Google OSS Expert Prize: May Winners Announcement

[Kaggle] Hi {NAME}, The Kaggle community has tremendous expertise regarding [Google's open source ML software ecosystem]( and [this $1,000 prize]( is intended to celebrate some of our favorite recently shared tutorials. Each month, up to three authors are selected for this award. May 2022 Google OSS Expert Prize Winners - In [awsaf49]('s notebook "[UW-Madison GI Tract Image Segmentation](", the author uses 2D projections of 3D image stacks as training data for a [TransUNet model](. The TransUNet model has an interesting architecture that incorporates concepts from Vision Transformer models (ViTs) into the more well-established [UNet]( approach for medical image segmentation. The submission makes use of [tf.Data]( / TFRecord files and [tf.Keras]( / TPUs, and contains well documented code that demonstrates strong expertise concerning Google's open source ML software ecosystem. - In [soumikrakshit]('s notebook "[Tensorflow Implementation of Zero-Reference Deep Curve Estimation](", the author uses [TensorFlow]( to train a [DCE-Net]( to estimate tonal curves such that the dynamic range of input images can be adjusted (in order to correct for insufficient brightness). You can test out the performance of this model using your own low-light images [here](. - In [aritrag]( and [spsayakpaul]('s notebook "[Investigating Vision Transformer Representations](", the authors demonstrate how to visualize attention heatmaps, projection filters, and positional embeddings from [Vision Transformer]( models that were built using [TensorFlow](. You can test the model and see the attention maps produced for your own input images [here](. You could be recognized yourself for the June 2022 [Google OSS prize](. To be considered, submit public content to Kaggle that makes use of Google's open source ML software ecosystem (e.g. TensorFlow, JAX, etc) [here](. A summary of previous winners can be found at [(. Best wishes, The Kaggle Team Kaggle, Inc 1600 Amphitheatre Pkwy Mountain View, CA 94043 This email was sent to {EMAIL} because you indicated that you'd like to receive news and updates about Kaggle. If you don't want to receive these emails in the future, please [unsubscribe here](. You can also change your preferences on your account's profile page by logging in at [kaggle.com.](