In a bid to assess racial and gender bias in its artificial intelligence/machine learning systems, Twitter is starting a new initiative called Responsible Machine Learning.

Terming it a long journey in its early days, Twitter said the initiative will assess any "unintentional harms" caused by its algorithms.

"When Twitter uses ML, it can impact hundreds of millions of Tweets per day and sometimes, the way a system was designed to help could start to behave differently than was intended," said Jutta Williams and Rumman Chowdhury from Twitter.

"These subtle shifts can then start to impact the people using Twitter and we want to make sure we're studying those changes and using them to build a better product," they said in a statement late on Thursday.

Twitter's 'Responsible ML' working group is interdisciplinary and is made up of people from across the company, including technical, research, trust and safety, and product teams.

"Leading this work is our ML Ethics, Transparency and Accountability (META) team: a dedicated group of engineers, researchers, and data scientists collaborating across the company to assess downstream or current unintentional harms in the algorithms we use and to help Twitter prioritise which issues to tackle first," the company elaborated.

Twitter said it will research and understand the impact of ML decisions, conduct in-depth analysis and studies to assess the existence of potential harms in the algorithms it uses.

Some of the tasks will be a gender and racial bias analysis of its image cropping (saliency) algorithm, a fairness assessment of our Home timeline recommendations across racial subgroups and an analysis of content recommendations for different political ideologies across seven countries.

"The most impactful applications of responsible ML will come from how we apply our learnings to build a better Twitter," the company said.

This may result in changing its product, such as removing an algorithm and giving people more control over the images they Tweet.

Twitter said it is also building explainable ML solutions so people can better understand its algorithms, what informs them, and how they impact what they see on the platform.

--IANS

na/

(Only the headline and picture of this report may have been reworked by the Business Standard staff; the rest of the content is auto-generated from a syndicated feed.)

Business Standard has always strived hard to provide up-to-date information and commentary on developments that are of interest to you and have wider political and economic implications for the country and the world. Your encouragement and constant feedback on how to improve our offering have only made our resolve and commitment to these ideals stronger. Even during these difficult times arising out of Covid-19, we continue to remain committed to keeping you informed and updated with credible news, authoritative views and incisive commentary on topical issues of relevance.We, however, have a request.

As we battle the economic impact of the pandemic, we need your support even more, so that we can continue to offer you more quality content. Our subscription model has seen an encouraging response from many of you, who have subscribed to our online content. More subscription to our online content can only help us achieve the goals of offering you even better and more relevant content. We believe in free, fair and credible journalism. Your support through more subscriptions can help us practise the journalism to which we are committed.

Support quality journalism and subscribe to Business Standard.

Digital Editor

View original post here:
Twitter analysing harmful impacts of its AI, machine learning algorithms - Business Standard

It's amazing how commonplace the term 'algorithm' has now become, with machine learning, algorithmic-defined systems now being used to filter information to us at an increasingly efficient rate, in order to keep us engaged, keep us clicking, and keep us scrolling through our social media feeds for hours on end.

But algorithms have also become a source of rising concern in recent times, with the goals of the platforms feeding us such information often at odds with broader societal aims of increased connection and community. Indeed, various studies have found that what sparks more engagement online is content that triggers strong emotional response, with anger, for one, being a powerful driver of such. Given this, algorithms, whether intentionally or not, are basically built to fuel division, via the more practical business aim of maximizing engagement.

Sure, partisan news coverage also plays a part, as does existing bias and division. But algorithms have arguably incentivized such to a significant enough degree that such approaches now largely define, or at least influence, everything that we see.

If it feels like the world is more divided than ever, that's probably because it is, and it's likely due to the algorithms which, in effect, keep us angry all of the time.

Every platform is examining this, and the impacts of algorithms in various respects. And today, Twitter has outlined its latest algorithmic research effort, which it's calling its 'Responsible Machine Learning Initiative', which will monitor the impacts of algorithmic shifts with a view to removing various negative elements, including bias, from how it applies machine learning systems.

As explained by Twitter:

"When Twitter uses ML, it can impact hundreds of millions of Tweets per day and sometimes, the way a system was designed to help could start to behave differently than was intended. These subtle shifts can then start to impact the people using Twitter and we want to make sure were studying those changes and using them to build a better product."

The project will address four key pillars:

The broader view is that by analyzing these elements, Twitter will be able to both maximize engagement, in line with its ambitious growth targets, while also taking into account, and minimizing potential societal harms. Which may lead to difficult conflicts across the two streams -but Twitter's hoping that by instituting more specific guidance as to how it applies such, it can build a more beneficial, inclusive platform through its increased learning and development.

"The META team works to study how our systems work and uses those findings to improve the experience people have on Twitter. This may result in changing our product, such as removing an algorithm and giving people more control over the images they Tweet, or in new standards into how we design and build policies when they have an outsized impact on one particular community."

The project will also include Twitter's ambitious 'BlueSky' initiative, which essentially aims to enable users to define their own algorithms at some point, as opposed to being guided by an overarching set of platform-wide rules.

"Were also building explainable ML solutions so you can better understand our algorithms, what informs them, and how they impact what you see on Twitter. Similarly, algorithmic choice will allow people to have more input and control in shaping what they want Twitter to be for them. Were currently in the early stages of exploring this and will share more soon."

That's a far broader-reaching project, with complexities that could make it impossible for day-to-day application or use by regular people. But the idea is that by exploring specific elements, Twitter will be able to make more informed, intelligent, and fair choices as to how it applies its machine-defined rules and systems.

It's good to see Twitter taking this element on, even with the amount of challenges it will face, and hopefully, it will help the platform weed out some of the more concerning algorithmic elements, and create a better, more inclusive, less divisive system.

But I have my doubts.

The desires of idealists will generally always conflict in the demands of shareholders, and it seems like, at some stage, such investigations will lead to difficult choices that can only go one way or the other. But still, that's likely on a wider scale - maybe, by addressing at least some of these aspects, Twitter can build a better system, even if it's not perfect.

At the least, it will provide more insight into the effects of algorithms, and what that means for social platforms in general.

Read more:
Twitter Outlines Evolving Approach to Algorithms as Part of New 'Responsible Machine Learning Initiative' - Social Media Today

As machine learning (ML) integrates itself into almost every industry from automotive and healthcare to banking and manufacturing- the most exciting advancements look as if they are still yet to come. Machine learning as a subset of artificial intelligence (AI) have been among the most significant technological developments in recent history, with few fields possessing the same amount of potential to disrupt a wide range of industries.

And while many applications of ML technology go unseen, there are countless ways companies are harnessing its power in new and intriguing applications. That said, MLs revolutionary impact is most poised perhaps when put to use for age-old problems.

Hearing loss is not a new condition by any means, and people have suffered from it for centuries. The first electric hearing aid was designed in 1898 by Miller Reese Hutchison, with the first commercially manufactured hearing aids introduced in 1913.With an estimated 48 million Americans experiencing some sort of hearing loss, hearing aids can be a lifeline for many who struggle with the quality of their hearing.

And while it may seem hard to believe, todays most predominant hearing aids on the market can be painful to wear having been designed 50-100 years ago.In response to a stagnant area of development, ML is being leveraged with deep learning and advanced signal processing techniques at a level of detail previously impossible.

Through the application of software-based solutions, ML algorithms can power hearing aids to detect, predict, and suppress unwanted background noise. Neural network models take structured and unstructured data and augment it with other data sets relating to the spectrum of age, language, and voice types. The data is then refined by being fed into neural network training, which begins a process of ongoing product improvement.

In an interview with Forbes, Andre Esteva, Head of Medical AI at Salesforce says that the limits of traditional approaches have been emphasized by the manual processes involved in acquiring data to mold it into a usable format, before preparing basic algorithms to be deployed to devices. ML training protocols, on the other hand, Esteva says, automatically processes data before updating themselves to redeploy.

The effect is a significant reduction in product feedback cycles and an increase in the range of capabilities available. The beauty of this approach is that the underlying intelligence improves over time as the neural nets go through iterative training, added Esteva.

As of today, there are several companies providing AI-powered hearing aids. The most recent being Whisper, a startup that has recently obtained funds of around US$50 million as they prepare to go to production on their first product. The AI-powered hearing aids from Whisper self-tune over time, and continually improve with AI for better performance.

Elsewhere, MicroTech claims its Essentia Edge product scans environments to make changes to boost speech intelligibility, while Widexs Evoke hearing aids combine real-time input with previously learned sounds from users and millions of other listening domains.The goal of introducing machine learning technologies in healthcare is to enhance the experience of patients and users. As intelligent innovative solutions continue to emerge in a field full of noise, the buzz around revolutionary tech seems to only get louder.

See the rest here:
How machine learning can revolutionize the quality of hearing - TechHQ

Jason is working on a few of the complex processes weve been wanting to automate for some time now and hes seeing some positive results.

[Were] looking to automate the home loan process - very document driven - trying to condense that, trying to extract data they can send into our decision systems for me to make a decision, Jason says.

The really exciting part is in today's world, using the old school techniques [such as neutral networks and gradient boosted models], we can make a decision after all those processes have been conducted within four seconds.

A faster decision means customers dont need to find supplementary documentation or spend time waiting for approval. They can get their answer and focus on whats important: getting into their new home.

But its not just the home loan process thats seen the benefit of new technologies. Our Institutional team has been using machine learning for the past few years and Sreeram says even three years ago the team saw the promise the tool held. Now, theyre seeing results.

I'm excited because it is really good for our staff. You know, there's so much value added from an individual point of view because banking can be notoriously paper intensive, he says.

This is a combination of technologies and capabilities. The machine nowthe transfer of paper to image, the quality and accuracy of imaging, the ability to read, the ability to interpret and then the ability to process; this is coming together for the first time, at least in my career.

We have seen cases where 50 per cent of the manual effort before has been. We have seen cases where our internal times have improved roughly 40 to 50 per cent. So I think it's absolutely made things better.

Although Sreeram reminds us that comes with its challenges and caution and management of governance, as with any technology.

Original post:
PODCAST: rise of the machine (learning) - BlueNotes

SAN ANTONIO, April 08, 2021 (GLOBE NEWSWIRE) -- Rackspace Technology (NASDAQ: RXT), a leading end-to-end, multicloud technology solutions company, announced today its relationship with Brave Software which provides a free, open-source private and secure web browser for PC, Mac, and mobile environments.

Brave gives users a fast and private web experience, helps advertisers achieve better conversions and increases publishers revenue share. "Its machine learning functionality helps to match advertisements in the Brave Ads content categories for which Brave users would have the most interest, while preserving user privacy

Brave worked with AWS Premier Consulting Partner, Onica, a Rackspace Technology company, to improve the scalability of Braves software, increased the teams efficiency, and reduced infrastructure costs by 50 percent. Rackspace Technology used a wide range of AWS services to build cloud infrastructure tailored to Braves needs.

Before working with Rackspace Technology, Braves processes for training and deploying machine learning models were slower, involving manual steps spanning several days. Brave needed a more robust pipeline and fully automated processes.

Working with Rackspace Technology and AWS was beneficial to the continued success and scaling of Brave, said Jimmy Secretan, VP of Services and Operations, Brave Software. It substantially improved the way we created and deployed new models, which has helped us to be much more responsive to advertisers needs."

Rackspace Technology is one of only a few providers to have achieved AWS Machine Learning Competency status, said Jeff Deverter, CTO, Solutions at Rackspace Technology. This unique combination of expertise in AWS services and machine learning made us an ideal partner for Brave.

To learn more about Rackspace Technologys work and capabilities please visit http://www.rackspace.com.

About Rackspace TechnologyRackspace Technology is a leading end-to-end multicloud technology services company. We can design, build and operate our customers cloud environments across all major technology platforms, irrespective of technology stack or deployment model. We partner with our customers at every stage of their cloud journey, enabling them to modernize applications, build new products and adopt innovative technologies.

Story continues

Media ContactNatalie SilvaRackspace Technology Corporate Communicationspublicrelations@rackspace.com

Read the rest here:
Rackspace Technology Works with Brave Software to Improve Machine Learning Functionality in the Web Browser - Yahoo Finance