PinnedLeon SickinTowards Data ScienceFrom Supervised To Unsupervised Learning: A Paradigm Shift In Computer VisionSlowly removing the injection of human knowledge from the training process·6 min read·Jan 1, 2022--1--1
Leon SickinTowards Data SciencePaper explained: Exploring Plain Vision Transformer Backbones for Object DetectionThe power of ViTs as object detection backbones·5 min read·Nov 20, 2022--1--1
Leon SickinTowards Data SciencePaper explained: Pushing The Limits Of Self-Supervised ResNets: Can We Outperform Supervised…Exploring the novel approaches in ReLICv2·7 min read·Feb 26, 2022--1--1
Leon SickinTowards Data SciencePaper explained: Momentum Contrast for Unsupervised Visual Representation LearningGoing over the principles from the MoCo v1 & v2 papers·6 min read·Jan 10, 2022----
Leon SickinTowards Data SciencePaper explained: Masked Autoencoders Are Scalable Vision LearnersHow reconstructing masked parts of an image can be beneficial·7 min read·Dec 29, 2021--2--2
Leon SickinTowards Data ScienceTwo Simple Ways To Measure Your Model’s UncertaintyThe key to better understanding your model’s predictions·8 min read·Dec 28, 2021--1--1
Leon SickinTowards Data SciencePaper explained: Semi-Supervised Learning of Visual Features by Non-Parametrically Predicting View…Taking advantage of labeled support samples for semi-supervised learning·6 min read·Dec 20, 2021----
Leon SickinTowards Data ScienceHow Well Does Self-Supervised Learning Perform In The Real World?Pre-training a model on random internet images instead of ImageNet·7 min read·Dec 15, 2021----
Leon SickinTowards Data ScienceRegNet: The Most Flexible Network Architecture For Computer VisionA model design that scales for high-efficiency or high-accuracy·6 min read·Dec 13, 2021--1--1
Leon SickinTowards Data SciencePaper explained: A Simple Framework for Contrastive Learning of Visual RepresentationsGoing over the ideas presented in the SimCLR paper·6 min read·Dec 9, 2021--3--3