Machine Learning Roundup: 10/7/2016

Machine learning (ML) and deep learning (DL) content from the past 24 hours, plus a few older items.

Telesign, which offers a fraud intelligence API, announces that they now use machine learning. Makes you wonder what they used before.

Miles Pritchard argues that machine learning won’t contribute to the downfall of humanity. I was worried about that.

Good Reads

Adrian Colyer summarizes a paper on knowledge base construction (KBC) with Stanford’s DeepDive.

In MIT Technology Review, Will Knight explains why the partnership between industrial robot maker Fanuc and NVIDIA, reported here on Wednesday, is a big deal.

Jack Vaughan summarizes infrastructure options for deep learning software.


On Kaggle, Outbrain launches a competition to predict which content a user will click. Competition ends 11:59pm, Wednesday 18 January 2017 UTC. First place prize is $12,000.

Big Data Vendors Predictably Dispute Gartner’s Bogus Statistics

Alex Woodie surveys top Big Data vendors, who say they see no signs of the spending slowdown reported by Gartner in June. Several comments:

  • Gartner surveyed 191 executives and found a 6% difference in the percent who report plans to invest in Big Data. Margin of error, anyone? Response bias in online surveys?
  • Gartner’s survey findings contradict a different Gartner survey that the same Gartner analyst reported at the Spark Summit in June.
  • Vendor CEOs are by nature optimistic. If there were an actual slowdown, we wouldn’t hear it from them.
  • The article quotes a Teradata executive, who says sales are “brisk”. Teradata’s product revenue has declined for seven consecutive quarters.

Other than that, it’s a great story.

Autonomous Vehicles

The Daily Mail reports one accident and one traffic violation for Uber’s self-driving cars in Pittsburgh. No word on accidents and traffic violations by human-driven cars.

Meanwhile, MIT Media Lab publishes Moral Machine, a game that explores how an autonomous vehicle should decide which pedestrian to kill. Popular Science reports.

Methods and Techniques

William Vorhies explains Spiking Neural Nets (SNNs), which he characterizes as the next generation of neural nets.

Mitsubishi Electric announces what it describes as the first Automated Design Deep Learning Algorithm, a system that designs deep learning structures to speed development of AI applications.

Ed Burns summarizes the role of visualization tools in developing machine learning applications.


ProgrammableWeb publishes what they claim is a complete list of machine learning APIs. I spot-checked the list and it looks pretty complete. The latest addition is Neuron API, which uses machine learning to extract intelligence from textual data.


Marion Marking interviews Graham Neubig, an assistant professor at Carnegie Mellon’s Language Technologies Institute, who summarizes recent developments in real-time machine translation.


Seeking Alpha’s got a fever, and the only prescription is NVIDIA.

iPerceptions acquires Datacratic’s ML-driven ad tech business.

Global automotive supplier Denso announces an investment in THINCI, Inc., a deep learning and vision processing startup.

Amazon donates $10 million to the University of Washington to build a second Computer Science and Engineering building on campus.

Microsoft donates $5 million in Microsoft Azure credits to the Alan Turing Institute.

EBay acquires Corrigan, which specializes in visual search.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.