8️⃣  Unifying Architectures, Tasks, and Modalities

9️⃣  Human pose estimation via soft-gated skip connections

🔟  VOS: Learning What You Don't Know by Virtual Outlier Synthesis

此外，还有最值得关注的SOTA、数据集和工具等。

英文原文： https://paperswithcode.com/newsletter/26/