AI

An AI Wizard of Words

Marcel Gagné — Mon, 15 Jul 2019 11:00:00 +0000

A look at using OpenAI's Generative Pretrained Transformer 2 (GPT-2) to generate text.

It's probably fair to say that there's more than one person out there who is worried about some version of artificial intelligence, or AI, possibly in a robot body of some kind, taking people's jobs. Anything that is repetitive or easily described is considered fair game for a robot, so driving a car or working in a factory is fair game.

Until recently, we could tell ourselves that people like yours truly—the writers and those who create things using some form of creativity—were more or less immune to the march of the machines. Then came GPT-2, which stands for Generative Pretrained Transformer 2. I think you'll agree, that isn't the sexiest name imaginable for a civilization-ending text bot. And since it's version 2, I imagine that like Star Trek's M-5 computer, perhaps GPT-1 wasn't entirely successful. That would be the original series episode titled, "The Ultimate Computer", if you want to check it out.

So what does the name "GPT-2" stand for? Well, "generative" means pretty much what it sounds like. The program generates text based on a predictive model, much like your phone suggests the next word as you type. The "pretrained" part is also quite obvious in that the model released by OpenAI has been built and fine-tuned for a specific purpose. The last word, "Transformer", refers to the "transformer architecture", which is a neural network design architecture suited for understanding language. If you want to dig deeper into that last one, I've included a link from a Google AI blog that compares it to other machine learning architecture (see Resources).

On February 14, 2019, Valentine's Day, OpenAI released GPT-2 with a warning:

Our model, called GPT-2 (a successor to GPT), was trained simply to predict the next word in 40GB of Internet text. Due to our concerns about malicious applications of the technology, we are not releasing the trained model. As an experiment in responsible disclosure, we are instead releasing a much smaller model for researchers to experiment with, as well as a technical paper.

I've included a link to the blog in the Resources section at the end of this article. It's worth reading partly because it demonstrates a sample of what this software is capable of using the full model (see Figure 1 for a sample). We already have a problem with human-generated fake news; imagine a tireless machine capable of churning out vast quantities of news and posting it all over the internet, and you start to get a feel for the dangers. For that reason, OpenAI released a much smaller model to demonstrate its capabilities and to engage researchers and developers.

Go to Full Article

How Can We Bring FOSS to the Virtual World?

Doc Searls — Mon, 10 Dec 2018 12:30:00 +0000

by Doc Searls

Is there room for FOSS in the AI, VR, AR, MR, ML and XR revolutions—or vice versa?

Will the free and open-source revolution end when our most personal computing happens inside the walled gardens of proprietary AI VR, AR, MR, ML and XR companies? I ask, because that's the plan.

I could see that plan when I met the Magic Leap One at IIW in October (only a few days ago as I write this). The ML1 (my abbreviation) gave me an MR (mixed reality) experience when I wore all of this:

Lightwear (a headset).
Control (a handset).
Lightpack (electronics in a smooth disc about the size of a saucer).

So far, all Magic Leap offers is a Creator Edition. That was the one I met. Its price is $2,295, revealed only at the end of a registration gauntlet that requires name, email address, birth date and agreement with two click-wrap contracts totaling more than 7,000 words apiece. Here's what the page with the price says you get:

Magic Leap One Creator Edition is a lightweight, wearable computer that seamlessly blends the digital and physical worlds, allowing digital content to coexist with real world objects and the people around you. It sees what you see and uses its understanding of surroundings and context to create unbelievably believable experiences.

Also recommended on the same page are a shoulder strap ($30), a USB (or USB-like) dongle ($60) and a "fit kit" ($40), bringing the full price to $2,425.

Buying all this is the cost of entry for chefs working in the kitchen, serving apps and experiences to customers paying to play inside Magic Leap's walled garden: a market Magic Leaps hopes will be massive, given an investment sum that now totals close to $2 billion.

The experience it created for me, thanks to the work of one early developer, was with a school of digital fish swimming virtually in my physical world. Think of a hologram without a screen. I could walk through them, reach out and make them scatter, and otherwise interact with them. It was a nice demo, but far from anything I might crave.

But I wondered, given Magic Leap's secretive and far-advanced tech, if it could eventually make me crave things. I ask because immersive doesn't cover what this tech does. A better adjective might be invasive.

Go to Full Article

What Does "Ethical" AI Mean for Open Source?

Glyn Moody — Mon, 20 Aug 2018 11:30:00 +0000

by Glyn Moody

Artificial intelligence is a threat—and an opportunity—for open source.

It would be an understatement to say that artificial intelligence (AI) is much in the news these days. It's widely viewed as likely to usher in the next big step-change in computing, but a recent interesting development in the field has particular implications for open source. It concerns the rise of "ethical" AI.

In October 2016, the White House Office of Science and Technology Policy, the European Parliament's Committee on Legal Affairs and, in the UK, the House of Commons' Science and Technology Committee, all released reports on how to prepare for the future of AI, with ethical issues being an important component of those reports. At the beginning of last year, the Asilomar AI Principles were published, followed by the Montreal Declaration for a Responsible Development of Artificial Intelligence, announced in November 2017.

Abstract discussions of what ethical AI might or should mean became very real in March 2018. It was revealed then that Google had won a share of the contract for the Pentagon's Project Maven, which uses artificial intelligence to interpret huge quantities of video images collected by aerial drones in order to improve the targeting of subsequent drone strikes. When this became known, it caused a firestorm at Google. Thousands of people there signed an internal petition addressed to the company's CEO, Sundar Pichai, asking him to cancel the project. Hundreds of researchers and academics sent an open letter supporting them, and some Google employees resigned in protest.

Go to Full Article

Why the Failure to Conquer the Desktop Was Great for GNU/Linux

Glyn Moody — Mon, 06 Aug 2018 11:30:00 +0000

by Glyn Moody

AI: open source's next big win.

Canonical recently launched Ubuntu 18.04 LTS. It's an important release. In part, that's because Canonical will support it for five years, making it one of the relatively rare LTS products in Ubuntu's history. Ubuntu 18.04 also marks a high-profile return to GNOME as the default desktop, after a few years of controversial experimentation with Unity. The result is regarded by many as the best desktop Ubuntu so far (that's my view too, for what it's worth). And yet, the emphasis at launch lay elsewhere. Mark Shuttleworth, CEO of Canonical and founder of Ubuntu, said:

Multi-cloud operations are the new normal. Boot-time and performance-optimised images of Ubuntu 18.04 LTS on every major public cloud make it the fastest and most efficient OS for cloud computing, especially for storage and compute-intensive tasks like machine learning.

The bulk of the official 18.04 LTS announcement is about Ubuntu's cloud computing features. On the main web site, Ubuntu claims to be "The standard OS for cloud computing", citing (slightly old) research that shows "70% of public cloud workloads and 54% of OpenStack clouds" use it. Since Canonical is a privately held company, it doesn't publish a detailed breakdown of its operations, just a basic summary. That means it's hard to tell just how successful the cloud computing strategy is proving. But, the fact that Shuttleworth is now openly talking about an IPO—not something to be undertaken lightly—suggests that there is enough good news to convince investors to throw plenty of money at Canonical when the prospectus spells out how the business is doing.

Go to Full Article

Engineers vs. Re-engineering

Doc Searls — Thu, 02 Aug 2018 12:07:07 +0000

by Doc Searls

In an age when people are being re-engineered into farm animals for AI ranchers, it's the job of engineers to save humanity through true personal agency.

A few months ago, I was driving through Los Angeles when the Waze app on my phone told me to take the Stadium Way exit off the 110 freeway. About five other cars peeled off with me, and we became a caravan, snaking through side streets and back onto the freeway a few miles later. I knew Waze had to be in charge of us, since Waze is the navigation app of choice in Los Angeles, and it was beyond coincidence that all these cars took the same wild maze run through streets only locals knew well.

What was Waze up to here, besides offering its users (or a subset of them) a way around a jam? Was it optimizing traffic by taking some cars off the highway and leaving others on? Running an experiment only some AI understood? There was no way to tell. I doubt anyone at Waze could say exactly what was going on either. Algorithms are like that. So are the large and constantly changing data sets informing algorithms most of us with mobile devices depend on every day.

In Re-engineering Humanity, Brett Frischmann and Evan Selinger have dug deeply into what's going on behind the "cheap bliss" in our fully connected world.

What they say is that we are all subjects of techno-social engineering. In other words, our algorithmic conveniences are re-making us, much as the technologies and techniques of agriculture re-makes farm animals. And, as with farming, there's an extraction business behind a lot of it.

They say "humanity's techno-social dilemma" is that "companies, institutions, and designers regularly treat us as programmable objects through personalized technologies that are attuned to our personal histories, present behavior and feelings, and predicted futures."

And we are not innocent of complicity in this. "We outsource memory, decision-making and even our interpersonal relations...we rely on the techno-social engineers' tools to train ourselves, and in doing so, let ourselves be trained."

There are obvious benefits to "delegating physical, cognitive, emotional and ethical labor to a third party", such as Waze, but there are downsides, which Brett and Evan number: 1) passivity, 2) decreased agency, 3) decreased responsibility, 4) increased ignorance, 5) detachment and 6) decreased independence. On the road to these diminished human states, we have "fetishised computers and idealized computation".

Go to Full Article