A Month of Machine Learning Paper Summaries

At the end of October 2018 I set myself a challenge (inspired by similar efforts) to write up 30 ML paper summaries, one for each day in the month of November. Their main home is a Google doc—the formatting will be more consistent there—but they’re also listed by topic below.

If you find errors—and there will inevitably be some—please comment with corrections. I hope that these summaries will be useful to others who are studying machine learning, so if any of this is helpful to you I’d love to hear about it.

Images: Style Transfer, Interpolation, and Segmentation

• A Neural Algorithm of Artistic Style
• Enhanced Deep Residual Networks for Single Image Super-Resolution
• U-Net: Convolutional Networks for Biomedical Image Segmentation
• Super SloMo: High Quality Estimation of Multiple Intermediate Frames for Video Interpolation
• Fast R-CNN
• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• Instance-aware Semantic Segmentation via Multi-task Network Cascades
• Spatial Transformer Networks

NLP & Language Models

• Efficient Estimation of Word Representations in Vector Space (word2vec part 1)
• Distributed representations of words and phrases and their compositionality (word2vec part 2)
• Enriching Word Vectors with Subword Information (fastText)
• Learned in Translation: Contextualized Word Vectors (CoVe)
• Neural Machine Translation by Jointly Learning to Align and Translate
• Regularizing and Optimizing LSTM Language Models
• Deep contextualized word representations (ELMo)
• Skip-Thought Vectors
• Universal Language Model Fine-tuning for Text Classification
• Attention Is All You Need
• Improving Language Understanding by Generative Pre-Training
• BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

Multi-Modal Models

• DeViSE: A Deep Visual-Semantic Embedding Model

Adversarial Examples

• Explaining and Harnessing Adversarial Examples
• Distillation as a Defense to Adversarial Perturbations against Deep Neural Networks
• Practical Black-Box Attacks against Machine Learning
• Universal adversarial perturbations
• Adversarial Examples Are Not Easily Detected: Bypassing Ten Detection Methods
• Synthesizing Robust Adversarial Examples
• Robust Physical-World Attacks on Deep Learning Models
• Towards Deep Learning Models Resistant to Adversarial Attacks
• Adversarial Spheres