Tag Archives: condition
#437673 Can AI and Automation Deliver a COVID-19 ...
Illustration: Marysia Machulska
Within moments of meeting each other at a conference last year, Nathan Collins and Yann Gaston-Mathé began devising a plan to work together. Gaston-Mathé runs a startup that applies automated software to the design of new drug candidates. Collins leads a team that uses an automated chemistry platform to synthesize new drug candidates.
“There was an obvious synergy between their technology and ours,” recalls Gaston-Mathé, CEO and cofounder of Paris-based Iktos.
In late 2019, the pair launched a project to create a brand-new antiviral drug that would block a specific protein exploited by influenza viruses. Then the COVID-19 pandemic erupted across the world stage, and Gaston-Mathé and Collins learned that the viral culprit, SARS-CoV-2, relied on a protein that was 97 percent similar to their influenza protein. The partners pivoted.
Their companies are just two of hundreds of biotech firms eager to overhaul the drug-discovery process, often with the aid of artificial intelligence (AI) tools. The first set of antiviral drugs to treat COVID-19 will likely come from sifting through existing drugs. Remdesivir, for example, was originally developed to treat Ebola, and it has been shown to speed the recovery of hospitalized COVID-19 patients. But a drug made for one condition often has side effects and limited potency when applied to another. If researchers can produce an antiviral that specifically targets SARS-CoV-2, the drug would likely be safer and more effective than a repurposed drug.
There’s one big problem: Traditional drug discovery is far too slow to react to a pandemic. Designing a drug from scratch typically takes three to five years—and that’s before human clinical trials. “Our goal, with the combination of AI and automation, is to reduce that down to six months or less,” says Collins, who is chief strategy officer at SRI Biosciences, a division of the Silicon Valley research nonprofit SRI International. “We want to get this to be very, very fast.”
That sentiment is shared by small biotech firms and big pharmaceutical companies alike, many of which are now ramping up automated technologies backed by supercomputing power to predict, design, and test new antivirals—for this pandemic as well as the next—with unprecedented speed and scope.
“The entire industry is embracing these tools,” says Kara Carter, president of the International Society for Antiviral Research and executive vice president of infectious disease at Evotec, a drug-discovery company in Hamburg. “Not only do we need [new antivirals] to treat the SARS-CoV-2 infection in the population, which is probably here to stay, but we’ll also need them to treat future agents that arrive.”
There are currentlyabout 200 known viruses that infect humans. Although viruses represent less than 14 percent of all known human pathogens, they make up two-thirds of all new human pathogens discovered since 1980.
Antiviral drugs are fundamentally different from vaccines, which teach a person’s immune system to mount a defense against a viral invader, and antibody treatments, which enhance the body’s immune response. By contrast, antivirals are chemical compounds that directly block a virus after a person has become infected. They do this by binding to specific proteins and preventing them from functioning, so that the virus cannot copy itself or enter or exit a cell.
The SARS-CoV-2 virus has an estimated 25 to 29 proteins, but not all of them are suitable drug targets. Researchers are investigating, among other targets, the virus’s exterior spike protein, which binds to a receptor on a human cell; two scissorlike enzymes, called proteases, that cut up long strings of viral proteins into functional pieces inside the cell; and a polymerase complex that makes the cell churn out copies of the virus’s genetic material, in the form of single-stranded RNA.
But it’s not enough for a drug candidate to simply attach to a target protein. Chemists also consider how tightly the compound binds to its target, whether it binds to other things as well, how quickly it metabolizes in the body, and so on. A drug candidate may have 10 to 20 such objectives. “Very often those objectives can appear to be anticorrelated or contradictory with each other,” says Gaston-Mathé.
Compared with antibiotics, antiviral drug discovery has proceeded at a snail’s pace. Scientists advanced from isolating the first antibacterial molecules in 1910 to developing an arsenal of powerful antibiotics by 1944. By contrast, it took until 1951 for researchers to be able to routinely grow large amounts of virus particles in cells in a dish, a breakthrough that earned the inventors a Nobel Prize in Medicine in 1954.
And the lag between the discovery of a virus and the creation of a treatment can be heartbreaking. According to the World Health Organization, 71 million people worldwide have chronic hepatitis C, a major cause of liver cancer. The virus that causes the infection was discovered in 1989, but effective antiviral drugs didn’t hit the market until 2014.
While many antibiotics work on a range of microbes, most antivirals are highly specific to a single virus—what those in the business call “one bug, one drug.” It takes a detailed understanding of a virus to develop an antiviral against it, says Che Colpitts, a virologist at Queen’s University, in Canada, who works on antivirals against RNA viruses. “When a new virus emerges, like SARS-CoV-2, we’re at a big disadvantage.”
Making drugs to stop viruses is hard for three main reasons. First, viruses are the Spartans of the pathogen world: They’re frugal, brutal, and expert at evading the human immune system. About 20 to 250 nanometers in diameter, viruses rely on just a few parts to operate, hijacking host cells to reproduce and often destroying those cells upon departure. They employ tricks to camouflage their presence from the host’s immune system, including preventing infected cells from sending out molecular distress beacons. “Viruses are really small, so they only have a few components, so there’s not that many drug targets available to start with,” says Colpitts.
Second, viruses replicate quickly, typically doubling in number in hours or days. This constant copying of their genetic material enables viruses to evolve quickly, producing mutations able to sidestep drug effects. The virus that causes AIDS soon develops resistance when exposed to a single drug. That’s why a cocktail of antiviral drugs is used to treat HIV infection.
Finally, unlike bacteria, which can exist independently outside human cells, viruses invade human cells to propagate, so any drug designed to eliminate a virus needs to spare the host cell. A drug that fails to distinguish between a virus and a cell can cause serious side effects. “Discriminating between the two is really quite difficult,” says Evotec’s Carter, who has worked in antiviral drug discovery for over three decades.
And then there’s the money barrier. Developing antivirals is rarely profitable. Health-policy researchers at the London School of Economics recently estimated that the average cost of developing a new drug is US $1 billion, and up to $2.8 billion for cancer and other specialty drugs. Because antivirals are usually taken for only short periods of time or during short outbreaks of disease, companies rarely recoup what they spent developing the drug, much less turn a profit, says Carter.
To change the status quo, drug discovery needs fresh approaches that leverage new technologies, rather than incremental improvements, says Christian Tidona, managing director of BioMed X, an independent research institute in Heidelberg, Germany. “We need breakthroughs.”
Putting Drug Development on Autopilot
Earlier this year, SRI Biosciences and Iktos began collaborating on a way to use artificial intelligence and automated chemistry to rapidly identify new drugs to target the COVID-19 virus. Within four months, they had designed and synthesized a first round of antiviral candidates. Here’s how they’re doing it.
1/5
STEP 1: Iktos’s AI platform uses deep-learning algorithms in an iterative process to come up with new molecular structures likely to bind to and disable a specific coronavirus protein. Illustrations: Chris Philpot
2/5
STEP 2: SRI Biosciences’s SynFini system is a three-part automated chemistry suite for producing new compounds. Starting with a target compound from Iktos, SynRoute uses machine learning to analyze and optimize routes for creating that compound, with results in about 10 seconds. It prioritizes routes based on cost, likelihood of success, and ease of implementation.
3/5
STEP 3: SynJet, an automated inkjet printer platform, tests the routes by printing out tiny quantities of chemical ingredients to see how they react. If the right compound is produced, the platform tests it.
4/5
STEP 4: AutoSyn, an automated tabletop chemical plant, synthesizes milligrams to grams of the desired compound for further testing. Computer-selected “maps” dictate paths through the plant’s modular components.
5/5
STEP 5: The most promising compounds are tested against live virus samples.
Previous
Next
Iktos’s AI platform was created by a medicinal chemist and an AI expert. To tackle SARS-CoV-2, the company used generative models—deep-learning algorithms that generate new data—to “imagine” molecular structures with a good chance of disabling a key coronavirus protein.
For a new drug target, the software proposes and evaluates roughly 1 million compounds, says Gaston-Mathé. It’s an iterative process: At each step, the system generates 100 virtual compounds, which are tested in silico with predictive models to see how closely they meet the objectives. The test results are then used to design the next batch of compounds. “It’s like we have a very, very fast chemist who is designing compounds, testing compounds, getting back the data, then designing another batch of compounds,” he says.
The computer isn’t as smart as a human chemist, Gaston-Mathé notes, but it’s much faster, so it can explore far more of what people in the field call “chemical space”—the set of all possible organic compounds. Unexplored chemical space is huge: Biochemists estimate that there are at least 1063 possible druglike molecules, and that 99.9 percent of all possible small molecules or compounds have never been synthesized.
Still, designing a chemical compound isn’t the hardest part of creating a new drug. After a drug candidate is designed, it must be synthesized, and the highly manual process for synthesizing a new chemical hasn’t changed much in 200 years. It can take days to plan a synthesis process and then months to years to optimize it for manufacture.
That’s why Gaston-Mathé was eager to send Iktos’s AI-generated designs to Collins’s team at SRI Biosciences. With $13.8 million from the Defense Advanced Research Projects Agency, SRI Biosciences spent the last four years automating the synthesis process. The company’s automated suite of three technologies, called SynFini, can produce new chemical compounds in just hours or days, says Collins.
First, machine-learning software devises possible routes for making a desired molecule. Next, an inkjet printer platform tests the routes by printing out and mixing tiny quantities of chemical ingredients to see how they react with one another; if the right compound is produced, the platform runs tests on it. Finally, a tabletop chemical plant synthesizes milligrams to grams of the desired compound.
Less than four months after Iktos and SRI Biosciences announced their collaboration, they had designed and synthesized a first round of antiviral candidates for SARS-CoV-2. Now they’re testing how well the compounds work on actual samples of the virus.
Out of 10
63 possible druglike molecules, 99.9 percent have never been synthesized.
Theirs isn’t the only collaborationapplying new tools to drug discovery. In late March, Alex Zhavoronkov, CEO of Hong Kong–based Insilico Medicine, came across a YouTube video showing three virtual-reality avatars positioning colorful, sticklike fragments in the side of a bulbous blue protein. The three researchers were using VR to explore how compounds might bind to a SARS-CoV-2 enzyme. Zhavoronkov contacted the startup that created the simulation—Nanome, in San Diego—and invited it to examine Insilico’s AI-generated molecules in virtual reality.
Insilico runs an AI platform that uses biological data to train deep-learning algorithms, then uses those algorithms to identify molecules with druglike features that will likely bind to a protein target. A four-day training sprint in late January yielded 100 molecules that appear to bind to an important SARS-CoV-2 protease. The company recently began synthesizing some of those molecules for laboratory testing.
Nanome’s VR software, meanwhile, allows researchers to import a molecular structure, then view and manipulate it on the scale of individual atoms. Like human chess players who use computer programs to explore potential moves, chemists can use VR to predict how to make molecules more druglike, says Nanome CEO Steve McCloskey. “The tighter the interface between the human and the computer, the more information goes both ways,” he says.
Zhavoronkov sent data about several of Insilico’s compounds to Nanome, which re-created them in VR. Nanome’s chemist demonstrated chemical tweaks to potentially improve each compound. “It was a very good experience,” says Zhavoronkov.
Meanwhile, in March, Takeda Pharmaceutical Co., of Japan, invited Schrödinger, a New York–based company that develops chemical-simulation software, to join an alliance working on antivirals. Schrödinger’s AI focuses on the physics of how proteins interact with small molecules and one another.
The software sifts through billions of molecules per week to predict a compound’s properties, and it optimizes for multiple desired properties simultaneously, says Karen Akinsanya, chief biomedical scientist and head of discovery R&D at Schrödinger. “There’s a huge sense of urgency here to come up with a potent molecule, but also to come up with molecules that are going to be well tolerated” by the body, she says. Drug developers are seeking compounds that can be broadly used and easily administered, such as an oral drug rather than an intravenous drug, she adds.
Schrödinger evaluated four protein targets and performed virtual screens for two of them, a computing-intensive process. In June, Google Cloud donated the equivalent of 16 million hours of Nvidia GPU time for the company’s calculations. Next, the alliance’s drug companies will synthesize and test the most promising compounds identified by the virtual screens.
Other companies, including Amazon Web Services, IBM, and Intel, as well as several U.S. national labs are also donating time and resources to the Covid-19 High Performance Computing Consortium. The consortium is supporting 87 projects, which now have access to 6.8 million CPU cores, 50,000 GPUs, and 600 petaflops of computational resources.
While advanced technologies could transform early drug discovery, any new drug candidate still has a long road after that. It must be tested in animals, manufactured in large batches for clinical trials, then tested in a series of trials that, for antivirals, lasts an average of seven years.
In May, the BioMed X Institute in Germany launched a five-year project to build a Rapid Antiviral Response Platform, which would speed drug discovery all the way through manufacturing for clinical trials. The €40 million ($47 million) project, backed by drug companies, will identify outside-the-box proposals from young scientists, then provide space and funding to develop their ideas.
“We’ll focus on technologies that allow us to go from identification of a new virus to 10,000 doses of a novel potential therapeutic ready for trials in less than six months,” says BioMed X’s Tidona, who leads the project.
While a vaccine will likely arrive long before a bespoke antiviral does, experts expect COVID-19 to be with us for a long time, so the effort to develop a direct-acting, potent antiviral continues. Plus, having new antivirals—and tools to rapidly create more—can only help us prepare for the next pandemic, whether it comes next month or in another 102 years.
“We’ve got to start thinking differently about how to be more responsive to these kinds of threats,” says Collins. “It’s pushing us out of our comfort zones.”
This article appears in the October 2020 print issue as “Automating Antivirals.” Continue reading
Posted in Human Robots
#437603 Throwable Robot Car Always Lands on Four ...
Throwable or droppable robots seem like a great idea for a bunch of applications, including exploration and search and rescue. But such robots do come with some constraints—namely, if you’re going to throw or drop a robot, you should be prepared for that robot to not land the way you want it to land. While we’ve seen some creative approaches to this problem, or more straightforward self-righting devices, usually you’re in for significant trade-offs in complexity, mobility, and mass.
What would be ideal is a robot that can be relied upon to just always land the right way up. A robotic cat, of sorts. And while we’ve seen this with a tail, for wheeled vehicles, it turns out that a tail isn’t necessary: All it takes is some wheel spin.
The reason that AGRO (Agile Ground RObot), developed at the U.S. Military Academy at West Point, can do this is because each of its wheels is both independently driven and steerable. The wheels are essentially reaction wheels, which are a pretty common way to generate forces on all kinds of different robots, but typically you see such reaction wheels kludged onto these robots as sort of an afterthought—using the existing wheels of a wheeled robot is a more elegant way to do it.
Four steerable wheels with in-hub motors provide control in all three axes (yaw, pitch, and roll). You’ll notice that when the robot is tossed, the wheels all toe inwards (or outwards, I guess) by 45 degrees, positioning them orthogonal to the body of the robot. The front left and rear right wheels are spun together, as are the front right and rear left wheels. When one pair of wheels spins in the same direction, the body of the robot twists in the opposite way along an axis between those wheels, in a combination of pitch and roll. By combining different twisting torques from both pairs of wheels, pitch and roll along each axis can be adjusted independently. When the same pair of wheels spin in directions opposite to each other, the robot yaws, although yaw can also be derived by adjusting the ratio between pitch authority and roll authority. And lastly, if you want to sacrifice pitch control for more roll control (or vice versa) the wheel toe-in angle can be changed. Put all this together, and you get an enormous amount of mid-air control over your robot.
Image: Robotics Research Center/West Point
The AGRO robot features four steerable wheels with in-hub motors, which provide control in all three axes (yaw, pitch, and roll).
According to a paper that the West Point group will present at the 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), the overall objective here is for the robot to reach a state of zero pitch or roll by the time the robot impacts with the ground, to distribute the impact as much as possible. AGRO doesn’t yet have a suspension to make falling actually safe, so in the short term, it lands on a foam pad, but the mid-air adjustments it’s currently able to make result in a 20 percent reduction of impact force and a 100 percent reduction in being sideways or upside-down.
The toss that you see in the video isn’t the most aggressive, but lead author Daniel J. Gonzalez tells us that AGRO can do much better, theoretically stabilizing from an initial condition of 22.5 degrees pitch and 22.5 degrees roll in a mere 250 milliseconds, with room for improvement beyond that through optimizing the angles of individual wheels in real time. The limiting factor is really the amount of time that AGRO has between the point at which it’s released and the point at which it hits the ground, since more time in the air gives the robot more time to change its orientation.
Given enough height, the current generation of AGRO can recover from any initial orientation as long as it’s spinning at 66 rpm or less. And the only reason this is a limitation at all is because of the maximum rotation speed of the in-wheel hub motors, which can be boosted by increasing the battery voltage, as Gonzalez and his colleagues, Mark C. Lesak, Andres H. Rodriguez, Joseph A. Cymerman, and Christopher M. Korpela from the Robotics Research Center at West Point, describe in the IROS paper, “Dynamics and Aerial Attitude Control for Rapid Emergency Deployment of the Agile Ground Robot AGRO.”
Image: Robotics Research Center/West Point
AGRO 2 will include a new hybrid wheel-leg and non-pneumatic tire design that will allow it to hop up stairs and curbs.
While these particular experiments focus on a robot that’s being thrown, the concept is potentially effective (and useful) on any wheeled robot that’s likely to find itself in mid-air. You can imagine it improving the performance of robots doing all sorts of stunts, from driving off ramps or ledges to being dropped out of aircraft. And as it turns out, being able to self-stabilize during an airdrop is an important skill that some Humvees could use to keep themselves from getting tangled in their own parachute lines and avoid mishaps.
Before they move on to Humvees, though, the researchers are working on the next version of AGRO named AGRO 2. AGRO 2 will include a new hybrid wheel-leg and non-pneumatic tire design that will allow it to hop up stairs and curbs, which sounds like a lot of fun to us. Continue reading
Posted in Human Robots
#437477 If a Robot Is Conscious, Is It OK to ...
In the Star Trek: The Next Generation episode “The Measure of a Man,” Data, an android crew member of the Enterprise, is to be dismantled for research purposes unless Captain Picard can argue that Data deserves the same rights as a human being. Naturally the question arises: What is the basis upon which something has rights? What gives an entity moral standing?
The philosopher Peter Singer argues that creatures that can feel pain or suffer have a claim to moral standing. He argues that nonhuman animals have moral standing, since they can feel pain and suffer. Limiting it to people would be a form of speciesism, something akin to racism and sexism.
Without endorsing Singer’s line of reasoning, we might wonder if it can be extended further to an android robot like Data. It would require that Data can either feel pain or suffer. And how you answer that depends on how you understand consciousness and intelligence.
As real artificial intelligence technology advances toward Hollywood’s imagined versions, the question of moral standing grows more important. If AIs have moral standing, philosophers like me reason, it could follow that they have a right to life. That means you cannot simply dismantle them, and might also mean that people shouldn’t interfere with their pursuing their goals.
Two Flavors of Intelligence and a Test
IBM’s Deep Blue chess machine was successfully trained to beat grandmaster Gary Kasparov. But it could not do anything else. This computer had what’s called domain-specific intelligence.
On the other hand, there’s the kind of intelligence that allows for the ability to do a variety of things well. It is called domain-general intelligence. It’s what lets people cook, ski, and raise children—tasks that are related, but also very different.
Artificial general intelligence, AGI, is the term for machines that have domain-general intelligence. Arguably no machine has yet demonstrated that kind of intelligence. This summer, a startup called OpenAI released a new version of its Generative Pre-Training language model. GPT-3 is a natural language processing system, trained to read and write so that it can be easily understood by people.
It drew immediate notice, not just because of its impressive ability to mimic stylistic flourishes and put together plausible content, but also because of how far it had come from a previous version. Despite this impressive performance, GPT-3 doesn’t actually know anything beyond how to string words together in various ways. AGI remains quite far off.
Named after pioneering AI researcher Alan Turing, the Turing test helps determine when an AI is intelligent. Can a person conversing with a hidden AI tell whether it’s an AI or a human being? If he can’t, then for all practical purposes, the AI is intelligent. But this test says nothing about whether the AI might be conscious.
Two Kinds of Consciousness
There are two parts to consciousness. First, there’s the what-it’s-like-for-me aspect of an experience, the sensory part of consciousness. Philosophers call this phenomenal consciousness. It’s about how you experience a phenomenon, like smelling a rose or feeling pain.
In contrast, there’s also access consciousness. That’s the ability to report, reason, behave, and act in a coordinated and responsive manner to stimuli based on goals. For example, when I pass the soccer ball to my friend making a play on the goal, I am responding to visual stimuli, acting from prior training, and pursuing a goal determined by the rules of the game. I make the pass automatically, without conscious deliberation, in the flow of the game.
Blindsight nicely illustrates the difference between the two types of consciousness. Someone with this neurological condition might report, for example, that they cannot see anything in the left side of their visual field. But if asked to pick up a pen from an array of objects in the left side of their visual field, they can reliably do so. They cannot see the pen, yet they can pick it up when prompted—an example of access consciousness without phenomenal consciousness.
Data is an android. How do these distinctions play out with respect to him?
The Data Dilemma
The android Data demonstrates that he is self-aware in that he can monitor whether or not, for example, he is optimally charged or there is internal damage to his robotic arm.
Data is also intelligent in the general sense. He does a lot of distinct things at a high level of mastery. He can fly the Enterprise, take orders from Captain Picard and reason with him about the best path to take.
He can also play poker with his shipmates, cook, discuss topical issues with close friends, fight with enemies on alien planets, and engage in various forms of physical labor. Data has access consciousness. He would clearly pass the Turing test.
However, Data most likely lacks phenomenal consciousness—he does not, for example, delight in the scent of roses or experience pain. He embodies a supersized version of blindsight. He’s self-aware and has access consciousness—can grab the pen—but across all his senses he lacks phenomenal consciousness.
Now, if Data doesn’t feel pain, at least one of the reasons Singer offers for giving a creature moral standing is not fulfilled. But Data might fulfill the other condition of being able to suffer, even without feeling pain. Suffering might not require phenomenal consciousness the way pain essentially does.
For example, what if suffering were also defined as the idea of being thwarted from pursuing a just cause without causing harm to others? Suppose Data’s goal is to save his crewmate, but he can’t reach her because of damage to one of his limbs. Data’s reduction in functioning that keeps him from saving his crewmate is a kind of nonphenomenal suffering. He would have preferred to save the crewmate, and would be better off if he did.
In the episode, the question ends up resting not on whether Data is self-aware—that is not in doubt. Nor is it in question whether he is intelligent—he easily demonstrates that he is in the general sense. What is unclear is whether he is phenomenally conscious. Data is not dismantled because, in the end, his human judges cannot agree on the significance of consciousness for moral standing.
Should an AI Get Moral Standing?
Data is kind; he acts to support the well-being of his crewmates and those he encounters on alien planets. He obeys orders from people and appears unlikely to harm them, and he seems to protect his own existence. For these reasons he appears peaceful and easier to accept into the realm of things that have moral standing.
But what about Skynet in the Terminator movies? Or the worries recently expressed by Elon Musk about AI being more dangerous than nukes, and by Stephen Hawking on AI ending humankind?
Human beings don’t lose their claim to moral standing just because they act against the interests of another person. In the same way, you can’t automatically say that just because an AI acts against the interests of humanity or another AI it doesn’t have moral standing. You might be justified in fighting back against an AI like Skynet, but that does not take away its moral standing. If moral standing is given in virtue of the capacity to nonphenomenally suffer, then Skynet and Data both get it even if only Data wants to help human beings.
There are no artificial general intelligence machines yet. But now is the time to consider what it would take to grant them moral standing. How humanity chooses to answer the question of moral standing for nonbiological creatures will have big implications for how we deal with future AIs—whether kind and helpful like Data, or set on destruction, like Skynet.
This article is republished from The Conversation under a Creative Commons license. Read the original article.
Image Credit: Ico Maker / Shutterstock.com Continue reading
Posted in Human Robots
#436258 For Centuries, People Dreamed of a ...
This is part six of a six-part series on the history of natural language processing.
In February of this year, OpenAI, one of the foremost artificial intelligence labs in the world, announced that a team of researchers had built a powerful new text generator called the Generative Pre-Trained Transformer 2, or GPT-2 for short. The researchers used a reinforcement learning algorithm to train their system on a broad set of natural language processing (NLP) capabilities, including reading comprehension, machine translation, and the ability to generate long strings of coherent text.
But as is often the case with NLP technology, the tool held both great promise and great peril. Researchers and policy makers at the lab were concerned that their system, if widely released, could be exploited by bad actors and misappropriated for “malicious purposes.”
The people of OpenAI, which defines its mission as “discovering and enacting the path to safe artificial general intelligence,” were concerned that GPT-2 could be used to flood the Internet with fake text, thereby degrading an already fragile information ecosystem. For this reason, OpenAI decided that it would not release the full version of GPT-2 to the public or other researchers.
GPT-2 is an example of a technique in NLP called language modeling, whereby the computational system internalizes a statistical blueprint of a text so it’s able to mimic it. Just like the predictive text on your phone—which selects words based on words you’ve used before—GPT-2 can look at a string of text and then predict what the next word is likely to be based on the probabilities inherent in that text.
GPT-2 can be seen as a descendant of the statistical language modeling that the Russian mathematician A. A. Markov developed in the early 20th century (covered in part three of this series).
GPT-2 used cutting-edge machine learning algorithms to do linguistic analysis with over 1.5 million parameters.
What’s different with GPT-2, though, is the scale of the textual data modeled by the system. Whereas Markov analyzed a string of 20,000 letters to create a rudimentary model that could predict the likelihood of the next letter of a text being a consonant or a vowel, GPT-2 used 8 million articles scraped from Reddit to predict what the next word might be within that entire dataset.
And whereas Markov manually trained his model by counting only two parameters—vowels and consonants—GPT-2 used cutting-edge machine learning algorithms to do linguistic analysis with over 1.5 million parameters, burning through huge amounts of computational power in the process.
The results were impressive. In their blog post, OpenAI reported that GPT-2 could generate synthetic text in response to prompts, mimicking whatever style of text it was shown. If you prompt the system with a line of William Blake’s poetry, it can generate a line back in the Romantic poet’s style. If you prompt the system with a cake recipe, you get a newly invented recipe in response.
Perhaps the most compelling feature of GPT-2 is that it can answer questions accurately. For example, when OpenAI researchers asked the system, “Who wrote the book The Origin of Species?”—it responded: “Charles Darwin.” While only able to respond accurately some of the time, the feature does seem to be a limited realization of Gottfried Leibniz’s dream of a language-generating machine that could answer any and all human questions (described in part two of this series).
After observing the power of the new system in practice, OpenAI elected not to release the fully trained model. In the lead up to its release in February, there had been heightened awareness about “deepfakes”—synthetic images and videos, generated via machine learning techniques, in which people do and say things they haven’t really done and said. Researchers at OpenAI worried that GPT-2 could be used to essentially create deepfake text, making it harder for people to trust textual information online.
Responses to this decision varied. On one hand, OpenAI’s caution prompted an overblown reaction in the media, with articles about the “dangerous” technology feeding into the Frankenstein narrative that often surrounds developments in AI.
Others took issue with OpenAI’s self-promotion, with some even suggesting that OpenAI purposefully exaggerated GPT-2s power in order to create hype—while contravening a norm in the AI research community, where labs routinely share data, code, and pre-trained models. As machine learning researcher Zachary Lipton tweeted, “Perhaps what's *most remarkable* about the @OpenAI controversy is how *unremarkable* the technology is. Despite their outsize attention & budget, the research itself is perfectly ordinary—right in the main branch of deep learning NLP research.”
OpenAI stood by its decision to release only a limited version of GPT-2, but has since released larger models for other researchers and the public to experiment with. As yet, there has been no reported case of a widely distributed fake news article generated by the system. But there have been a number of interesting spin-off projects, including GPT-2 poetry and a webpage where you can prompt the system with questions yourself.
Mimicking humans on Reddit, the bots have long conversations about a variety of topics, including conspiracy theories and
Star Wars movies.
There’s even a Reddit group populated entirely with text produced by GPT-2-powered bots. Mimicking humans on Reddit, the bots have long conversations about a variety of topics, including conspiracy theories and Star Wars movies.
This bot-powered conversation may signify the new condition of life online, where language is increasingly created by a combination of human and non-human agents, and where maintaining the distinction between human and non-human, despite our best efforts, is increasingly difficult.
The idea of using rules, mechanisms, and algorithms to generate language has inspired people in many different cultures throughout history. But it’s in the online world that this powerful form of wordcraft may really find its natural milieu—in an environment where the identity of speakers becomes more ambiguous, and perhaps, less relevant. It remains to be seen what the consequences will be for language, communication, and our sense of human identity, which is so bound up with our ability to speak in natural language.
This is the sixth installment of a six-part series on the history of natural language processing. Last week’s post explained how an innocent Microsoft chatbot turned instantly racist on Twitter.
You can also check out our prior series on the untold history of AI. Continue reading
Posted in Human Robots