rpg.pbem.online

Search

Items tagged with: AI

"AlphaZero had the advantage of knowing the rules of games it was tasked with playing. In pursuit of a performant machine learning model capable of teaching itself the rules, a team at DeepMind devised MuZero, which combines a tree-based search with a learned model. MuZero predicts the quantities most relevant to game planning, such that it achieves industry-leading performance on 57 different Atari games and matches the performance of AlphaZero in Go, chess, and shogi."

"The researchers say MuZero paves the way for learning methods in a host of real-world domains, particularly those lacking a simulator that communicates rules or environment dynamics."

"With respect to Go, MuZero slightly exceeded the performance of AlphaZero despite using less overall computation, which the researchers say is evidence it might have gained a deeper understanding of its position. As for Atari, MuZero achieved a new state of the art for both mean and median normalized score across the 57 games, outperforming the previous state-of-the-art method (R2D2) in 42 out of 57 games and outperforming the previous best model-based approach in all games."

DeepMind's MuZero teaches itself how to win at Atari, chess, shogi, and Go

#solidstatelife #ai #reinforcementlearning #deepmind #alphazero #muzero
DeepMind’s MuZero teaches itself how to win at Atari, chess, shogi, and Go
 
"AlphaZero had the advantage of knowing the rules of games it was tasked with playing. In pursuit of a performant machine learning model capable of teaching itself the rules, a team at DeepMind devised MuZero, which combines a tree-based search with a learned model. MuZero predicts the quantities most relevant to game planning, such that it achieves industry-leading performance on 57 different Atari games and matches the performance of AlphaZero in Go, chess, and shogi."

"The researchers say MuZero paves the way for learning methods in a host of real-world domains, particularly those lacking a simulator that communicates rules or environment dynamics."

"With respect to Go, MuZero slightly exceeded the performance of AlphaZero despite using less overall computation, which the researchers say is evidence it might have gained a deeper understanding of its position. As for Atari, MuZero achieved a new state of the art for both mean and median normalized score across the 57 games, outperforming the previous state-of-the-art method (R2D2) in 42 out of 57 games and outperforming the previous best model-based approach in all games."

DeepMind's MuZero teaches itself how to win at Atari, chess, shogi, and Go

#solidstatelife #ai #reinforcementlearning #deepmind #alphazero #muzero
DeepMind’s MuZero teaches itself how to win at Atari, chess, shogi, and Go
 

Cortana app will stop working on phones in some countries (updated) | Engadget

Microsoft's changing Cortana strategy is about to have consequences for some phone users. The company has revealed that the Cortana app will stop working after January 31st, 2020 for people in Australia, Canada and the UK. Lists, reminders and other Cortana content won't pop up in either the app or Microsoft Launcher, and a version of the Launcher will arrive after January with Cortana removed. You can still expect Cortana material to show up in the To Do app, but you won't have a central place to create it on your phone.
Will Cortana be the new Clippy?

#technology #smartphone #assistant #AI #ArtificialAssistant #Microsoft #Cortana
 

Future versions of Apple's Siri may interpret your emotions | Apple Insider

Future versions of Apple's Siri may go beyond voice recognition to enhance accuracy, taping into FaceTime cameras in the company's devices to simultaneously analyze facial reactions and emotions as users engages in dialog with the voice assistant.
#technology #computing #AI #ArtificialIntelligence #assistant #Siri
 
Learning is optimized when we fail 15% of the time. "We learn best when we are challenged to grasp something just outside the bounds of our existing knowledge. When a challenge is too simple, we don't learn anything new; likewise, we don't enhance our knowledge when a challenge is so difficult that we fail entirely or give up."

"So where does the sweet spot lie? According to the new study in the journal Nature Communications, it's when failure occurs 15% of the time. Put another way, it's when the right answer is given 85% of the time."

You would think, from this description, that they got this from doing empirical observations on thousands of human beings. But that's not what they did. They looked at neural networks. Specially those trained by the mathematical algorithm known as gradient descent. And they didn't even look at thousands of neural networks. They studied gradient descent itself and worked out an exact solution from first principles. That solution is:

Ideal failure rate = (1 - erf(1 / sqrt(2))) / 2

where erf is the Gauss error function, which is related to the Gaussian (normal) distribution. The answer comes out to about 0.1586553. That's where they get the 15%.

Learning is Optimized When We Fail 15% of the Time

#solidstatelife #ai
 
Learning is optimized when we fail 15% of the time. "We learn best when we are challenged to grasp something just outside the bounds of our existing knowledge. When a challenge is too simple, we don't learn anything new; likewise, we don't enhance our knowledge when a challenge is so difficult that we fail entirely or give up."

"So where does the sweet spot lie? According to the new study in the journal Nature Communications, it's when failure occurs 15% of the time. Put another way, it's when the right answer is given 85% of the time."

You would think, from this description, that they got this from doing empirical observations on thousands of human beings. But that's not what they did. They looked at neural networks. Specially those trained by the mathematical algorithm known as gradient descent. And they didn't even look at thousands of neural networks. They studied gradient descent itself and worked out an exact solution from first principles. That solution is:

Ideal failure rate = (1 - erf(1 / sqrt(2))) / 2

where erf is the Gauss error function, which is related to the Gaussian (normal) distribution. The answer comes out to about 0.1586553. That's where they get the 15%.

Learning is Optimized When We Fail 15% of the Time

#solidstatelife #ai
 

Facebook alters video to make people invisible to facial recognition | VentureBeat

Facebook AI Research says it’s created a machine learning system for de-identification of individuals in video. Startups like D-ID and a number of previous works have made de-identification technology for still images, but this is the first one that works on video. In initial tests, the method was able to thwart state-of-the-art facial recognition systems.
#technology #AI #ArtificialIntelligence #FacialRecognition #privacy #security
Facebook alters video to make people invisible to facial recognition
 

Amazon Echo and Google Home owners spied on by apps | BBC News

Amazon Echo and Google Home speakers have been compromised by apps modified to spy on users after being approved by the technology companies.
Berlin-based Security Research Labs (SRL) built the eight "smart spies", which were promoted as a way to deliver horoscopes and generate random numbers.
Once approved, the researchers updated the Echo Skills and Home Actions to eavesdrop and steal passwords.
They then alerted the US companies, which blocked the software.
#technology #SmartSpeaker #AI #ArtificialIntelligence #assistant #Amazon #Echo #Google #Home
 
"Multilingual machine translation processes multiple languages using a single translation model. The success of multilingual training for data-scarce languages has been demonstrated for automatic speech recognition and text-to-speech systems, and by prior research on multilingual translation. We previously studied the effect of scaling up the number of languages that can be learned in a single neural network, while controlling the amount of training data per language. But what happens once all constraints are removed? Can we train a single model using all of the available data, despite the huge differences across languages in data size, scripts, complexity and domains?"

"We push the limits of research on multilingual neural machine translation by training a single neural machine translation model on 25+ billion sentence pairs, from 100+ languages to and from English, with 50+ billion parameters."

"Once trained using all of the available data (25+ billion examples from 103 languages), we observe strong positive transfer towards low-resource languages, dramatically improving the translation quality of 30+ languages at the tail of the distribution by an average of 5 bilingual evaluation understudy (BLEU) points."

Exploring Massively Multilingual, Massive Neural Machine Translation

#solidstatelife #ai #nlp #neuralmachinetranslation
 
"Multilingual machine translation processes multiple languages using a single translation model. The success of multilingual training for data-scarce languages has been demonstrated for automatic speech recognition and text-to-speech systems, and by prior research on multilingual translation. We previously studied the effect of scaling up the number of languages that can be learned in a single neural network, while controlling the amount of training data per language. But what happens once all constraints are removed? Can we train a single model using all of the available data, despite the huge differences across languages in data size, scripts, complexity and domains?"

"We push the limits of research on multilingual neural machine translation by training a single neural machine translation model on 25+ billion sentence pairs, from 100+ languages to and from English, with 50+ billion parameters."

"Once trained using all of the available data (25+ billion examples from 103 languages), we observe strong positive transfer towards low-resource languages, dramatically improving the translation quality of 30+ languages at the tail of the distribution by an average of 5 bilingual evaluation understudy (BLEU) points."

Exploring Massively Multilingual, Massive Neural Machine Translation

#solidstatelife #ai #nlp #neuralmachinetranslation
 

Apple’s promised Siri recording opt-in feature arrives in iOS 13.2 beta | The Verge

Apple’s promised Siri recording opt-in feature is starting to roll out with the latest beta for the upcoming iOS 13.2 update, as spotted by TechCrunch, allowing customers to decide if they’d like to give audio recordings of requests to Apple to improve Siri.
#technology #computing #ArtificialIntelligence #AI #assistant #Apple #Siri
 

'Alexa, are you invading my privacy?' – the dark side of our voice assistants | The Guardian

There are more than 100m Alexa-enabled devices in our homes. But are they fun time-savers or the beginning of an Orwellian nightmare
#technology #AI #AritificialIntelligence #assistant #Alexa #security #privacy
 

Facebook to create VR world called Horizon | BBC News

Facebook is creating an immersive environment called Horizon to tempt people into spending more time in virtual reality.
The VR app will be a mix of social places where users can mingle and chat, and other areas where they can play games against each other.
People will inhabit and explore the virtual spaces via a cartoon avatar.
So, it's a sort of cartoon Second Life, then. And, presumably, you can't move vertically.

#technology #Facebook #ArtificialReality #AI #Horizon

https://www.bbc.co.uk/news/technology-49851798
 

This AI reads privacy policies so you don't have to | TNW

Don’t you absolutely hate how dense and confusing privacy policies are? Considering they’re full of gotchas and intentionally obscure legalese, it’s no surprise that hardly anyone bothers to even read them — we’ve simply accepted we’re giving up our data, and with it, our sense of privacy.

But thanks to this new policy-reading AI, things won’t have to be this way for much longer. Guard is a recurrent neural network-based app that reads and analyzes privacy terms, so you don’t have to. While it can’t yet examine policies on request, the AI has rated the privacy terms of a slew of popular services like Netflix, Instagram, Twitter, and Tinder.
#technology #AI #ArtificialIntelligence #privacy
 
Aha, so that's why Laura Nolan quit Google. She wouldn't tell me at last year's SREcon Europe, but apparently she was on the make-drones-killier project.
“I am not saying that missile-guided systems or anti-missile defence systems should be banned. They are after all under full human control and someone is ultimately accountable. These autonomous weapons however are an ethical as well as a technological step change in warfare. Very few people are talking about this but if we are not careful one or more of these weapons, these killer robots, could accidentally start a flash war, destroy a nuclear power station and cause mass atrocities.”
Ex-Google worker fears 'killer robots' could cause mass atrocities

#ai #ethics #robotics
 

AI can read your emotions. Should it? | The Guardian

Advertisers, tech giants and border forces are using face tracking software to monitor our moods – whether we like it or not

Do we really want our emotions to be machine-readable? How can we know that this data will be used in a way that will benefit citizens? Would we be happy for our employers to profile us at work, and perhaps make judgments on our stress management and overall competence?


Dr Brent Mittelstadt, a philosopher specialising in data ethics at the University of Oxford’s Internet Institute, finds that he has come across “more troubling examples than positive ones” with regard to emotion AI. “What if emotion AI is used to evaluate the honesty of an individual, or whether they pose a threat?” he says.
#technology #AI #ArtificialIntelligence
 

Could ‘fake text’ be the next global political threat? | The Guardian

Every day, millions of these unremarkable conversations unfold on Reddit, spanning from cooking techniques to geopolitics in the Western Sahara to birds with arms. But what made this conversation about egg whites noteworthy is that it was not taking place among people, but artificial intelligence (AI) bots.
#ai #deepfake #fakenews
 

MIT AI tool can predict breast cancer up to 5 years early, works equally well for white and black patients | TechCrunch

MIT’s Computer Science and Artificial Intelligence Lab has developed a new deep learning-based AI prediction model that can anticipate the development of breast cancer up to five years in advance. Researchers working on the product also recognized that other similar projects have often had inherent bias because they were based overwhelmingly on white patient populations, and specifically designed their own model so that it is informed by “more equitable” data that ensures it’s “equally accurate for white and black women.”
#health #medicine #AI #cancer #BreastCancer
MIT AI tool can predict breast cancer up to 5 years early, works equally well for white and black patients
 
Both interesting and concerning to see the advancement of technology and techniques that can provide greater accessibility yet also fake lip sync that is so difficult to detect as fake.

OMFG! That's such a good #deepfake ... i love it!

Canny AI: Imagine world leaders singing


#video #music #johnlennon #imagine #deeplearning #ai #art
Canny AI: Imagine world leaders singing
 

This Company Is Betting the Future Is Personal AI Avatars, So It Made Me One | Gizmodo

In January 2019, when China Central Television, the largest broadcast network in the most populous nation in the world, aired a special to celebrate the Lunar New Year, the hosts welcomed four life-sized “personal artificial intelligences” to share the stage with them. Called PAIs, they were three-dimensional holographic replicas of the presenters that moved, spoke, and sang to the delight of the cheering live audience. The program was viewed some 1.8 billion times. One of the most-watched TV shows in the world had been hosted by AI avatars.
And one day we might all have one or more PAIs. And I can imagine a dystopian future in which only our PAIs talk to each other...

#technology #ai #ArtificialIntelligence #avatar
 

Microsoft's Cortana will eventually sound more like a real assistant | Engadget

Virtual assistants like Microsoft's Cortana, Amazon's Alexa and Google's have finally made voice-controlled computing a reality. But talking to them still feels basic -- shouting commands isn't exactly how you'd interact with another human being. At its Build developer conference today, Microsoft gave us a glimpse at how Cortana could improve on that.
But will it work for people with a Scottish accent? 😉

#technology #computing #assistant #ai #nlp
 
#AI #surveillance #escape

This colorful printed patch makes you pretty much invisible to AI - The Verge



The rise of AI-powered surveillance is extremely worrying. The ability of governments to track and identify citizens en masse could spell an end to public anonymity. But as researchers have shown time and time again, there are ways to trick such systems.

The latest example comes from a group of engineers from the university of KU Leuven in Belgium. In a paper shared last week on the preprint server arXiv, these students show how simple printed patterns can fool an AI system that’s designed to recognize people in images.

If you print off one of the students’ specially designed patches and hang it around your neck, from an AI’s point of view, you may as well have slipped under an invisibility cloak.

As the researchers write: “We believe that, if we combine this technique with a sophisticated clothing simulation, we can design a T-shirt print that can make a person virtually invisible for automatic surveillance cameras.” (They don’t mention it, but this is, famously, an important plot device in the sci-fi novel Zero History by William Gibson.)
 

Facebook Working on AI Virtual Assistant to Be Competitor to Siri and Alexa | Make Tech Easier

If at first you don’t succeed, try, try again. Facebook is undeterred it seems after a prior failed attempt to launch a Siri competitor in 2015. It’s being reported that the social media company is working on an artificial intelligence virtual assistant once again and have been for the past year. They’re still looking to compete with Siri and Alexa.
"Hey Zuck! Delete my account."

#technology #computing #Facebook #ai #VIrtualAssistant
Facebook Working on AI Virtual Assistant to Be Competitor to Siri and Alexa
 


GauGAN turns your doodles into photorealistic landscapes

NVIDIA's deep learning AI model GauGAN, cleverly named after post-impressionist painter Paul Gauguin, turns simple sketches into realistic scenes in seconds by leveraging generative adversarial networks, or GANs, to convert segmentation maps into lifelike images.

GauGAN allows users to draw their own segmentation maps and manipulate the scene, labeling each segment with labels like sand, sky, sea or snow. The tool also allows users to add a style filter, changing a generated image to adapt the style of a particular painter, or change a daytime scene to sunset.

Source: https://blogs.nvidia.com/blog/2019/03/18/gaugan-photorealistic-landscapes-nvidia-research/

#AI #deep learning #art
 
Image/Photo
GauGAN turns your doodles into photorealistic landscapes

NVIDIA's deep learning AI model GauGAN, cleverly named after post-impressionist painter Paul Gauguin, turns simple sketches into realistic scenes in seconds by leveraging generative adversarial networks, or GANs, to convert segmentation maps into lifelike images.

GauGAN allows users to draw their own segmentation maps and manipulate the scene, labeling each segment with labels like sand, sky, sea or snow. The tool also allows users to add a style filter, changing a generated image to adapt the style of a particular painter, or change a daytime scene to sunset.

Source: https://blogs.nvidia.com/blog/2019/03/18/gaugan-photorealistic-landscapes-nvidia-research/

#AI #deep learning #art
 
Later posts Earlier posts