This paper presents a technical approach to robot learning of motor skills which combines active intrinsically motivated learning with imitation learning. Our architecture, called SGIM-D, allows efficient learning of high-dimensional continuous sensorimotor inverse models in robots, and in particular learns distributions of parameterised motor policies that solve a corresponding distribution of parameterised goals/tasks. This is made possible by the technical integration of imitation learning techniques within an algorithm for learning inverse models that relies on active goal babbling. After reviewing social learning and intrinsic motivation approaches to action learning, we describe the general framework of our algorithm, before detailing its architecture. In an experiment where a robot arm has to learn to use a flexible fishing line , we illustrate that SGIM-D efficiently combines the advantages of social learning and intrinsic motivation and benefits from human demonstration properties to learn how to produce varied outcomes in the environment, while developing more precise control policies in large spaces.

For most deep learning practitioners, sequence modeling is synonymous with recurrent networks. Yet recent results indicate that convolutional architectures can outperform recurrent networks on tasks such as audio synthesis and machine translation. Given a new sequence modeling task or dataset, which architecture should one use? We conduct a systematic evaluation of generic convolutional and recurrent architectures for sequence modeling. The models are evaluated across a broad range of standard tasks that are commonly used to benchmark recurrent networks. Our results indicate that a simple convolutional architecture outperforms canonical recurrent networks such as LSTMs across a diverse range of tasks and datasets, while demonstrating longer effective memory. We conclude that the common association between sequence modeling and recurrent networks should be reconsidered, and convolutional networks should be regarded as a natural starting point for sequence modeling tasks. To assist related work, we have made code available at http://github.com/locuslab/TCN .

Launched last year, Google's AIY Projects are simple hardware kits for building AI-powered devices like an Assistant speaker and a camera with image recognition capabilities. With a clear use case in schools and STEM, Google is releasing updated 2018 kits that are easier to assemble and more widely available. Since launching in 2017, Google notes that they've seen "continued demand" for the Voice and Vision Kits "especially from the STEM audience where parents and teachers alike have found the products to be great tools for the classroom." We're taking the first of many steps to help educators integrate AIY into STEM lesson plans and help prepare students for the challenges of the future by launching a new version of our AIY kits. With "Made by you with Google" branding on the slick packaging, the AIY Voice Kit v2 includes a new Raspberry Pi Zero WH, USB connector cables, and a pre-provisioned SD card.

I am into my first term of Udacity's Self Driving Car Nanodegree and I want to share my experiences regarding one of my recent projects. The objective of this project is to basically apply the concepts of Deep Learning and Convolutional Neural Networks to teach the machine to drive car autonomously. How is this even possible? First things first, it is not magic but it really feels like magic. With just a bunch of Python libraries, some lines of Code and huge amount of Data we can teach a car to drive itself.

If you got excited for the Women of NASA and Saturn V rocket Lego sets, you'll dig this new offering from the building brick company from Denmark. Lego's education arm just announced two new robotic kits that can be used in the First Lego League series of robotics competitions, the Mission Moon and Into Orbit sets were designed in partnership with astronauts and space experts to celebrate the 20th anniversary of the Lego league itself. The new robotics kits fit right into Lego League Jr. and Lego League Mindstorms systems, which are made to help kids explore their science, technology, engineering and math (STEM) skills. The Mission Moon challenge will reach over 86,000 students ages 6 to 10 from 41 countries around the globe, while the Into Orbit Challenge will offer more than 280,000 kids from age 9 to 16 in almost 90 different countries. The younger kids will build the space-themed model and use Lego's WeDo software to make it move, while the older students will design, build and code the more advanced Mindstorms robot to complete multi-step missions on a playing field.

Optimizing the performance of marketing campaigns in the digital age relies on being able to analyze and interpret large quantities of data. The more platforms and measurement capabilities have grown, the more data there is for marketers to consume. Handing over the job of gathering, cleaning, processing and interpreting vast quantities of data from multiple sources to AI is a huge opportunity for marketers. Machine learning algorithms can analyze, understand and identify nuanced patterns in big data that would be undetectable to humans. They can also perform unlimited split tests in real-time to continue learning and adapting throughout the course of a campaign.

Machine learning and artificial intelligence (AI) are two of the most exciting technologies currently being developed and applied within industries around the world. Their potential is almost limitless and their gradual implementation in various fields and sectors have been mostly met with widespread acclaim. The education sector is one of the industries set to benefit from the implementation of machine learning technologies in and around the classroom and various applications already exist that are helping schools, colleges and universities in making the transition to these new, ground breaking technologies. In this article, we'll be looking at five ways in which machine learning is transforming both education and the classroom, however, before we get into that, let's take a brief look at how machine learning came to be one of the most exciting technologies to ever hit the classroom. The idea of artificial intelligence being used within the education sector is an idea almost as old as the concept of AI itself.

Machine learning algorithms excel at finding complex patterns within big data, so researchers often use them to make predictions. Researchers are pushing this emerging technology beyond finding correlations to help uncover hidden cause-effect relationships and drive scientific discoveries. At the University of South Florida, researchers are integrating machine learning techniques into their work studying proteins. The researchers report that one of their main challenges has been a lack of methods to identify cause-effect relationships in data obtained from molecular dynamics simulations. Machine learning-based analysis of the signaling pathways found inside amino acids found in human proteins.

Experienced lawyers in the US have been left behind by AI when it came to reviewing legal documents according to a new report – with the lawyers exhibiting 85% average accuracy compared to 94% average accuracy rate achieved by AI software. This revelation is based on a study carried out by professors at Duke Law, University of Southern California, and Stanford Law School. Metaphorically, the study was a race between LawGeex, an AI contract review platform provider, and a team of 20 top corporate lawyers with notable experience particularly in reviewing Non-Disclosure Agreements (NDAs). For the study, the lawyers and the LawGeex AI had to analyse five previously unseen contracts with 153 paragraphs of technical legal language, under controlled conditions precisely prepared the way lawyers review and approve everyday contracts. The highest performing lawyer stood in line with LawGeex AI by achieving 94% accuracy but the average accuracy achieved by the least performing lawyer stood at just 67%.

Somewhere between my eighth and eighteenth turmeric lattes, I realized I was dangerously close to falling for TED. The annual conference, which gathers elite technologists, thought leaders, scientists, economists, futurists, visionaries, activists, physicists, poets, enthusiasts, academics, entertainers and billionaires has a binary reputation: For anyone who hasn't been, it's an object of easy mockery. For anyone who has, it's a religion. After five days in the garden of TED, downing blueberry mint kombucha, champagne gummy bears and green juice described as "good for when you feel like you're being chased by a cheetah," I had seen the light. The ideas felt exciting (flying cars!