Connect with us


Jigs, reels, bits and bytes — these students are teaching a machine to play Irish music





A group of computer science students at Memorial University are hoping to use computers and artificial intelligence to generate some jigs and reels.

Zach Northcott and Caleb Graves, two of the six students working on the project, say they’re using a type of neural network called a general adversarial network, or GAN — a project they’re calling GANs and Reels.

“Basically, you can think of it almost like a cluster of nodes that are connected in similar ways to the brain, and when one activates it triggers activations elsewhere,” Northcott said.

“So, I guess it’s a bunch of little bits and bobs strung together to try and emulate knowledge or thought or learning in some sort of way.”

It learns over time … to try and generate music that convincingly passes for traditional Irish music. – Zach Northcott

Graves said they’ve had to feed the network a huge amount of traditional Irish music for it to begin to emulate the style.

“It learns over time the characteristics of the music — it learns to recognize what Irish music is, what it thinks Irish music should sound like, and it uses that knowledge to try and generate music that convincingly passes for traditional Irish music.”

And while they haven’t settled on a name for the network, they’re calling it GAN Morrison for now.

Vast amounts of Irish traditional music are fed into the GAN in hopes that it will learn how to produce its own music. (GANs & Reels/MUN Computer Science)

But does the music it generates really sound like traditional Irish music?

“Absolutely not,” said Graves.

Have a listen for yourself in the audio clip below.

A group of computer science students at Memorial University have built a neural network and fed it Irish music, in hopes that it would one day generate a jig. Here’s what it came up with. 1:02

Northcott said they believe part of the problem is in how the music is being encoded and decoded.

“You can’t give a computer a pitch or sheet music and expect it to read it like a person would, so we’ve basically taken all the notes turned them into numbers, and then have the GAN work on that,” he said.

A further challenge comes in getting the computer to play nothing at all.

“To try and represent a rest in music in our number system is difficult, because if you have middle C, you could say that’s 60, and then C sharp is 61,” said Northcott.

“But with a rest, it doesn’t actually have a pitch, so we think that representing it as a zero is skewing what it thinks of in terms of pitches.”

Improvements in leaps and bounds

Graves and Northcott said they have been working on the GAN since May, and improvements come in leaps and bounds in each successive iteration.

“Some of the changes we make will improve it significantly and some of them will sometimes even decrease it,” said Graves.

To help track their progress, the students have also developed visual representations, representing pitches as grey scale pixels to compare with real Irish tunes put through the same visualization process.

Graves and Northcott use grey scale pixels to visually represent the music the GAN produces. (GANs & Reels/MUN Computer Science)

“We can kind of use just our intuition to say, is this working, is this not working? Is it getting better, is it not getting better?” Northcott said.

“We have some basic idea of music theory and what not, but we’re not professional musicians by any regard, so having the visual cues helps us kind of refine things in a different way.” 

A group of computer science students at Memorial University have built a neural network and fed it Irish music, in hopes that it would one day generate a jig. Here’s what it came up with. 1:02 

The hope is to push the use of neural networks to be used it ways they never have before and advance the use of computers to “learn.” 

“We just want to see if there’s, I guess, applications that haven’t been considered, because traditionally, it is used for image manipulation,” said Northcott.

“But if we can perhaps represent different data in a similar way to images, perhaps we have a way to manage data and draw conclusions from it.” 

With files from On the Go

Read more articles from CBC Newfoundland and Labrador


Source link

قالب وردپرس

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *


Top 5 Analytics Trends That Are Shaping The Future





Digital transformation is increasingly becoming the focus for many CIOs around the world today—with analytics playing a fundamental role in driving the future of the digital economy.

While data is important to every business, it is necessary for businesses to have a firm grip on data analytics to allow them transform raw pieces of data into important insights. However, unlike the current trends in business intelligence—which is centred around data visualization—the future of data analytics would encompass a more contextual experience.

“The known data analytics development cycle is described in stages: from descriptive (what happened) to diagnostic (why did it happen), to discovery (what can we learn from it), to predictive (what is likely to happen), and, finally, to prescriptive analytics (what action is the best to take),” said Maurice op het Veld is a partner at KPMG Advisory in a report.

“Another way of looking at this is that data analytics initially “supported” the decision-making process but is now enabling “better” decisions than we can make on our own.”

Here are some of the current trends that arealready shaping the future of data analytics in individuals and businesses.

  1. Growth in mobile devices

With the number of mobile devices expanding to include watches, digital personal assistants, smartphones, smart glasses, in-car displays, to even video gaming systems, the final consumption plays a key role on the level of impact analytics can deliver.

Previously, most information consumers accessed were on a computer with sufficient room to view tables, charts and graphs filled with data, now, most consumers require information delivered in a format well optimized for whatever device they are currently viewing it on.

Therefore, the content must be personalized to fit the features of the user’s device and not just the user alone.

  1. Continuous Analytics

More and more businesses are relying on the Internet of Things (IoT) and their respective streaming data—which in turn shortens the time it takes to capture, analyze and react to the information gathered. Therefore, while analytics programspreviously were termed successful when results were delivered within days or weeks of processing, the future of analytics is bound to drastically reduce this benchmark to hours, minutes, seconds—and even milliseconds.

“All devices will be connected and exchange data within the “Internet of Things” and deliver enormous sets of data. Sensor data like location, weather, health, error messages, machine data, etc. will enable diagnostic and predictive analytics capabilities,” noted Maurice.

“We will be able to predict when machines will break down and plan maintenance repairs before it happens. Not only will this be cheaper, as you do not have to exchange supplies when it is not yet needed, but you can also increase uptime.”

  1. Augmented Data Preparation

During the process of data preparation, machine learning automation will begin to augment data profiling and data quality, enrichment, modelling, cataloguing and metadata development.

Newer techniques would include supervised, unsupervised and reinforcement learning which is bound to enhance the entire data preparation process. In contrast to previous processes—which depended on rule-based approach to data transformation—this current trend would involve advanced machine learning processes that would evolve based on recent data to become more precise at responding to changes in data.

  1. Augmented Data Discovery

Combined with the advancement in data preparation, a lot of these newer algorithms now allow information consumers to visualize and obtain relevant information within the data with more ease. Enhancements such as automatically revealing clusters, links, exceptions, correlation and predictions with pieces of data, eliminate the need for end users to build data models or write algorithms themselves.

This new form of augmented data discovery will lead to an increase in the number of citizen data scientist—which include information users who, with the aid of augmented assistance can now identify and respond to various patterns in data faster and a more distributed model.

  1. AugmentedData Science

It is important to note that the rise of citizen data scientist will not in any way eliminate the need for a data scientist who gathers and analyze data to discover profitable opportunities for the growth of a business. However, as these data scientists give room for citizen data scientists to perform the easier tasks, their overall analysis becomes more challenging and equally valuable to the business.

As time goes by, machine learning would be applied in other areas such as feature and model selection. This would free up some of the tasks performed by data scientist and allow them focus on the most important part of their job, which is to identify specific patterns in the data that can potentially transform business operations and ultimately increase revenue.

Continue Reading


Waterloo drone-maker Aeryon Labs bought by U.S. company for $265M






Waterloo’s Aeryon Labs has been bought by Oregon-based FLIR Systems Inc. for $256 million, or $200 million US.

The acquisition was announced Monday. 

Dave Kroetsch, co-founder and chief technology officer of Aeryon Labs, says not much will change in the foreseeable future.

“The Waterloo operations of Aeryon Labs will actually continue as they did yesterday with manufacturing, engineering and all the functions staying intact in Waterloo and ultimately, we see growing,” he said.

“The business here is very valuable to FLIR and our ability to sell internationally is a key piece of keeping these components of the business here in Canada.”

Aeroyn Labs builds high-performance drones that are sold to a variety of customers including military, police services and commercial businesses. The drones can provide high-resolution images for surveillance and reconnaissance.

The drones already include cameras and thermal technology from FLIR. Jim Cannon, president and CEO of FLIR Systems, said acquiring Aeryon Labs is part of the company’s strategy to move beyond sensors “to the development of complete solutions that save lives and livelihoods.”

‘A piece of a bigger solution’

Kroetsch said this is a good way for the company to grow into something bigger.

“We see the business evolving in much the direction our business has been headed over the last couple of years. And that’s moving beyond the drone as a product in and of itself as a drone as a piece of a bigger solution,” he said.

For example, FLIR bought a drone company that builds smaller drones that look like little helicopters.

“We can imagine integrating those with our drones, perhaps having ours carry their drones and drop them off,” he said.

FLIR also does border security systems, which Kroetsch says could use the drones to allow border agents to look over a hill where there have been issues.

“We see the opportunity there as something that we never could have done on our own but being involved with and part of a larger company that’s already providing these solutions today gives us access not only to these great applications, but also to some fantastic technologies,” he said.

Aeryon Labs has done a lot of work during emergency disasters, including in Philippines after Typhoon Hagupit in 2014, Ecuador after an earthquake in 2016 and the Fort McMurray wildfire in 2016.


Source link

قالب وردپرس

Continue Reading


Inuvik infrastructure may not be ready for climate change, says study






The Arctic is expected to get warmer and wetter by the end of this century and new research says that could mean trouble for infrastructure in Inuvik.

The study from Global Water Futures looked at how climate change could impact Havipak Creek — which crosses the Dempster Highway in Inuvik, N.W.T. — and it predicts some major water changes.

“They were quite distressing,” John Pomeroy, director of Global Water Futures and the study’s lead author, said of the findings.

Researchers used a climate model and a hydrological model to predict future weather and climate patterns in the region. They also looked at data gathered from 1960 to the present. 

If greenhouse gas emissions continue at their current rate — which Pomeroy said they are on track to do — the study projects the region will be 6.1 C warmer by 2099 and precipitation, particularly rain, will increase by almost 40 per cent.

The study also found that the spring flood will be earlier and twice as large, and the permafrost will thaw an additional 25 centimetres. While the soil is expected to be wetter early in the summer, the study said it will be drier in late summer, meaning a higher risk of wildfires.

John Pomeroy is the director of Global Water Futures. (Erin Collins/CBC)

“The model’s painting kind of a different world than we’re living in right now for the Mackenzie Delta region,” Pomeroy said.

He noted these changes are not only expected for Havipak Creek, but also for “many, many creeks along the northern part of the Dempster [Highway].”

Pomeroy said the deeper permafrost thaw and a bigger spring flood could pose challenges for buildings, roads, culverts and crossings in the area that were designed with the 20th century climate in mind.

He said the projected growth of the snowpack and the spring flood are “of grave concern because that’s what washes out the Dempster [Highway] and damages infrastructure in the area.”

Culverts and bridges may have to be adjusted to allow room for greater stream flows, Pomeroy said. And building foundations that are dependent upon the ground staying frozen will have to be reinforced or redesigned.

Pomeroy said the ultimate solution is for humans to reduce greenhouse gas emissions.

“This study is the future we’re heading for, but it’s not the future we necessarily have if we can find a way to reduce those gases,” he said.  

“It’d be far smarter to get those emissions under control than to pay the terrible expenses for infrastructure and endangered safety of humans and destroyed ecosystems.”


Source link

قالب وردپرس

Continue Reading