In the Artificial Intelligence Bulletin – OpenAI’s Robot Hand Learns Dexterity we explore some of the latest AI news, as well as returning to look at a recent topic, AI and music production.
We will also look at some recent developments involving AI and Google Maps as well as at big hitting A.I. companies such as IBM.
A theme that runs through some of the articles below is that AI algorithms are becoming more user-friendly and open for more widespread adoption. This can be seen in the articles about Amper Sound and IBM.
OpenAI’s Robot Hand Learns Dexterity
In this video from OpenAI, we see first hand the progress being made in one of the trickiest tasks robots can be given to replicate human behaviour… Being able to show dexterity when picking up and manipulating different sized objects.
OpenAI is a non-profit research organization co-founded by Elon Musk in order to keep some of the key knowledge and research in AI in the public open source domain. It is backed by some of the big players in the AI field, such as NVIDIA.
It’s mission is to build ‘safe’ AGI (artificial general intelligence). AGI is the stage at which computers will become as ‘smart’ as humans and could potentially build their own technology.
Watch the video below about OpenAI’s robot, Dactyl:
Learn more at OpenAI’s blog article on Dactyl
How AI-Generated Music is Changing the Way Hits are Made
Let’s catch up with a recent topic we explored concerning AI and music production.
This is an interesting article from the Verge including interviews with Taryn Southern, the musician who created a whole album using AI. They also talk to the co-founder of Amper Music, Michael Hobe, who offer the most user-friendly AI music producer on the market.
One thing both Southern and Hobe emphasize is that AI is not meant to replace human musicians or producers. It is meant to help creative people more easily find solutions when working with new instruments or needing new ideas.
Amper Music sounds like a fun tool to try out due to its ease of use, as described in the article:
All you have to do is go to the website and pick a genre of music and a mood. That’s it. You don’t have to know code or composition or even music theory in order to make a song with it. It builds tracks from prerecorded samples and spits out actual audio, not MIDI. From there, you can change the tempo, the key; mute individual instruments, or switch out entire instrument kits to shift the mood of the song its made. This audio can then be exported as a whole or as individual layers of instruments (known as “stems”). Stems can then be further manipulated in DAWs like Ableton or Logic.
Read more and listen to clips created in Amper at the Verge
Check out the recent blog post regarding the Future of Music below: (article continues beneath)
How Artificial Intelligence Estimates Obesity Levels From Google Map Photos
We’ve seen AI algorithms use deep learning to analyze and recognize cats versus dogs to a very accurate degree. Now, researchers have shown how AI, using a convolutional neural network (CNN or ConvNet) can quite accurately predict the levels of obesity in six US cities based on the geography (number and size of parks, waterways, greenland etc.)
ConvNet is a type of AI algorithm that uses visual learning based on biology to recognize patterns.
The algorithm does not give perfect results since there are more variables involved than simply access to certain types of recreational areas. However, it does a pretty accurate job of mapping out which parts of a city would be more likely to have higher and less levels of obesity.
Read more at Forbes
IBM’s New System Automatically Selects the Optimal AI Algorithm
Pretty fascinating stuff here. Now we have an AI algorithm that work as matchmakers for AI algorithms and research projects.
The problem IBM is trying to fix is that sometimes researchers have to take a long time to find the suitable AI algorithm for their projects.
IBM have now trialed an AI algorithm that does the job for you.
“At IBM, engineers and scientists select the best architecture for a deep learning model from a large set of possible candidates. Today this is a time-consuming manual process; however, using a more powerful automated AI solution to select the neural network can save time and enable non-experts to apply deep learning faster.” – Martin Wistuba, Data Scientist at IBM Research Ireland
Read more at Venture Beat
Nikolas is a world-leading Futurist Speaker that drives leaders to take action in creating a better world for humanity. He promotes exponential thinking along with a critical, honest, and optimistic view that empowers you with knowledge to plan for today, tomorrow, and for the future.
Read previous related posts from Nikolas Badminton: