The 2017 Top programming languages by IEEE Spectrum

Python jumps to number one and Swift burst in top ten
01 August 2017   1500

Spectrum is a "flagship" of Institute of Electrical and Electronics Engineers - an international non-profit association of experts in the field of engineering, the world leader in the development of standards for radio electronics, electrical engineering and hardware for computing systems and networks. 

An IEEE Spectrum team researched 12 metrics from 10 carefully chosen online sources to rank 48 languages. The main feature of the Spectrum's ranking is that it is interactive, it contains 5 ranking for 4 platforms. 

Rankings:

  • IEEE Spectrum average
  • Trending - growing rapidly
  • Jobs - in demand by employers
  • Open - popular in open-source hubs
  • Custom  - here you can create your own ranking, based on date and 12 different criteria (number of indexed online resources on Google, GitHub repos, Reddit posts, etc.)

Platforms:

  • Web
  • Mobile
  • Enterprise 
  • Embedded

Python jumped two places to the No. 1 slot, though the top four—Python, C, Java, and C++—all remain very close in popularity.

C# has reentered the top five. Ruby has fallen all the way down to 12th position, but in doing so it has given Apple’s Swift the chance to join Google’s Go in the Top Ten. This is impressive, as Swift debuted on the rankings just two years ago. (Outside the Top Ten, Apple’s Objective-C mirrors the ascent of Swift, dropping down to 26th place.) 

IEEE Spectrum Top 30 Web programming languages
IEEE Spectrum Top 30 Web programming languages

You can view an interactive ranking at Spectrum.

Students to Beat Google’s Machine-Learning Code

Student programmers' image classification algorithm successfully identifies the object in 93% of cases
13 August 2018   395

Developers-students from Fast.ai which organize free online computer training courses have created an image classification algorithm that successfully identifies the object in 93% of cases and copes with it faster than a similar Google algorithm with a similar configuration. The authors argue that "the creation of breakthrough technologies is not just for big companies". This is reported by MIT Technology Review.

When evaluating performance, the DAWNBench test was used, which calculates the speed and cost of teaching the neural network. During the Fast.ai experiment, the neural network was launched on 16 virtual AWS nodes, each contained 8 NVIDIA V100 graphics cards. This configuration achieved accuracy of 93% in 18 minutes, and the cost of machine time was estimated at $ 40. The result of Fast.ai is faster than the development of Google engineers by 40%, but the corporation uses its own clusters TPU Pod, so the comparison is not entirely objective.

The developers used the PyTorch Python library, as well as their own development - fastai. They were able to achieve this learning speed with the new method of cropping images from the ImageNet dataset: instead of square pictures, they began to use rectangular:

Fast AI
Fast AI

State-of-the-art results are not the exclusive domain of big companies. These are the obvious, dumb things that many researchers wouldn’t even think to do.
 

Jeremy Howard

Founder, Fast.AI

The authors tried to make the project accessible to everyone, so they simplified its infrastructure, refusing to use distributed computing systems and containers. To implement it, developers teamed up with engineers from the innovative division of the Pentagon (DIU) to release software to quickly create and support distributed models on AWS.