Monthly Archives: April 2020

Did you know TNW Conference has a track fully dedicated to exploring new design trends this year? Check out the full Sprint program here.

Our intelligence is what makes us human, and AI is an extension of that quality. Yann LeCun

Thehuman species has performed incredible feats of ingenuity. We have created beautiful sculptures from a single block of marble, written enchanting sonnets that have stood for centuries and landed a craft on the face of a distant rock orbiting our planet. It is sobering then to think, that what separates us from our close, albeit far less superior cousins the chimpanzee, is a45% difference in our genomes.

I propose to you, however, that natures insatiable thirst for balance has ultimately led us to create a potential rival to our dominance as a species on this planetArtificial Intelligence. The pertinent question then becomes, what aspects of our infamous ingenuity will AI augment, and perhaps ultimately surpass?

What is AI & Machine Learning?

Essentially what some really smart people out there are trying to achieve, is a computer system that emulates human intelligence. This is the ability to make decisions that maximize the chance of the system achieving its goals. Even more important is the ability of the system to learn and evolve.

To achieve this, every system needs a starting point massive amounts of data. For example, in order to train a computer system to tell the difference between a cat and a dog, you would have to feed it with thousands of images of cats and dogs.

Read: [AI will never replace good old human creativity]

What is creativity?

Creativity is seeing what everyone else saw, and thinking what no one else thought Albert Einstein

Ive heard many people say a computer system could never be creative, and that to create art, music,or an ad campaign, one needs to feel, have a soul, and a lifetime of experiences to draw from.

Having spent over a decade in the advertising industry, I can confidently say that the best creatives I have seen, were usually the ones with the most exposure. The more you have seen, traveled or experienced, the more creative you tend to be.

Creativity is about challenging the norm,thinking differently, being the square pegs in the round holes, and evoking specific emotions in your audience. So how difficult can that be for AI to achieve? It certainly seems that in todays world, creativity is actually very arbitrary. Why? Because both this

and this

are considered valuable works of art.

The current state of AI vs Creatives

Link:

Will product designers survive the AI revolution? - The Next Web

Last month, Microsoft announced that Teams, its competitor to Slack, Facebooks Workplace, and Googles Hangouts Chat, had passed 44 million daily active users. The milestone overshadowed its unveiling of a few new features coming later this year. Most were straightforward: a hand-raising feature to indicate you have something to say, offline and low-bandwidth support to read chat messages and write responses even if you have poor or no internet connection, and an option to pop chats out into a separate window. But one feature, real-time noise suppression, stood out Microsoft demoed how the AI minimized distracting background noise during a call.

Weve all been there. How many times have you asked someone to mute themselves or to relocate from a noisy area? Real-time noise suppression will filter out someone typing on their keyboard while in a meeting, the rustling of a bag of chips (as you can see in the video above), and a vacuum cleaner running in the background. AI will remove the background noise in real time so you can hear only speech on the call. But how exactly does it work? We talked to Robert Aichner, Microsoft Teams group program manager, to find out.

The use of collaboration and video conferencing tools is exploding as the coronavirus crisis forces millions to learn and work from home. Microsoft is pushing Teams as the solution for businesses and consumers as part of its Microsoft 365 subscription suite. The company is leaning on its machine learning expertise to ensure AI features are one of its big differentiators. When it finally arrives, real-time background noise suppression will be a boon for businesses and households full of distracting noises. Additionally, how Microsoft built the feature is also instructive to other companies tapping machine learning.

Of course, noise suppression has existed in the Microsoft Teams, Skype, and Skype for Business apps for years. Other communication tools and video conferencing apps have some form of noise suppression as well. But that noise suppression covers stationary noise, such as a computer fan or air conditioner running in the background. The traditional noise suppression method is to look for speech pauses, estimate the baseline of noise, assume that the continuous background noise doesnt change over time, and filter it out.

Going forward, Microsoft Teams will suppress non-stationary noises like a dog barking or somebody shutting a door. That is not stationary, Aichner explained. You cannot estimate that in speech pauses. What machine learning now allows you to do is to create this big training set, with a lot of representative noises.

In fact, Microsoft open-sourced its training set earlier this year on GitHub to advance the research community in that field. While the first version is publicly available, Microsoft is actively working on extending the data sets. A company spokesperson confirmed that as part of the real-time noise suppression feature, certain categories of noises in the data sets will not be filtered out on calls, including musical instruments, laughter, and singing.

Microsoft cant simply isolate the sound of human voices because other noises also happen at the same frequencies. On a spectrogram of speech signal, unwanted noise appears in the gaps between speech and overlapping with the speech. Its thus next to impossible to filter out the noise if your speech and noise overlap, you cant distinguish the two. Instead, you need to train a neural network beforehand on what noise looks like and speech looks like.

To get his points across, Aichner compared machine learning models for noise suppression to machine learning models for speech recognition. For speech recognition, you need to record a large corpus of users talking into the microphone and then have humans label that speech data by writing down what was said. Instead of mapping microphone input to written words, in noise suppression youre trying to get from noisy speech to clean speech.

We train a model to understand the difference between noise and speech, and then the model is trying to just keep the speech, Aichner said. We have training data sets. We took thousands of diverse speakers and more than 100 noise types. And then what we do is we mix the clean speech without noise with the noise. So we simulate a microphone signal. And then you also give the model the clean speech as the ground truth. So youre asking the model, From this noisy data, please extract this clean signal, and this is how it should look like. Thats how you train neural networks [in] supervised learning, where you basically have some ground truth.

For speech recognition, the ground truth is what was said into the microphone. For real-time noise suppression, the ground truth is the speech without noise. By feeding a large enough data set in this case hundreds of hours of data Microsoft can effectively train its model. Its able to generalize and reduce the noise with my voice even though my voice wasnt part of the training data, Aichner said. In real time, when I speak, there is noise that the model would be able to extract the clean speech [from] and just send that to the remote person.

Comparing the functionality to speech recognition makes noise suppression sound much more achievable, even though its happening in real time. So why has it not been done before? Can Microsofts competitors quickly recreate it? Aichner listed challenges for building real-time noise suppression, including finding representative data sets, building and shrinking the model, and leveraging machine learning expertise.

We already touched on the first challenge: representative data sets. The team spent a lot of time figuring out how to produce sound files that exemplify what happens on a typical call.

They used audio books for representing male and female voices, since speech characteristics do differ between male and female voices. They used YouTube data sets with labeled data that specify that a recording includes, say, typing and music. Aichners team then combined the speech data and noises data using a synthesizer script at different signal to noise ratios. By amplifying the noise, they could imitate different realistic situations that can happen on a call.

But audiobooks are drastically different than conference calls. Would that not affect the model, and thus the noise suppression?

That is a good point, Aichner conceded. Our team did make some recordings as well to make sure that we are not just training on synthetic data we generate ourselves, but that it also works on actual data. But its definitely harder to get those real recordings.

Aichners team is not allowed to look at any customer data. Additionally, Microsoft has strict privacy guidelines internally. I cant just simply say, Now I record every meeting.'

So the team couldnt use Microsoft Teams calls. Even if they could say, if some Microsoft employees opted-in to have their meetings recorded someone would still have to mark down when exactly distracting noises occurred.

And so thats why we right now have some smaller-scale effort of making sure that we collect some of these real recordings with a variety of devices and speakers and so on, said Aichner. What we then do is we make that part of the test set. So we have a test set which we believe is even more representative of real meetings. And then, we see if we use a certain training set, how well does that do on the test set? So ideally yes, I would love to have a training set, which is all Teams recordings and have all types of noises people are listening to. Its just that I cant easily get the same number of the same volume of data that I can by grabbing some other open source data set.

I pushed the point once more: How would an opt-in program to record Microsoft employees using Teams impact the feature?

You could argue that it gets better, Aichner said. If you have more representative data, it could get even better. So I think thats a good idea to potentially in the future see if we can improve even further. But I think what we are seeing so far is even with just taking public data, it works really well.

The next challenge is to figure out how to build the neural network, what the model architecture should be, and iterate. The machine learning model went through a lot of tuning. That required a lot of compute. Aichners team was of course relying on Azure, using many GPUs. Even with all that compute, however, training a large model with a large data set could take multiple days.

A lot of the machine learning happens in the cloud, Aichner said. So, for speech recognition for example, you speak into the microphone, thats sent to the cloud. The cloud has huge compute, and then you run these large models to recognize your speech. For us, since its real-time communication, I need to process every frame. Lets say its 10 or 20 millisecond frames. I need to now process that within that time, so that I can send that immediately to you. I cant send it to the cloud, wait for some noise suppression, and send it back.

For speech recognition, leveraging the cloud may make sense. For real-time noise suppression, its a nonstarter. Once you have the machine learning model, you then have to shrink it to fit on the client. You need to be able to run it on a typical phone or computer. A machine learning model only for people with high-end machines is useless.

Theres another reason why the machine learning model should live on the edge rather than the cloud. Microsoft wants to limit server use. Sometimes, there isnt even a server in the equation to begin with. For one-to-one calls in Microsoft Teams, the call setup goes through a server, but the actual audio and video signal packets are sent directly between the two participants. For group calls or scheduled meetings, there is a server in the picture, but Microsoft minimizes the load on that server. Doing a lot of server processing for each call increases costs, and every additional network hop adds latency. Its more efficient from a cost and latency perspective to do the processing on the edge.

You want to make sure that you push as much of the compute to the endpoint of the user because there isnt really any cost involved in that. You already have your laptop or your PC or your mobile phone, so now lets do some additional processing. As long as youre not overloading the CPU, that should be fine, Aichner said.

I pointed out there is a cost, especially on devices that arent plugged in: battery life. Yeah, battery life, we are obviously paying attention to that too, he said. We dont want you now to have much lower battery life just because we added some noise suppression. Thats definitely another requirement we have when we are shipping. We need to make sure that we are not regressing there.

Its not just regression that the team has to consider, but progression in the future as well. Because were talking about a machine learning model, the work never ends.

We are trying to build something which is flexible in the future because we are not going to stop investing in noise suppression after we release the first feature, Aichner said. We want to make it better and better. Maybe for some noise tests we are not doing as good as we should. We definitely want to have the ability to improve that. The Teams client will be able to download new models and improve the quality over time whenever we think we have something better.

The model itself will clock in at a few megabytes, but it wont affect the size of the client itself. He said, Thats also another requirement we have. When users download the app on the phone or on the desktop or laptop, you want to minimize the download size. You want to help the people get going as fast as possible.

Adding megabytes to that download just for some model isnt going to fly, Aichner said. After you install Microsoft Teams, later in the background it will download that model. Thats what also allows us to be flexible in the future that we could do even more, have different models.

All the above requires one final component: talent.

You also need to have the machine learning expertise to know what you want to do with that data, Aichner said. Thats why we created this machine learning team in this intelligent communications group. You need experts to know what they should do with that data. What are the right models? Deep learning has a very broad meaning. There are many different types of models you can create. We have several centers around the world in Microsoft Research, and we have a lot of audio experts there too. We are working very closely with them because they have a lot of expertise in this deep learning space.

The data is open source and can be improved upon. A lot of compute is required, but any company can simply leverage a public cloud, including the leaders Amazon Web Services, Microsoft Azure, and Google Cloud. So if another company with a video chat tool had the right machine learners, could they pull this off?

The answer is probably yes, similar to how several companies are getting speech recognition, Aichner said. They have a speech recognizer where theres also lots of data involved. Theres also lots of expertise needed to build a model. So the large companies are doing that.

Aichner believes Microsoft still has a heavy advantage because of its scale. I think that the value is the data, he said. What we want to do in the future is like what you said, have a program where Microsoft employees can give us more than enough real Teams Calls so that we have an even better analysis of what our customers are really doing, what problems they are facing, and customize it more towards that.

See the article here:

How Microsoft Teams will use AI to filter out typing, barking, and other noise from video calls - VentureBeat

Artificial intelligence (AI) is one of the most talked-about new technologies in the business world today.

It's estimated that enterprise AI usage has increased 270% since 2015. This has coincided with a massive spike in investment, with the enterprise AI industry expected to grow to $6.1 billion by 2022.

Along with the technology's very real ability to transform the job market, exaggerated myths have also become common. The hype surrounding this branch of technology has led to a number of myths:

Myth No. 1: More Data Is The Key To AI's Success

While it's true that AI needs data in order to learn and operate efficiently, the idea that more data equals better outcomes is misleading. Not all data is created equal.

If the information fed to an AI program is labeled incorrectly or isn't relevant, it poisons the data pool. The more information AI has access to, the more precise its models and predictions will be. If the data itself is of poor quality, the outcome will be precise but not necessarily based on business reality. This can result in poor decision-making.

The truth is that the data fed to an AI solution needs to be curated and analyzed beforehand. Prioritize quality over quantity.

Myth No. 2: Companies See Immediate Value From AI investments

The integration of AI into standard operating procedures doesn't happen overnight. As seen in Myth No. 1, the data the AI uses needs to be curated and checked for relevance beforehand. This may significantly reduce the amount of information the AI has access to.

To obtain truly valuable returns, it's essential to continuously provide relevant data. Like humans, AI solutions need to be given time to learn. There may be a significant lag between when an AI-based initiative begins and when you see a return on investment.

Myth No. 3: AI Will Render Humans Obsolete

The purpose of AI is not to replace all human workers. AI is a tool businesses can use to achieve their goals. It can automate mundane processes and pull interesting insights from large data sets. When used correctly, it augments and aids human decision-making. AI provides recommendations based on trends gleaned from mountains of information. It may even pose new questions that have never been considered. A human still needs to weigh the information provided and make a final decision based on risk analysis.

Pointing out these myths in no way indicates that AI won't deliver on its transformational promise. It's easy to forget that enterprise AI adoption is still in its infancy. Even still, a 2018 Deloitte survey reported that 82% of executives said their AI projects had already led to a positive ROI. Those now implementing AI projects will be the case studies of the near future.

While there are sure to be growing pains, being on the cutting edge of this exciting technology should be beneficial. There's little doubt about how important it will be for the businesses of tomorrow. Getting a head start now, ironing out the wrinkles and locking down efficient processes will pay dividends.

Go here to see the original:

AI In The Enterprise: Reality Or Myth? - Forbes

AI streamlines acoustic ID of beluga whales

Scientists at the National Oceanic and Atmospheric Administration who study endangered beluga whales in Alaskas Cook Inlet used artificial intelligence to reduce the time they spend on analysis by 93%.

Researchers have acoustically monitored beluga whales in the waterway since 2008, but acoustic data analysis is labor-intensive because automated detection tools are relativelyarchaic in our field, Manuel Castellote, a NOAA affiliate scientist, told GCN. By improving the analysis process, we would provide resultssooner, and our research wouldbecome more efficient.

The analysis typically gets hung up in the process of validating the data because detectors pick up any acoustic signal that is similar to that of a beluga whales call or whistle. As a result, researchers get many false detections, including noise from vessel propellers, ice friction and even birds at the surface in shallow areas, Castellote said.

A machine learning model that could distinguish between actual whale calls and other sounds would provide highly accurate validation output and replace the effort of a human analyst going through thousands of detections to validatethe ones corresponding to beluga, he said.

The researchers used Microsoft AI products to develop a model with a deep neural network, a convolutional neural network, a deep residual network, and a densely connected convolutional neural network. The resulting detector that is an ensemble of these four AI models is more accurate than each of the independent models, Castellote said.

Heres how it works: Twice a year, researchers recover acoustic recorders from the seafloor. A semi-automated detector has been extracting the data and processing it, looking for tones in the recordings. It yields thousands sometimes hundreds of thousands of detections per dataset.

The team used the collection of recordings with annotated detections -- both actual beluga calls and false positives -- that it has amassed in the past 12 years to train the AI and ML tools.

Now, instead of having a data analyst sit in front of a computer for seven to 14 days to validate all these detections one by one, the unvalidated detection log is used by the ensemble model to check the recordings and validate all the detections in the log in four to five hours, Castellote said. The validated log is then used to generate plots of beluga seasonalpresence in each monitored location. These results are useful to inform management decisions.

With the significant time theyre saving, researchers can increase the number of recorders they send to the seafloor each season and focus on other aspects of data analysis, such as understanding where belugas feed based on the sounds they make when hunting prey, Castellote said. They can also study human-made noise to identify activity in the area that might harm the whales.

The team is now moving into the second phase of its collaboration with Microsoft, which involves cutting the semi-automated detector out of the process and instead applying ML directly to the sound recordings. The streamlined process will search for signals from raw data, rather than using a detection log to validate pre-detected signals.

This allows widening the detection process from beluga only to all cetaceans inhabiting Cook Inlet, Castellote said. Furthermore, it allows incorporating other target signals to be detected and classified [such as] human-made noise. Once the detection and classification processes are implemented, this approach will allow covering multiple objectives at once in our data analysis.

Castellotes colleague, Erin Moreland, will use AI this spring to monitor other mammals, too, including ice seals and polar bears. A NOAA turboprop airplane outfitted with AI-enabled cameras will fly over the Beaufort Sea scanning and classifying the imagery to produce a population count that will be ready in hours instead of months, according to a Microsoft blog post.

The work is in line with a larger NOAA push for more AI in research. On Feb. 18, the agency finalized the NOAA Artificial Intelligence Strategy. It lists five goals for using AI, including establishing organizational structures and processes to advance AI agencywide, using AI research in support of NOAAs mission and accelerating the transition of AI research to applications.

Castellote said the ensemble deep learning model hes using could easily be applied to other acoustic signal research.

A code module was built to allow retraining the ensemble, he said. Thus, any other project focused on different species (and soon human-made noise) can adapt the machine learningmodel to detect and classify signals of interest in their data.

Specifics about the model are available on GitHub.

About the Author

Stephanie Kanowitz is a freelance writer based in northern Virginia.

See more here:

AI streamlines acoustic ID of beluga whales - GCN.com