Deep Learning without Poor Local Minima This paper proves a conjecture about deep neural networks published in 1989 and addresses an open problem announced at COLT 2015. machine-learning mathematics papers theorems
Generative Adversarial Networks GAN originally presented by Goodfellow et al is a novel technique that uses a minmax two-player game to learn latent data distributions. gan machine-learning papers privacy-gans
Privacy Preserving Predictive Modeling GANs This paper examines a GAN architecture to generate private encodings by ensuring a three player min-max optimization that regulates information leakage. GAN machine-learning papers privacy-gans
Social Learning Networks A type of network among students, instructors and modules of learning encoding the relationships among people and learning processes. machine-learning papers sln
Breaking down Tesseract OCR Tesseract, an open source OCR project was originally developed by HP between 1984 and 1994 as a part of PhD research project at HP Labs, Bristol. vision ocr machine-learning papers
Introduction to Survival Analysis The term survival time is used to describe the length of time until a specified event. The widespread use of these models in medicine to analyze survival times leads to the name survival analysis. machine-learning papers featured
Google Smart Reply Google's smart reply, a feature available in Inbox, Gmail and Allo, saves time by suggesting quick responses to the messages. This feature already drives 12% of replies over mobile devices. NLP machine-learning papers
Bitcoin A Peer-to-Peer Electronic Cash System that prevents double spending using the hash-based proof-of-work driven network transactions, that are authenticated by global consensus. concept references papers
Improvements on Word2Vec Importance of optimization and its effect on training time for Word2Vec. Effective utilization of Hierarchical Softmax, Negative Sampling and Subsampling of frequent words. NLP machine-learning papers word-embedding