Machine Learning Takes The Embarrassment Out Of Video-conference Wardrobe Malfunctions

Telecommuters: tired of the constant embarrassment of showing up to video conferences wearing nothing but your underwear? Machine Learning Takes The Embarrassment Out Of Video-conference Wardrobe Malfunctions.

 Save the humiliation and all those pesky trips down to HR with Safe Meeting, the new system that uses the power of artificial intelligence to turn off your camera if you forget that casual Friday isn’t supposed to be that casual.

The following infomercial is brought to you by [Nick Bild], who says the whole thing is tongue-in-cheek but we sense a certain degree of “necessity is the mother of invention” here. It’s true that the sudden throng of remote-work newbies certainly increases the chance of videoconference mishaps and the resulting mortification, so whatever the impetus, Safe Meeting seems like a great idea. It uses a Pi cam connected to a Jetson Nano to capture images of you during videoconferences, which are conducted over another camera. The stream is classified by a convolutional neural net (CNN) that determines whether it can see your underwear. If it can, it makes a REST API call to the conferencing app to turn off the camera. The video below shows it in action, and that it douses the camera quickly enough to spare your modesty.

We shudder to think about how [Nick] developed an underwear-specific training set, but we applaud him for doing so and coming up with a neat application for machine learning. He’s been doing some fun work in this space lately, from monitoring where surfaces have been touched to a 6502-based gesture recognition system.

Thanks To Renewable And Machine Learning

Google currently Forecasts The Wind

Wind farms have historically created less cash for the electricity they turn out as a result of they need been unable to predict however windy it'll be tomorrow.

“The approach plenty of power markets work is you have got to schedule your assets every day ahead,” aforementioned archangel Terrell, the pinnacle of energy market strategy at Google. “And you tend to urge stipendiary higher after you try this than if you sell into the market time period.

“Well, however do variable assets like wind schedule every day ahead after you do not know the wind goes to blow?” Terrell asked, “and however are you able to really reserve your home in line?”

“We're not obtaining the complete profit and therefore the full worth of that power.”

Here’s how: Google and therefore the Google-owned AI firm DeepMind combined weather information with power information from 700 megawatts of wind energy that Google sources within the Central us. exploitation machine learning, they need been ready to higher predict wind production, higher predict electricity provide and demand, and as a result, cut back operative prices.

“What we've been doing is functioning in partnership with the DeepMind team to use machine learning to require the weather information that is out there in public, really forecast what we predict the wind production are succeeding day, and bid that wind into the day-ahead markets,” Terrell aforementioned in a very recent seminar hosted by the Stanford Precourt Institute of Energy. Stanford denote video of the seminar last week.

The result has been a twenty p.c increase in revenue for wind farms, Terrell aforementioned.

The Department of Energy listed improved wind prognostication as a primary priority in its 2015 Wind Vision report, mostly to enhance reliability: “Improve Wind Resource Characterization,” the report aforementioned at the highest of its list of goals. “Collect information and develop models to enhance wind prognostication at multiple temporal scales—e.g., minutes, hours, days, months, years.”

Google’s goal has been additional sweeping: to clean carbon entirely from its energy portfolio, that consumes the maximum amount power as 2 San Franciscos.

Google achieved associate degree initial milestone by matching its annual energy use with its annual renewable-energy procural, Terrell aforementioned. however the corporate has not been carbon-free in each location at each hour, that is currently its new goal—what Terrell calls its “24x7 carbon-free” goal.

“We're extremely getting down to flip our efforts during this direction, and we're finding that it is not one thing that is simple to try to to. It's arguably a rocket launching, particularly in places wherever the renewable resources of these days don't seem to be as value effective as they're in different places.”

The scientists at London-based DeepMind have incontestible that AI will facilitate by increasing the market viability of renewables at Google and on the far side.

“Our hope is that this sort of machine learning approach will strengthen the business case for alternative energy and drive more adoption of carbon-free energy on electrical grids worldwide,” aforementioned DeepMind program manager Sims American Revolutionary leader and Google engineer Carl Elkin. in a very Deepmind diary post, they define however they boosted profits for Google’s wind farms within the Southwest Power Pool, associate degree energy market that stretches across the plains from the Canadian border to north Texas:

“Using a neural network trained on wide out there weather forecasts and historical rotary engine information, we tend to designed the DeepMind system to predict wind-power output thirty six hours prior to actual generation. supported these predictions, our model recommends the way to create optimum hourly delivery commitments to the ability grid a full day prior to.”

The DeepMind system predicts wind-power output thirty six hours prior to, permitting power producers to form ... [+] additional remunerative advance bids to produce power to the grid.

Google

Artificial Intelligence - The Future Of Communication

The future of work is no longer merely a concept, but a reality — Covid-19 has made sure of that.

What role, then, does artificial intelligence (AI) have to play in this drastic shift?

For some time now, I’ve firmly maintained the belief that AI would take over the vast majority of process-driven work within 15 years. However, with years of key developments in the world of work having recently been crammed into a matter of months, the future has unfolded very differently than we imagined.

Rather than coming about through careful planning, companies have been thrust into this new way of working.

Without doubt, many were unprepared for it and have had to move quickly to put in place remote working solutions to keep business going.

They simply didn’t have the time to manage change and implement AI-driven solutions. However, many predict that we’ll see remote working becoming part of the ‘new normal’ even after lockdown measures are eased.

Companies like Twitter, for example, have already announced that their employees can work from home indefinitely.

Assuming that the growing trend towards permanent remote working continues, organisations will need to carefully consider the AI solutions they turn to for automating process-driven work.

Yet, how does this affect the security considerations required to make remote working effective for companies?

Every time we send a message to a colleague or share a company file, we share bits of data electronically.

Data, of course, is the lifeblood of every modern organisation, so when it is shared, it must be done so securely.

If we add AI into the mix, we run into a potential data security issue.

This is because, fundamentally, AI needs data to work properly.

Its purpose is to access data, analyse it and generate better outcomes for organisations through automation.

In doing so, it replaces certain tasks, but also enables employees to perform existing jobs more effectively.

Yet, despite the productivity gains, this new era of remote working doesn’t necessarily lend itself to AI being used appropriately when it comes to the secure transmission of data.

AI-driven communications are just as vulnerable to security flaws as those performed by humans.

This is especially evident when we look closely at the kind of technology enterprises use to communicate internally – a crucial component of any business model outside an office’s four walls.

According to Morten Brøgger, CEO of messaging and collaboration platform Wire, AI may not be as intertwined with the future of work as we think: “AI and the future definitely not in the collaboration and communication market or in the future of work 2.0.”

Brøgger continues: “The reason is that if you start building a lot of AI into communication tools, it means surveillance for your users, which is a clear breach of their privacy. If you do need an AI, you need to have the data to examine behavioural patterns. This means breaking end-to-end encryption, because there will be a machine that receives a copy of everything a user is doing.”

Wire is one of a crop of collaboration tools that have enabled many organisations to continue operating during COVID-19.

It found its niche by specifically targeting large enterprises back in 2017, who Brøgger believes have an advanced understanding of the importance of security and privacy.

With its leadership team comprising ex-Skype employees, the company now positions itself as being on a mission to change the way employees communicate in the workplace.

Though organisations can safely integrate AI technology, this abrupt new iteration of the workplace that we’ve suddenly found ourselves in has perhaps gone some way to exposing why AI isn’t the panacea for everything - at least not without careful security planning at the outset.

As Brøgger notes: “There were a lot of companies who were basically caught in this situation that weren’t ready for it. So, how do they put infrastructure in place that is sufficiently secure to allow people to work from home and work on things that are absolutely confidential? There are no longer any global rules – no one size fits all. That’s not how the world is, even with collaboration.”

It’s clear that the world of work is changing, though it took a pandemic to accelerate that change.

Companies need to take stock of how they can make the most of the tools available to them to reduce inefficiencies, and I still maintain that AI is, in

But it must be done in a way that doesn’t come at the expense of breaching a company’s security and putting its most valuable data assets at risk.

Alternative data and artificial intelligence - the future fuel for investors

Tomas Franczyk, Managing Director, Head of Global Information Services, Asia Pacific, Nasdaq. This is the undiscovered data from non-traditional data sources that give investors information and unique insights to help them evaluate investment opportunities.

The alternative data market, expected to become a $1.7 billion industry in the next few years, can encompass a range of sources.

These include natively digital information like web traffic, online buying habits, and social media activity, as well as more granular indicators of financial performance, such as ocean cargo and in the capital markets space, alternative data is viewed as increasingly important.

According to a 2019 survey by Greenwich Associates, 95% of trading professionals believe alternative data will become more valuable to the trading process, and 85% of banks, investors, and capital markets service providers plan to increase spending on data management.

As this new data becomes a powerful differentiator in the search for alpha, a rapidly growing community of buy-side firms have started using it to add power to quantitative and fundamental investment models with the aim of outperforming the market.

For example, Nasdaq's platform Quandl, which identifies datasets from local firms to build investment models, has partnered with large insurance companies in the United States to access insurance policies on this enables users to accurately measure car sales before automotive manufacturers report them.

This data would be extremely important, say, to hedge fund investors who need investment insights into the automotive sector.

Meanwhile, in Asia, Nasdaq is building regional-specific data products and has partnered with local fintechs and other innovative local vendors to migrate their core data and alternative data to the more sophisticated technologies mean organisations can create datasets that support managers with short-term trading strategies as well as those with a long-term approach, such as institutional investors.

Another report by Greenwich Associates last year found that 74% of firms surveyed agreed that alternative data has started to have a big impact on institutional investing, while nearly 30% of quantitative funds attribute at least 20% of their alpha to alternative data.

It can only work if it is properly interpreted and analysed.

By its nature, alternative data is harder to consume than financial data; it is often unstructured, does not follow patterns, and is created at a very fast rate.

Hence, investors now have a growing need for talent and technology, including analytics platforms, testing tools, fluid data architecture, and data science teams, to help them with their data management.

Advanced analytics and artificial intelligence (AI), such as machine learning and natural language processing, can be crucial to analysing data.

Machines can process events at roughly 2,000 times the speed of humans, digest vast datasets, and work around the clock.

During the investment process, AI-enabled data processing can increase the volume and quality of idea generation; this increase in data, including alternative data, when combined with computing power can help investment managers develop a long-lasting competitive advantage.

While some organisations are well on their way to introducing AI-based models, the industry is still understanding and identifying the operational, regulatory, and technological risks that come with the race effective risk management practices will be key for the successful adoption of AI.

Data providers have an opportunity to assist the asset management industry by making alternative data and AI the drivers of future investment research.

In fact, we may see active portfolio managers look less to the sell-side for their research needs and instead develop their own research, invest in data experts and technology, and partner with vendors to supply the information and analytical tools they need. Nasdaq offers comprehensive, bespoke, and timely data and insights to help clients build and protect assets.