Research at Google and ICLR 2017

Posted by Ian Goodfellow, Staff Research Scientist, Google Brain TeamThis week, Toulon, France hosts the 5th International Conference on Learning Representations (ICLR 2017), a conference focused on how one can learn meaningful and useful representations of data for Machine Learning. ICLR includes conference and workshop tracks, with invited talks along with oral and poster presentations of some of the latest research on deep learning, metric learning, kernel learning, compositional models, non-linear structured prediction, and issues regarding non-convex optimization.At the forefront of innovation in cutting-edge technology in Neural Networks and Deep Learning, Google focuses on both theory and application, developing learning approaches to understand and generalize. As Platinum Sponsor of ICLR 2017, Google will have a strong presence with over 50 researchers attending (many from the Google Brain team and Google Research Europe), contributing to and learning from the broader academic…
Original Post: Research at Google and ICLR 2017

Experimental Nighttime Photography with Nexus and Pixel

Posted by Florian Kainz, Software Engineer, Google DaydreamOn a full moon night last year I carried a professional DSLR camera, a heavy lens and a tripod up to a hilltop in the Marin Headlands just north of San Francisco to take a picture of the Golden Gate Bridge and the lights of the city behind it. A view of the Golden Gate Bridge from the Marin Headlands, taken with a DSLR camera (Canon 1DX, Zeiss Otus 28mm f/1.4 ZE). Click here for the full resolution image. I thought the photo of the moonlit landscape came out well so I showed it to my (then) teammates in Gcam, a Google Research team that focuses on computational photography – developing algorithms that assist in taking pictures, usually with smartphones and similar small cameras. Seeing my nighttime photo, one of the Gcam team…
Original Post: Experimental Nighttime Photography with Nexus and Pixel

Jupyter Digest: 24 April 2017

Jupyter Digest. Got a project you think we’d be interested in? Submit a link. Python Cheat Sheet. This handy guide from Julian Gaal (@juliangaal on GitHub), focused on data science, provides a quick and well-formatted reference for common NumPy and Matplotlib functions. So, if you can never remember how to add ticks to a plot, split an array, or you just want to share a great trick you’ve learned, this is worth checking out. Introduction to Artificial Neural Networks and Deep Learning: A Practical Guide with Applications in Python. If you’re wondering what deep learning is all about, this open source guide by Sebastian Raschka (@rasbt) hits all the right notes (TensorFlow, RNN, etc). It’s still in the early stages (mostly just appendices about math), but if the chapters are close to this quality, it’s going to be a great…
Original Post: Jupyter Digest: 24 April 2017

Fact over Fiction

Politics is a distracting affair which I generally believe it’s best to stay out of if you want to be able to concentrate on research. Nevertheless, the US presidential election looks like something that directly politicizes the idea and process of research by damaging the association of scientists & students, funding for basic research, and creating political censorship. A core question here is: What to do? Today’s March for Science is a good step, but I’m not sure it will change many minds. Unlike most scientists, I grew up in a a county (Linn) which voted overwhelmingly for Trump. As a consequence, I feel like I must translate the mindset a bit. For the median household left behind over my lifetime a march by relatively affluent people protesting the government cutting expenses will not elicit much sympathy. Discussion about the…
Original Post: Fact over Fiction

Introduction to scikit-learn

Colored dots (source: Daniel Perez Sutil).Got a project you think we’d be interested in? Submit a link. There are several Python libraries which provide solid implementations of a range of machine learning algorithms. One of the best known is Scikit-Learn, a package that provides efficient versions of a large number of common algorithms. Scikit-Learn is characterized by a clean, uniform, and streamlined API, as well as by very useful and complete online documentation. A benefit of this uniformity is that once you understand the basic use and syntax of Scikit-Learn for one type of model, switching to a new model or algorithm is very straightforward. This section provides an overview of the Scikit-Learn API; a solid understanding of these API elements will form the foundation for understanding the deeper practical discussion of machine learning algorithms and approaches in the following…
Original Post: Introduction to scikit-learn

Scaling machine learning

Efecto matrix. (source: EEIM on Wikimedia Commons).Reza Zadeh is giving a talk, “Scaling computer vision in the cloud,” at the O’Reilly Artificial Intelligence Conference, June 26-29, 2017, in New York City. Subscribe to the O’Reilly Data Show Podcast to explore the opportunities and techniques driving big data, data science, and AI. Find us on Stitcher, TuneIn, iTunes, SoundCloud, RSS. In this episode of the Data Show, I spoke with Reza Zadeh, adjunct professor at Stanford University, co-organizer of ScaledML, and co-founder of Matroid, a startup focused on commercial applications of deep learning and computer vision. Zadeh also is the co-author of the forthcoming book TensorFlow for Deep Learning (now in early release). Our conversation took place on the eve of the recent ScaledML conference, and much of our conversation was focused on practical and real-world strategies for scaling machine learning.…
Original Post: Scaling machine learning

A new value chain for next-generation mobility

Gustave Trouvé’s tricycle, the first ever electric automobile to be shown in public, 1881-1883. (source: Bibliothèque Nationale de France via Jacques CATTELIN on Wikimedia Commons).For more on building strategies and data-driven business models, check out the Strata Business Summit at Strata Data London, May 22-25, 2017. In my book, The Big Data Opportunity in Our Driverless Future, I make two arguments: 1) that societal and urban challenges are accelerating the adoption of on-demand mobility, and 2) technology advances, including big data and machine intelligence, are making Autonomous Connected and Electrified (ACE) vehicles a reality. ACE vehicles and on-demand mobility will cause three major shifts that can lead to the disruption of the automotive and transportation industries: a consumer shift, an automotive industry shift, and a mobility services shift. In this post, I examine what is causing these shifts, the value…
Original Post: A new value chain for next-generation mobility

Cacophony for the whole family

Grade school band at Louviers, Colorado, USA (source: Named Faces from the Past, via Flickr).Got a project you think we’d be interested in? Submit a link. This is an example that demonstrates some of the features in the Think DSP library. It is inspired by the performance of a grade school band I witnessed recently. My goal is to simulate the sound of a beginner band. from future import print_function, division import thinkdsp import thinkplot import random %matplotlib inline First, a function that translates from a MIDI number to a frequency: def midi_to_freq(midi_num): “””Converts MIDI note number to frequency. midi_num: int MIDI note number returns: float frequency in Hz “”” x = (midi_num – 69) / 12.0 freq = 440.0 * 2**x return freq Now here’s a randomized version that simulates three kinds of errors: poor tuning, playing the wrong…
Original Post: Cacophony for the whole family

How do I compare document similarity using Python?

Poster (source: OReilly).Learn more about common NLP tasks in the new video training course from Jonathan Mugan, Natural Language Text Processing with Python. How do I find documents similar to a particular document? We will use a library in Python called gensim. import gensim print(dir(gensim)) Let’s create some documents. raw_documents = [“I’m taking the show on the road.”, “My socks are a force multiplier.”, “I am the barber who cuts everyone’s hair who doesn’t cut their own.”, “Legend has it that the mind is a mad monkey.”, “I make my own fun.”] print(“Number of documents:”,len(raw_documents)) We will use NLTK to tokenize. A document will now be a list of tokens. from nltk.tokenize import word_tokenize gen_docs = [[w.lower() for w in word_tokenize(text)] for text in raw_documents] print(gen_docs) We will create a dictionary from a list of documents. A dictionary maps every…
Original Post: How do I compare document similarity using Python?