Amazon's sexist AI recruiter, Nvidia gets busy, Waymo cars rack up 10 million road miles
Your two-minute guide to this week in machine-learning world
Roundup Hello, here's a quick roundup of what's been happening in AI outside of the headlines.
Machine learning is increasingly being applied to new domains, and human resources is one of them. It's a controversial area and looks like Amazon ran into the problem of creating biased models. Nvidia announced new software integrations for its GPUs, and Waymo cars have reached 10 million miles of autonomous driving.
Amazon’s sexist AI: Amazon built an AI tool to help recruit candidates, but the program was scrapped after it was found to be biased against women.
A team of developers working for Amazon’s Edinburgh hub were tasked with building hundreds of models to scout out potential candidates by inspecting resumes. The systems learned to rate people based on certain words.
For example, it preferred applications containing words more commonly found in men’s resumes such as “executed” or “captured” over words like “women’s”, according to Reuters.
This means that even if the model doesn’t explicitly take into account gender, it can still penalize female candidates for including things like “women’s netball team” on their resumes. It also looked down on graduates from two all-women’s colleges.
Machine learning systems learn how to perform specific tasks, whether its facial recognition or processing resumes, by picking up patterns from the data it’s trained on. It’s no surprise that Amazon’s system was biased against women considering it was trained on the company’s internal hiring data collected over a ten-year period. Tech companies are well known for being male dominated, especially in technical roles like software engineering.
Amazon, apparently, never used the software and canned the whole project after realizing it was sexist. The Register has asked Amazon for comment but has yet to hear back.
Nvidia GTC Europe: Nvidia’s CEO Jensen Huang was back on stage for GTC Europe to announce RAPIDS, new open source software optimized for GPUs.
It’s aimed at data scientists hoping to speed up data processing with GPUs. Nvidia claims that developers can expect up to a 50x speedup using Nvidia’s DGX-2 based on 16 Volta V100 chips, compared to CPU-only systems.
“The world’s largest industries run algorithms written by machine learning on a sea of servers to sense complex patterns in their market and environment, and make fast, accurate predictions that directly impact their bottom line,” Huang said.
“Building on CUDA and its global ecosystem, and working closely with the open-source community, we have created the RAPIDS GPU-acceleration platform. It integrates seamlessly into the world’s most popular data science libraries and workflows to speed up machine learning.”
RAPIDS is being integrated with Apache Spark, an analytics engine commonly used by companies to process data on clusters of computers.
You can start using it here.
And I will drive 10 million miles: Waymo’s autonomous cars, tested on roads across 25 different US cities, have collectively driven a total of 10 million miles.
It’s a huge number, but it pales in comparison to testing in simulation where it can rack up that number in a single day. Waymo has driven over seven billion miles in simulation. Creating fake environments allows it to test rare scenarios that are potentially dangerous in real life.
“While we’ve made great strides thanks to these 10 million miles, the next 10 million will focus on turning our advanced technology into a service that people will use and love. To best serve our riders and make it possible for more people to benefit from this technology, we need to be safe and also capable, comfortable, and convenient,” Waymo announced this week.
It hopes to expand its robo-taxi service to areas beyond Phoenix, Arizona. Right now, it’s only available to 400 specially chosen people dubbed “early riders”.
“Our applicants come from all corners of the community: senior citizens, high school students, families with young children, people living with disabilities or without a driver’s license. All said they wanted to try truly self-driving technology, and many mentioned their hopes for a future of safer streets and roadways,” it said in a previous statement. ®
We'll be examining machine learning, artificial intelligence, and data analytics, and what they mean for you, at Minds Mastering Machines in London, between October 15 and 17. Head to the website for the full agenda and ticket information.