By Jessica Montgomery, Senior Policy Adviser
Advances in artificial intelligence (AI) technologies over the last five years have generated great excitement, especially in the areas of health and care.
I was at RoboCup last week in Sydney as part of the Junior League. Here’s a quick video tour of RoboCup, and a behind the scenes look at the Rapidly Manufactured Robot Challenge (RMRC), which is part of the RoboCupRescue Robot competition.
For the final stage of the @Home competition, the top 2 teams were selected in each of the sub-leagues (domestic, social and open platforms) to perform demos inside a 10 minute window followed by 5 minutes of questions from the judges. The scoring judges were comprised by trustee board members, external judges and members of the technical and organizing committees.
The Standard Platform League’s Competition Day 3 at RoboCup 2019 was busy with main competition games, mixed team games and a directional whistle challenge.
After three test blocks, about half of the teams in all three sub-leagues advanced to the 2nd round of the competition.
The Standard Platform League’s Competition Day 2 at RoboCup 2019 was busy with main competition games, mixed team games and an open challenge.
Twenty teams are competing in the Standard Platform League (SPL) at RoboCup 2019 in Sydney this week. However, the experience and skills of the league’s teams vary greatly, from six-time champion B-Human to new teams from Russia and Colombia. Although teams have been onsite for two days preparing, today was the first day of competition.
RoboCup 2019 in Sydney opens to the public tomorrow, but participants have already been preparing onsite for two days. There are fifteen RoboCup leagues that are grouped into five categories — RoboCupSoccer, RoboCupRescue, RoboCup@Home, RoboCupIndustrial and RoboCupJunior — and live coverage on AIHub will touch each league.
However, I’ll personally be covering the Standard Platform League (SPL) with one post per day throughout RoboCup 2019.
The International Conference on Machine Learning took place this month in Long Beach, California. Over one week, the conference was home to 6000+ machine learning experts from academia and industry.
Almost all content at ICML 2019 was live streamed – which means you can hear talks from the likes of Jeff Clune, Raia Hadsell, or Yann Lecun. I’ve embedded as many streams as I could below for ease of navigation, but you should also check out the ICML Facebook Page for the full list.
The International Conference on Machine Learning took place this month in Long Beach, California. Over one week, the conference was home to 6000+ machine learning experts from academia and industry.
Almost all content at ICML 2019 was live streamed – which means you can hear talks from the likes of Jeff Clune, Raia Hadsell, or Yann Lecun. I’ve embedded as many streams as I could below for ease of navigation, but you should also check out the ICML Facebook Page for the full list.
In this blog post we introduce Population Based Augmentation (PBA), an algorithm that quickly and efficiently learns a state-of-the-art approach to augmenting data for neural network training. PBA matches the previous best result on CIFAR and SVHN but uses one thousand times less compute, enabling researchers and practitioners to effectively learn new augmentation policies using a single workstation GPU. You can use PBA broadly to improve deep learning performance on image recognition tasks.
We discuss the PBA results from our recent paper and then show how to easily run PBA for yourself on a new data set in the Tune framework.
