Issue 288: CognitionX Data Science, AI and Machine Learning Briefing

What if AI could help you land your dream job?

A new HR tech startup, Leap.ai is claiming just that. They’re using machine learning to analyse data about jobseekers, including their dream employer, their aspirations, and ideal role, and all the usual core competencies to match them with the right company and “to assemble a more complete picture of a candidate’s career aspirations”.

In other HR tech news, Cazar (HR platform) is teaming up with Microsoft Azure to add AI to their recruitment technology. In the world of funding, GoodTime, which is using AI to help recruiting teams match job applicants with the right interviewers, has recently secured $2M of seed funding.

Our research team has been working hard on mapping out the HR tech space and has identified almost 200 companies which are using
AI
to turbocharge candidate acquisition, employee analytics, general HR processes, and more.

Check out our research subscription service to see how we can help you better understand this space and turbocharge your business.

Best,

Tabitha UntilTheBotsTakeOver Goldstaub

Future of Health

App detects pancreatic cancer from the whites of your eyes

Pancreatic cancer has a very low survival rate, with just nine percent of patients surviving past five years. A major contributor to this rate is the fact that once those with pancreatic cancer start to show symptoms, the disease is usually already quite advanced. But researchers at the University of Washington have come up with a simple and incredibly accurate way to test for the cancer that people can administer themselves.

The team developed an app called BiliScreen and with a smartphone’s camera, it uses computer vision algorithms to detect levels of the chemical bilirubin in the whites of a person’s eyes. BiliScreen is able to detect very low levels of bilirubin and provide users with an assessment of whether their levels are high enough to indicate possible disease. This is easier and cheaper than a blood test, which is the traditional test for the cancer and can be done before any symptoms start to show.

Innovation

Future of underwater warfare

Manta ray motherships, self-driving eel submarines which can dissolve to avoid enemy detection and fish-shaped torpedoes, are all concept designs imagining the future of underwater warfare.  The series of futuristic submarine designs, which mimic real marine lifeforms, have been created for a Royal Navy project to show how sea defence vessels could look in 50 years’ time.

Commander Peter Pipkin, fleet robotics officer, added: “With more than 70% of the planet’s surface covered by water, the oceans remain one of the world’s great mysteries and untapped resources.” The project, named Nautilus 100, was set up to mark the 100th anniversary of the launch of the USS Nautilus, the world’s first nuclear-powered submarine.

Open Source

Google’s Tensorflow team open sources speech recognition dataset for DIY

To get you started on using deep learning for speech recognition, the TensorFlow and AIY  teams have created the Speech Commands Dataset, and used it to add training* and inference sample code to TensorFlow. The dataset has 65,000 one-second long utterances of 30 short words, by thousands of different people, contributed by members of the public through the AIY website. It’s released under a Creative Commons BY 4.0 license, and will continue to grow in future releases as more contributions are received.

The dataset is designed to let you build basic but useful voice interfaces for applications, with common words like “Yes”, “No”, digits, and directions included. The infrastructure they used to create the data has been open sourced too. You can also learn how to train your own version of this model through the new audio recognition tutorial on TensorFlow.org.

Research

When neural networks misbehave

In this paper the authors show that outsourced training introduces new security risks: an adversary can create a maliciously trained network (a backdoored neural network, or a BadNet) that has state-of-the art performance on the user’s training and validation samples, but behaves badly on specific attacker-chosen inputs. They demonstrate backdoors in a realistic scenario by creating a US street sign classifier that identifies stop signs as speed limits when a special sticker is added to the stop sign.

These results demonstrate that backdoors in neural networks are both powerful and—because the behaviour of neural networks is difficult to explicate— stealthy. This work provides motivation for further research into techniques for verifying and inspecting neural networks, just as we have developed tools for verifying and debugging software. Check out this article for more information.

Future of Transportation

Domino’s and Ford will test self-driving pizza delivery cars

Ford and Domino’s Pizza are teaming up to test self-driving pizza delivery cars in Michigan, as part of an effort to better understand how customers respond to and interact with autonomous vehicles. In the coming weeks, randomly selected Domino’s customers in Ann Arbor, Michigan will have the option to accept pizza deliveries from a Ford Fusion Hybrid autonomous research vehicle (the car will be manned by a safety engineer).

Sherif Marakby, Vice President of Ford Autonomous Vehicles and Electrification, described the project as ethnographic research in an interview with The Verge. “We don’t want to wait until we get everything done on the tech and remove the driver. We’re trying to start doing the research. We still are working on the technology, because it’s not ready to be put on public streets,” he said. “It’s simulating that the vehicle is in autonomous mode.”

Something to Get Involved In

Finalists announced for the oldest running Turing test competition

The Loebner Prize is the oldest Turing Test contest, started in 1991 by Hugh Loebner and the Cambridge Centre for Behavioural studies. Since then, a number of institutions across the globe have hosted the competition including recently, the Universities of Reading, Exeter and Ulster. From 2014, the contest is run under the aegis of the AISB, the world’s first AI society (founded 1964) at Bletchley Park where Alan Turing worked as a code-breaker during World War 2. Check out the finalists.

Impact of AI on the Environment

Drones will start patrolling for sharks in Australia

Drones using AI will be deployed off beaches in New South Wales, Australia, to detect sharks and warn swimmers.

The new drones will patrol beaches and use artificial intelligence to detect sharks in the water. It’s not the first time drones have been used in Australia to detect sharks, but what makes the new drones interesting is that they do not rely on human operators controlling the drones to look for sharks. Instead, they use AI so they can spot sharks by themselves. To improve the results, additional software is also deployed to analyse the video feed from the drones to doubly make sure that sharks aren’t missed.

I’ve been making some changes based on Feedback. Would love to hear from more of you. Please do click to share your thoughts!

Published in
0 Comments

Leave a reply

Processing...
Thank you! Your subscription has been confirmed. You'll hear from us soon.
ErrorHere

Log in with your credentials

or    

Forgot your details?

Create Account