Visual aesthetics are very personal, often subconscious, and hard to express.  In a world with an overload of photographic content, a lot of time and effort is spent manually curating photographs, and it’s often hard to separate the good images from the visual noise. The question we put forward at EyeEm is: can a machine learn personalized aesthetics embodied in a set of chosen photos, and recreate them in a different set?


https://devblogs.nvidia.com/parallelforall/personalized-aesthetics-machine-learning/

#cv #deeplearning

Ruslan Salakhutdinov, Apple AI leader, thoughts on deep learning:

https://www.technologyreview.com/s/603912/apples-ai-director-heres-how-to-supercharge-deep-learning/

#apple #deeplearning #theory

Google came up with special TPU (Tensor Processing Unit). So we can expect more efficient solutions, involving AI.

https://www.wired.com/2017/04/building-ai-chip-saved-google-building-dozen-new-data-centers/

Shopping cart sequence learning explained with emojis

https://tech.instacart.com/deep-learning-with-emojis-not-math-660ba1ad6cdc

#deeplearning #medium

Unsupervised sentiment neuron which learns an excellent representation of sentiment, despite being trained only to predict the next character in the text of Amazon reviews.


https://blog.openai.com/unsupervised-sentiment-neuron/

#nlp #sentiment #openai

Karpathy on ML trends for last 5 years:

https://medium.com/@karpathy/a-peek-at-trends-in-machine-learning-ab8a1085a106

#meta #arxiv

OpenAI released $5,000 grant announcement. You can recieve a grant if you develop open source AI project.

https://nat.org/aigrant-get-5-000-for-your-open-source-ai-project-1118dd7db083?gi=f68a512e1c1

#grant #openai

Google released a tool for drawing search, which is a nice attempt to get some hype from recent pix2pix architecture demo.

Basically now you can draw a doodle and let Google search for similar pictures.

https://www.blog.google/topics/machine-learning/fast-drawing-everyone/

Nice post about how you can start solving ML tasks:

Leaf Classification Competition: 1st Place Winner's Interview, Ivan Sosnovik

http://blog.kaggle.com/2017/03/24/leaf-classification-competition-1st-place-winners-interview-ivan-sosnovik/

#kaggle #ml #interview #novice2master

T-SNE visualization by Vadim Markovtsev

Jonker-Volgenant Algorithm + t-SNE = Super Powers: https://blog.sourced.tech/post/lapjv/

#tsne #visualization

Google has set up a new milestone for speech generation: "Tacotron: A Fully End-to-End Text-To-Speech Synthesis Model"

You can listen to generated samples at: https://google.github.io/tacotron/

Paper: https://arxiv.org/abs/1703.10135

#audio #arxiv #google #breakthrough #generative

The Verge’s piece on how Alibaba uses AI to study playing Starcraft. Probably this is the way to attract new brilliant researches into the company, netherless, results look really promising.

http://www.theverge.com/2017/4/3/15164490/alibaba-ai-starcraft-combat

#reinforcementlearning #alibaba #starcraft

Tech Review on Facebook incredible chatbot:

Facebook is quietly trying to develop the most useful virtual assistant ever, in a project that illustrates the current limitations of artificial intelligence.

https://www.technologyreview.com/s/604117/facebooks-perfect-impossible-chatbot/

#lyrics #mit #nlp #qa #chatbot

The Conversational Intelligence Challenge

NIPS 2017 Live Competition

Recent advances in machine learning have sparked a renewed interest for dialogue systems in the research community. In addition to the growing real-world applications, the ability to converse is closely related to the overall goal of AI. This NIPS Live Competition aims to unify the community around the challenging task: building systems capable of intelligent conversations. Teams are expected to submit dialogue systems able to carry out intelligent and natural conversations about specific news articles with humans. At the final stage of the competition participants, as well as volunteers, will be randomly matched with a bot or a human to chat and evaluate answers of a peer. We expect the competition to have two major outcomes: (1) a measure of quality of state-of-the-art dialogue systems, and (2) an open-source dataset collected from evaluated dialogues.

Organizers

Mikhail Burtsev, Valentin Malykh, MIPT, Moscow
Ryan Lowe, McGill University, Montreal
Iulian Serban, Yoshua Bengio, University of Montreal, Montreal
Alexander Rudnicky, Alan W. Black, Carnegie Mellon University, Pittsburgh

http://convai.io

#nlp #qa #nips

Outdated, but still valuable blog to dive into deep learning:

http://deeplearning.net/tutorial/

#tutorial #wheretostart

Nothing special, just siberian mathematicians went down to streets on national Labour Day holiday.

Spotify + Emoji = ❤️

https://insights.spotify.com/es/2017/05/02/spotify-emoji-music/

#insights #emoji #spotify

Deep Learning Papers by task

https://github.com/sbrugman/deep-learning-papers

#deeplearning #list #github