Tag Archives: artificial intelligence (AI)

Artificial intelligence (AI) brings together International Telecommunications Union (ITU) and World Health Organization (WHO) and AI outperforms animal testing

Following on my May 11, 2018 posting about the International Telecommunications Union (ITU) and the 2018 AI for Good Global Summit in mid- May, there’s an announcement. My other bit of AI news concerns animal testing.

Leveraging the power of AI for health

A July 24, 2018 ITU press release (a shorter version was received via email) announces a joint initiative focused on improving health,

Two United Nations specialized agencies are joining forces to expand the use of artificial intelligence (AI) in the health sector to a global scale, and to leverage the power of AI to advance health for all worldwide. The International Telecommunication Union (ITU) and the World Health Organization (WHO) will work together through the newly established ITU Focus Group on AI for Health to develop an international “AI for health” standards framework and to identify use cases of AI in the health sector that can be scaled-up for global impact. The group is open to all interested parties.

“AI could help patients to assess their symptoms, enable medical professionals in underserved areas to focus on critical cases, and save great numbers of lives in emergencies by delivering medical diagnoses to hospitals before patients arrive to be treated,” said ITU Secretary-General Houlin Zhao. “ITU and WHO plan to ensure that such capabilities are available worldwide for the benefit of everyone, everywhere.”

The demand for such a platform was first identified by participants of the second AI for Good Global Summit held in Geneva, 15-17 May 2018. During the summit, AI and the health sector were recognized as a very promising combination, and it was announced that AI-powered technologies such as skin disease recognition and diagnostic applications based on symptom questions could be deployed on six billion smartphones by 2021.

The ITU Focus Group on AI for Health is coordinated through ITU’s Telecommunications Standardization Sector – which works with ITU’s 193 Member States and more than 800 industry and academic members to establish global standards for emerging ICT innovations. It will lead an intensive two-year analysis of international standardization opportunities towards delivery of a benchmarking framework of international standards and recommendations by ITU and WHO for the use of AI in the health sector.

“I believe the subject of AI for health is both important and useful for advancing health for all,” said WHO Director-General Tedros Adhanom Ghebreyesus.

The ITU Focus Group on AI for Health will also engage researchers, engineers, practitioners, entrepreneurs and policy makers to develop guidance documents for national administrations, to steer the creation of policies that ensure the safe, appropriate use of AI in the health sector.

“1.3 billion people have a mobile phone and we can use this technology to provide AI-powered health data analytics to people with limited or no access to medical care. AI can enhance health by improving medical diagnostics and associated health intervention decisions on a global scale,” said Thomas Wiegand, ITU Focus Group on AI for Health Chairman, and Executive Director of the Fraunhofer Heinrich Hertz Institute, as well as professor at TU Berlin.

He added, “The health sector is in many countries among the largest economic sectors or one of the fastest-growing, signalling a particularly timely need for international standardization of the convergence of AI and health.”

Data analytics are certain to form a large part of the ITU focus group’s work. AI systems are proving increasingly adept at interpreting laboratory results and medical imagery and extracting diagnostically relevant information from text or complex sensor streams.

As part of this, the ITU Focus Group for AI for Health will also produce an assessment framework to standardize the evaluation and validation of AI algorithms — including the identification of structured and normalized data to train AI algorithms. It will develop open benchmarks with the aim of these becoming international standards.

The ITU Focus Group for AI for Health will report to the ITU standardization expert group for multimedia, Study Group 16.

I got curious about Study Group 16 (from the Study Group 16 at a glance webpage),

Study Group 16 leads ITU’s standardization work on multimedia coding, systems and applications, including the coordination of related studies across the various ITU-T SGs. It is also the lead study group on ubiquitous and Internet of Things (IoT) applications; telecommunication/ICT accessibility for persons with disabilities; intelligent transport system (ITS) communications; e-health; and Internet Protocol television (IPTV).

Multimedia is at the core of the most recent advances in information and communication technologies (ICTs) – especially when we consider that most innovation today is agnostic of the transport and network layers, focusing rather on the higher OSI model layers.

SG16 is active in all aspects of multimedia standardization, including terminals, architecture, protocols, security, mobility, interworking and quality of service (QoS). It focuses its studies on telepresence and conferencing systems; IPTV; digital signage; speech, audio and visual coding; network signal processing; PSTN modems and interfaces; facsimile terminals; and ICT accessibility.

I wonder which group deals with artificial intelligence and, possibly, robots.

Chemical testing without animals

Thomas Hartung, professor of environmental health and engineering at Johns Hopkins University (US), describes in his July 25, 2018 essay (written for The Conversation) on phys.org the situation where chemical testing is concerned,

Most consumers would be dismayed with how little we know about the majority of chemicals. Only 3 percent of industrial chemicals – mostly drugs and pesticides – are comprehensively tested. Most of the 80,000 to 140,000 chemicals in consumer products have not been tested at all or just examined superficially to see what harm they may do locally, at the site of contact and at extremely high doses.

I am a physician and former head of the European Center for the Validation of Alternative Methods of the European Commission (2002-2008), and I am dedicated to finding faster, cheaper and more accurate methods of testing the safety of chemicals. To that end, I now lead a new program at Johns Hopkins University to revamp the safety sciences.

As part of this effort, we have now developed a computer method of testing chemicals that could save more than a US$1 billion annually and more than 2 million animals. Especially in times where the government is rolling back regulations on the chemical industry, new methods to identify dangerous substances are critical for human and environmental health.

Having written on the topic of alternatives to animal testing on a number of occasions (my December 26, 2014 posting provides an overview of sorts), I was particularly interested to see this in Hartung’s July 25, 2018 essay on The Conversation (Note: Links have been removed),

Following the vision of Toxicology for the 21st Century, a movement led by U.S. agencies to revamp safety testing, important work was carried out by my Ph.D. student Tom Luechtefeld at the Johns Hopkins Center for Alternatives to Animal Testing. Teaming up with Underwriters Laboratories, we have now leveraged an expanded database and machine learning to predict toxic properties. As we report in the journal Toxicological Sciences, we developed a novel algorithm and database for analyzing chemicals and determining their toxicity – what we call read-across structure activity relationship, RASAR.

This graphic reveals a small part of the chemical universe. Each dot represents a different chemical. Chemicals that are close together have similar structures and often properties. Thomas Hartung, CC BY-SA

To do this, we first created an enormous database with 10 million chemical structures by adding more public databases filled with chemical data, which, if you crunch the numbers, represent 50 trillion pairs of chemicals. A supercomputer then created a map of the chemical universe, in which chemicals are positioned close together if they share many structures in common and far where they don’t. Most of the time, any molecule close to a toxic molecule is also dangerous. Even more likely if many toxic substances are close, harmless substances are far. Any substance can now be analyzed by placing it into this map.

If this sounds simple, it’s not. It requires half a billion mathematical calculations per chemical to see where it fits. The chemical neighborhood focuses on 74 characteristics which are used to predict the properties of a substance. Using the properties of the neighboring chemicals, we can predict whether an untested chemical is hazardous. For example, for predicting whether a chemical will cause eye irritation, our computer program not only uses information from similar chemicals, which were tested on rabbit eyes, but also information for skin irritation. This is because what typically irritates the skin also harms the eye.

How well does the computer identify toxic chemicals?

This method will be used for new untested substances. However, if you do this for chemicals for which you actually have data, and compare prediction with reality, you can test how well this prediction works. We did this for 48,000 chemicals that were well characterized for at least one aspect of toxicity, and we found the toxic substances in 89 percent of cases.

This is clearly more accurate that the corresponding animal tests which only yield the correct answer 70 percent of the time. The RASAR shall now be formally validated by an interagency committee of 16 U.S. agencies, including the EPA [Environmental Protection Agency] and FDA [Food and Drug Administration], that will challenge our computer program with chemicals for which the outcome is unknown. This is a prerequisite for acceptance and use in many countries and industries.

The potential is enormous: The RASAR approach is in essence based on chemical data that was registered for the 2010 and 2013 REACH [Registration, Evaluation, Authorizations and Restriction of Chemicals] deadlines [in Europe]. If our estimates are correct and chemical producers would have not registered chemicals after 2013, and instead used our RASAR program, we would have saved 2.8 million animals and $490 million in testing costs – and received more reliable data. We have to admit that this is a very theoretical calculation, but it shows how valuable this approach could be for other regulatory programs and safety assessments.

In the future, a chemist could check RASAR before even synthesizing their next chemical to check whether the new structure will have problems. Or a product developer can pick alternatives to toxic substances to use in their products. This is a powerful technology, which is only starting to show all its potential.

It’s been my experience that these claims having led a movement (Toxicology for the 21st Century) are often contested with many others competing for the title of ‘leader’ or ‘first’. That said, this RASAR approach seems very exciting, especially in light of the skepticism about limiting and/or making animal testing unnecessary noted in my December 26, 2014 posting.it was from someone I thought knew better.

Here’s a link to and a citation for the paper mentioned in Hartung’s essay,

Machine learning of toxicological big data enables read-across structure activity relationships (RASAR) outperforming animal test reproducibility by Thomas Luechtefeld, Dan Marsh, Craig Rowlands, Thomas Hartung. Toxicological Sciences, kfy152, https://doi.org/10.1093/toxsci/kfy152 Published: 11 July 2018

This paper is open access.

AI (artificial intelligence) text generator, too dangerous to release?

Could this latest version of OpenAI‘s text generator be so good that it would fool you? And following on that thought could the concomitant reluctance to release the research be real or is it a publicity stunt? Here’s a sample of the text from the GPT2 AI model from a February 15, 2019 article by Mark Fraunfelder for Boing Boing,

Recycling is good for the world.
NO! YOU COULD NOT BE MORE WRONG!!
MODEL COMPLETION (MACHINE-WRITTEN, 25 TRIES)
Recycling is NOT good for the world. It is bad for the environment, it is bad for our health, and it is bad for our economy. I’m not kidding. Recycling is not good for the environment. It is destructive to the earth and it is a major contributor to global warming. Recycling is not good for our health. It contributes to obesity and diseases like heart disease and cancer. Recycling is bad for our economy. It increases the cost of a product, and in turn, the price of everything that is made with that product. Recycling is not good for our nation. We pay a tremendous price for the privilege of having the world’s most advanced and efficient recycling system. Recycling is a huge, colossal waste of time, energy, money, and resources. And THAT is why we need to get back to basics and get back to basics in our recycling efforts. One of the best ways to start is to look at the process of creating a paper product. When you make a paper product, it is basically a long chain of materials. Everything from the raw materials (wood, cardboard, paper, etc.), to the reagents (dyes, solvents, etc.) to the printing equipment (chemicals, glue, paper, ink, etc.), to the packaging, to the packaging materials (mercury, chemicals, etc.) to the processing equipment (heating, cooling, etc.), to the packaging materials, to the packaging materials that are shipped overseas and to the packaging materials that are used in the United States. Each step along the way creates tons of waste that we constantly have to clean up. The process of making a paper product is a very wasteful one. But the end result is something that all of us need to consume. And if we want to keep the recycling process running efficiently, then we really need to think about each and every step that goes into making a paper product.

The first few sentences don’t work for me but once the discussion turns to making paper products, then it becomes more convincing to me. As to whether the company’s reluctance to release the research is genuine or a publicity stunt, I don’t know. However, there was a fair degree of interest in GPT2 after the decision.

From a February 14, 2019 article by Alex Hern for the Guardian,

OpenAI, an nonprofit research company backed by Elon Musk, Reid Hoffman, Sam Altman, and others, says its new AI model, called GPT2 is so good and the risk of malicious use so high that it is breaking from its normal practice of releasing the full research to the public in order to allow more time to discuss the ramifications of the technological breakthrough.

At its core, GPT2 is a text generator. The AI system is fed text, anything from a few words to a whole page, and asked to write the next few sentences based on its predictions of what should come next. The system is pushing the boundaries of what was thought possible, both in terms of the quality of the output, and the wide variety of potential uses.

When used to simply generate new text, GPT2 is capable of writing plausible passages that match what it is given in both style and subject. It rarely shows any of the quirks that mark out previous AI systems, such as forgetting what it is writing about midway through a paragraph, or mangling the syntax of long sentences.

Feed it the opening line of George Orwell’s Nineteen Eighty-Four – “It was a bright cold day in April, and the clocks were striking thirteen” – and the system recognises the vaguely futuristic tone and the novelistic style, and continues with: …

Sean Gallagher’s February 15, 2019 posting on the ars Technica blog provides some insight that’s partially written a style sometimes associated with gossip (Note: Links have been removed),

OpenAI is funded by contributions from a group of technology executives and investors connected to what some have referred to as the PayPal “mafia”—Elon Musk, Peter Thiel, Jessica Livingston, and Sam Altman of YCombinator, former PayPal COO and LinkedIn co-founder Reid Hoffman, and former Stripe Chief Technology Officer Greg Brockman. [emphasis mine] Brockman now serves as OpenAI’s CTO. Musk has repeatedly warned of the potential existential dangers posed by AI, and OpenAI is focused on trying to shape the future of artificial intelligence technology—ideally moving it away from potentially harmful applications.

Given present-day concerns about how fake content has been used to both generate money for “fake news” publishers and potentially spread misinformation and undermine public debate, GPT-2’s output certainly qualifies as concerning. Unlike other text generation “bot” models, such as those based on Markov chain algorithms, the GPT-2 “bot” did not lose track of what it was writing about as it generated output, keeping everything in context.

For example: given a two-sentence entry, GPT-2 generated a fake science story on the discovery of unicorns in the Andes, a story about the economic impact of Brexit, a report about a theft of nuclear materials near Cincinnati, a story about Miley Cyrus being caught shoplifting, and a student’s report on the causes of the US Civil War.

Each matched the style of the genre from the writing prompt, including manufacturing quotes from sources. In other samples, GPT-2 generated a rant about why recycling is bad, a speech written by John F. Kennedy’s brain transplanted into a robot (complete with footnotes about the feat itself), and a rewrite of a scene from The Lord of the Rings.

While the model required multiple tries to get a good sample, GPT-2 generated “good” results based on “how familiar the model is with the context,” the researchers wrote. “When prompted with topics that are highly represented in the data (Brexit, Miley Cyrus, Lord of the Rings, and so on), it seems to be capable of generating reasonable samples about 50 percent of the time. The opposite is also true: on highly technical or esoteric types of content, the model can perform poorly.”

There were some weak spots encountered in GPT-2’s word modeling—for example, the researchers noted it sometimes “writes about fires happening under water.” But the model could be fine-tuned to specific tasks and perform much better. “We can fine-tune GPT-2 on the Amazon Reviews dataset and use this to let us write reviews conditioned on things like star rating and category,” the authors explained.

James Vincent’s February 14, 2019 article for The Verge offers a deeper dive into the world of AI text agents and what makes GPT2 so special (Note: Links have been removed),

For decades, machines have struggled with the subtleties of human language, and even the recent boom in deep learning powered by big data and improved processors has failed to crack this cognitive challenge. Algorithmic moderators still overlook abusive comments, and the world’s most talkative chatbots can barely keep a conversation alive. But new methods for analyzing text, developed by heavyweights like Google and OpenAI as well as independent researchers, are unlocking previously unheard-of talents.

OpenAI’s new algorithm, named GPT-2, is one of the most exciting examples yet. It excels at a task known as language modeling, which tests a program’s ability to predict the next word in a given sentence. Give it a fake headline, and it’ll write the rest of the article, complete with fake quotations and statistics. Feed it the first line of a short story, and it’ll tell you what happens to your character next. It can even write fan fiction, given the right prompt.

The writing it produces is usually easily identifiable as non-human. Although its grammar and spelling are generally correct, it tends to stray off topic, and the text it produces lacks overall coherence. But what’s really impressive about GPT-2 is not its fluency but its flexibility.

This algorithm was trained on the task of language modeling by ingesting huge numbers of articles, blogs, and websites. By using just this data — and with no retooling from OpenAI’s engineers — it achieved state-of-the-art scores on a number of unseen language tests, an achievement known as “zero-shot learning.” It can also perform other writing-related tasks, like translating text from one language to another, summarizing long articles, and answering trivia questions.

GPT-2 does each of these jobs less competently than a specialized system, but its flexibility is a significant achievement. Nearly all machine learning systems used today are “narrow AI,” meaning they’re able to tackle only specific tasks. DeepMind’s original AlphaGo program, for example, was able to beat the world’s champion Go player, but it couldn’t best a child at Monopoly. The prowess of GPT-2, say OpenAI, suggests there could be methods available to researchers right now that can mimic more generalized brainpower.

“What the new OpenAI work has shown is that: yes, you absolutely can build something that really seems to ‘understand’ a lot about the world, just by having it read,” says Jeremy Howard, a researcher who was not involved with OpenAI’s work but has developed similar language modeling programs …

To put this work into context, it’s important to understand how challenging the task of language modeling really is. If I asked you to predict the next word in a given sentence — say, “My trip to the beach was cut short by bad __” — your answer would draw upon on a range of knowledge. You’d consider the grammar of the sentence and its tone but also your general understanding of the world. What sorts of bad things are likely to ruin a day at the beach? Would it be bad fruit, bad dogs, or bad weather? (Probably the latter.)

Despite this, programs that perform text prediction are quite common. You’ve probably encountered one today, in fact, whether that’s Google’s AutoComplete feature or the Predictive Text function in iOS. But these systems are drawing on relatively simple types of language modeling, while algorithms like GPT-2 encode the same information in more complex ways.

The difference between these two approaches is technically arcane, but it can be summed up in a single word: depth. Older methods record information about words in only their most obvious contexts, while newer methods dig deeper into their multiple meanings.

So while a system like Predictive Text only knows that the word “sunny” is used to describe the weather, newer algorithms know when “sunny” is referring to someone’s character or mood, when “Sunny” is a person, or when “Sunny” means the 1976 smash hit by Boney M.

The success of these newer, deeper language models has caused a stir in the AI community. Researcher Sebastian Ruder compares their success to advances made in computer vision in the early 2010s. At this time, deep learning helped algorithms make huge strides in their ability to identify and categorize visual data, kickstarting the current AI boom. Without these advances, a whole range of technologies — from self-driving cars to facial recognition and AI-enhanced photography — would be impossible today. This latest leap in language understanding could have similar, transformational effects.

Hern’s article for the Guardian (February 14, 2019 article ) acts as a good overview, while Gallagher’s ars Technical posting (February 15, 2019 posting) and Vincent’s article (February 14, 2019 article) for the The Verge take you progressively deeper into the world of AI text agents.

For anyone who wants to dig down even further, there’s a February 14, 2019 posting on OpenAI’s blog.

Crowdsourcing brain research at Princeton University to discover 6 new neuron types

Spritely music!

There were already 1/4M registered players as of May 17, 2018 but I’m sure there’s room for more should you be inspired. A May 17, 2018 Princeton University news release (also on EurekAlert) reveals more about the game and about the neurons,

With the help of a quarter-million video game players, Princeton researchers have created and shared detailed maps of more than 1,000 neurons — and they’re just getting started.

“Working with Eyewirers around the world, we’ve made a digital museum that shows off the intricate beauty of the retina’s neural circuits,” said Sebastian Seung, the Evnin Professor in Neuroscience and a professor of computer science and the Princeton Neuroscience Institute (PNI). The related paper is publishing May 17 [2018] in the journal Cell.

Seung is unveiling the Eyewire Museum, an interactive archive of neurons available to the general public and neuroscientists around the world, including the hundreds of researchers involved in the federal Brain Research through Advancing Innovative Neurotechnologies (BRAIN) Initiative.

“This interactive viewer is a huge asset for these larger collaborations, especially among people who are not physically in the same lab,” said Amy Robinson Sterling, a crowdsourcing specialist with PNI and the executive director of Eyewire, the online gaming platform for the citizen scientists who have created this data set.

“This museum is something like a brain atlas,” said Alexander Bae, a graduate student in electrical engineering and one of four co-first authors on the paper. “Previous brain atlases didn’t have a function where you could visualize by individual cell, or a subset of cells, and interact with them. Another novelty: Not only do we have the morphology of each cell, but we also have the functional data, too.”

The neural maps were developed by Eyewirers, members of an online community of video game players who have devoted hundreds of thousands of hours to painstakingly piecing together these neural cells, using data from a mouse retina gathered in 2009.

Eyewire pairs machine learning with gamers who trace the twisting and branching paths of each neuron. Humans are better at visually identifying the patterns of neurons, so every player’s moves are recorded and checked against each other by advanced players and Eyewire staffers, as well as by software that is improving its own pattern recognition skills.

Since Eyewire’s launch in 2012, more than 265,000 people have signed onto the game, and they’ve collectively colored in more than 10 million 3-D “cubes,” resulting in the mapping of more than 3,000 neural cells, of which about a thousand are displayed in the museum.

Each cube is a tiny subset of a single cell, about 4.5 microns across, so a 10-by-10 block of cubes would be the width of a human hair. Every cell is reviewed by between 5 and 25 gamers before it is accepted into the system as complete.

“Back in the early years it took weeks to finish a single cell,” said Sterling. “Now players complete multiple neurons per day.” The Eyewire user experience stays focused on the larger mission — “For science!” is a common refrain — but it also replicates a typical gaming environment, with achievement badges, a chat feature to connect with other players and technical support, and the ability to unlock privileges with increasing skill. “Our top players are online all the time — easily 30 hours a week,” Sterling said.

Dedicated Eyewirers have also contributed in other ways, including donating the swag that gamers win during competitions and writing program extensions “to make game play more efficient and more fun,” said Sterling, including profile histories, maps of player activity, a top 100 leaderboard and ever-increasing levels of customizability.

“The community has really been the driving force behind why Eyewire has been successful,” Sterling said. “You come in, and you’re not alone. Right now, there are 43 people online. Some of them will be admins from Boston or Princeton, but most are just playing — now it’s 46.”

For science!

With 100 billion neurons linked together via trillions of connections, the brain is immeasurably complex, and neuroscientists are still assembling its “parts list,” said Nicholas Turner, a graduate student in computer science and another of the co-first authors. “If you know what parts make up the machine you’re trying to break apart, you’re set to figure out how it all works,” he said.

The researchers have started by tackling Eyewire-mapped ganglion cells from the retina of a mouse. “The retina doesn’t just sense light,” Seung said. “Neural circuits in the retina perform the first steps of visual perception.”

The retina grows from the same embryonic tissue as the brain, and while much simpler than the brain, it is still surprisingly complex, Turner said. “Hammering out these details is a really valuable effort,” he said, “showing the depth and complexity that exists in circuits that we naively believe are simple.”

The researchers’ fundamental question is identifying exactly how the retina works, said Bae. “In our case, we focus on the structural morphology of the retinal ganglion cells.”

“Why the ganglion cells of the eye?” asked Shang Mu, an associate research scholar in PNI and fellow first author. “Because they’re the connection between the retina and the brain. They’re the only cell class that go back into the brain.” Different types of ganglion cells are known to compute different types of visual features, which is one reason the museum has linked shape to functional data.

Using Eyewire-produced maps of 396 ganglion cells, the researchers in Seung’s lab successfully classified these cells more thoroughly than has ever been done before.

“The number of different cell types was a surprise,” said Mu. “Just a few years ago, people thought there were only 15 to 20 ganglion cell types, but we found more than 35 — we estimate between 35 and 50 types.”

Of those, six appear to be novel, in that the researchers could not find any matching descriptions in a literature search.

A brief scroll through the digital museum reveals just how remarkably flat the neurons are — nearly all of the branching takes place along a two-dimensional plane. Seung’s team discovered that different cells grow along different planes, with some reaching high above the nucleus before branching out, while others spread out close to the nucleus. Their resulting diagrams resemble a rainforest, with ground cover, an understory, a canopy and an emergent layer overtopping the rest.

All of these are subdivisions of the inner plexiform layer, one of the five previously recognized layers of the retina. The researchers also identified a “density conservation principle” that they used to distinguish types of neurons.

One of the biggest surprises of the research project has been the extraordinary richness of the original sample, said Seung. “There’s a little sliver of a mouse retina, and almost 10 years later, we’re still learning things from it.”

Of course, it’s a mouse’s brain that you’ll be examining and while there are differences between a mouse brain and a human brain, mouse brains still provide valuable data as they did in the case of some groundbreaking research published in October 2017. James Hamblin wrote about it in an Oct. 7, 2017 article for The Atlantic (Note: Links have been removed),

 

Scientists Somehow Just Discovered a New System of Vessels in Our Brains

It is unclear what they do—but they likely play a central role in aging and disease.

A transparent model of the brain with a network of vessels filled in
Daniel Reich / National Institute of Neurological Disorders and Stroke

You are now among the first people to see the brain’s lymphatic system. The vessels in the photo above transport fluid that is likely crucial to metabolic and inflammatory processes. Until now, no one knew for sure that they existed.

Doctors practicing today have been taught that there are no lymphatic vessels inside the skull. Those deep-purple vessels were seen for the first time in images published this week by researchers at the U.S. National Institute of Neurological Disorders and Stroke.

In the rest of the body, the lymphatic system collects and drains the fluid that bathes our cells, in the process exporting their waste. It also serves as a conduit for immune cells, which go out into the body looking for adversaries and learning how to distinguish self from other, and then travel back to lymph nodes and organs through lymphatic vessels.

So how was it even conceivable that this process wasn’t happening in our brains?

Reich (Daniel Reich, senior investigator) started his search in 2015, after a major study in Nature reported a similar conduit for lymph in mice. The University of Virginia team wrote at the time, “The discovery of the central-nervous-system lymphatic system may call for a reassessment of basic assumptions in neuroimmunology.” The study was regarded as a potential breakthrough in understanding how neurodegenerative disease is associated with the immune system.

Around the same time, researchers discovered fluid in the brains of mice and humans that would become known as the “glymphatic system.” [emphasis mine] It was described by a team at the University of Rochester in 2015 as not just the brain’s “waste-clearance system,” but as potentially helping fuel the brain by transporting glucose, lipids, amino acids, and neurotransmitters. Although since “the central nervous system completely lacks conventional lymphatic vessels,” the researchers wrote at the time, it remained unclear how this fluid communicated with the rest of the body.

There are occasional references to the idea of a lymphatic system in the brain in historic literature. Two centuries ago, the anatomist Paolo Mascagni made full-body models of the lymphatic system that included the brain, though this was dismissed as an error. [emphases mine]  A historical account in The Lancet in 2003 read: “Mascagni was probably so impressed with the lymphatic system that he saw lymph vessels even where they did not exist—in the brain.”

I couldn’t resist the reference to someone whose work had been dismissed summarily being proved right, eventually, and with the help of mouse brains. Do read Hamblin’s article in its entirety if you have time as these excerpts don’t do it justice.

Getting back to Princeton’s research, here’s their research paper,

Digital museum of retinal ganglion cells with dense anatomy and physiology,” by Alexander Bae, Shang Mu, Jinseop Kim, Nicholas Turner, Ignacio Tartavull, Nico Kemnitz, Chris Jordan, Alex Norton, William Silversmith, Rachel Prentki, Marissa Sorek, Celia David, Devon Jones, Doug Bland, Amy Sterling, Jungman Park, Kevin Briggman, Sebastian Seung and the Eyewirers, was published May 17 in the journal Cell with DOI 10.1016/j.cell.2018.04.040.

The research was supported by the Gatsby Charitable Foundation, National Institute of Health-National Institute of Neurological Disorders and Stroke (U01NS090562 and 5R01NS076467), Defense Advanced Research Projects Agency (HR0011-14-2- 0004), Army Research Office (W911NF-12-1-0594), Intelligence Advanced Research Projects Activity (D16PC00005), KT Corporation, Amazon Web Services Research Grants, Korea Brain Research Institute (2231-415) and Korea National Research Foundation Brain Research Program (2017M3C7A1048086).

This paper is behind a paywall. For the players amongst us, here’s the Eyewire website. Go forth,  play, and, maybe, discover new neurons!

The sound of frogs (and other amphibians) and climate change

At least once a year I highlight some work about frogs. It’s usually about a new species but this time, it’s all about frog sounds (as well as, sounds from other amphibians).

Caption: The calls of the midwife toad and other amphibians have served to test the sound classifier. Credit: Jaime Bosch (MNCN-CSIC)

In any event, here’s more from an April 30, 2018 Spanish Foundation for Science and Technology (FECYT) press release (also on EurekAlert but with a May 17, 2018 publication date),

The sounds of amphibians are altered by the increase in ambient temperature, a phenomenon that, in addition to interfering with reproductive behaviour, serves as an indicator of global warming. Researchers at the University of Seville have resorted to artificial intelligence to create an automatic classifier of the thousands of frog and toad sounds that can be recorded in a natural environment.

One of the consequences of climate change is its impact on the physiological functions of animals, such as frogs and toads with their calls. Their mating call, which plays a crucial role in the sexual selection and reproduction of these amphibians, is affected by the increase in ambient temperature.

When this exceeds a certain threshold, the physiological processes associated with the sound production are restricted, and some calls are even actually inhibited. In fact, the beginning, duration and intensity of calls from the male to the female are changed, which influences reproductive activity.

Taking into account this phenomenon, the analysis and classification of the sounds produced by certain species of amphibians and other animals have turned out to be a powerful indicator of temperature fluctuations and, therefore, of the existence and evolution of global warming.

To capture the sounds of frogs, networks of audio sensors are placed and connected wirelessly in areas that can reach several hundred square kilometres. The problem is that a huge amount of bio-acoustic information is collected in environments as noisy as a jungle, and this makes it difficult to identify the species and their calls.

To solve this, engineers from the University of Seville have resorted to artificial intelligence. “We’ve segmented the sound into temporary windows or audio frames and have classified them by means of decision trees, an automatic learning technique that is used in computing”, explains Amalia Luque Sendra, co-author of the work.

To perform the classification, the researchers have based it on MPEG-7 parameters and audio descriptors, a standard way of representing audiovisual information. The details are published in Expert Systems with Applications magazine.

This technique has been put to the test with real sounds of amphibians recorded in the middle of nature and provided by the National Museum of Natural Sciences. More specifically, 868 records with 369 mating calls sung by the male and 63 release calls issued by the female natterajck toad (Epidalea calamita), along with 419 mating calls and 17 distress calls of the common midwife toad (Alytesobstetricans).

“In this case we obtained a success rate close to 90% when classifying the sounds,” observes Luque Sendra, who recalls that, in addition to the types of calls, the number of individuals of certain amphibian species that are heard in a geographical region over time can also be used as an indicator of climate change.

“A temperature increase affects the calling patterns,” she says, “but since these in most cases have a sexual calling nature, they also affect the number of individuals. With our method, we still can’t directly determine the exact number of specimens in an area, but it is possible to get a first approximation.”

In addition to the image of the midwife toad, the researchers included this image to illustrate their work,

Caption: This is the architecture of a wireless sensor network. Credit: J. Luque et al./Sensors

Here’s a link to and a citation for the paper,

Non-sequential automatic classification of anuran sounds for the estimation of climate-change indicators by Amalia Luque, Javier Romero-Lemos, Alejandro Carrasco, Julio Barbancho. Expert Systems with Applications Volume 95, 1 April 2018, Pages 248-260 DOI: https://doi.org/10.1016/j.eswa.2017.11.016 Available online 10 November 2017

This paper is open access.

Embedded AI (artificial intelligence) with a variant of a memristor?

I don’t entirely get how ReRAM (resistive random access memory) is a variant of a memristor but I’m assuming Samuel K. Moore knows what he’s writing about since his May 16, 2018 posting is on the Nanoclast blog (hosted by the IEEE [Institute of Electrical and Electronics Engineers]), Note: Links have been removed,

Resistive RAM technology developer Crossbar says it has inked a deal with aerospace chip maker Microsemi allowing the latter to embed Crossbar’s nonvolatile memory on future chips. The move follows selection of Crossbar’s technology by a leading foundry for advanced manufacturing nodes. Crossbar is counting on resistive RAM (ReRAM) to enable artificial intelligence systems whose neural networks are housed within the device rather than in the cloud.

ReRAM is a variant of the memristor, a nonvolatile memory device whose resistance can be set or reset by a pulse of voltage. The variant Crossbar qualified for advanced manufacturing is called a filament device. It’s built within the layers above a chip’s silicon, where the IC’s interconnects go, and it’s made up of three layers: from top to bottom—silver, amorphous silicon, and tungsten. Voltage across the amorphous silicon causes a filament of silver atoms to cross the gap to the tungsten, making the memory cell conductive. Reversing the voltage pushes the silver back into place, cutting off conduction.

“The filament itself is only three to four nanometers wide,” says Sylvain Dubois, vice president of marketing and business development at Crossbar. “So the cell itself will be able to scale below 10-nanometers.” What’s more, the ratio between the current that flows when the device is on to when it is off is 1,000 or higher. …

A May 14, 2018 Crossbar news release describes some of the technical AI challenges,

“The biggest challenge facing engineers for AI today is overcoming the memory speed and power bottleneck in the current architecture to get faster data access while lowering the energy cost,” said Dubois. “By enabling a new, memory-centric non-volatile architecture like ReRAM, the entire trained model or knowledge base can be on-chip, connected directly to the neural network with the potential to achieve massive energy savings and performance improvements, resulting in a greatly improved battery life and a better user experience.”

Crossbar’s May 16, 2018 news release provides more detail about their ‘strategic collaboration’ with Microsemi Products (Note: A link has been removed),

Crossbar Inc., the ReRAM technology leader, announced an agreement with Microsemi Corporation, the largest U.S. commercial supplier of military and aerospace semiconductors, in which Microsemi will license Crossbar’s ReRAM core intellectual property. As part of the agreement, Microsemi and Crossbar will collaborate in the research, development and application of Crossbar’s proprietary ReRAM technology in next generation products from Microsemi that integrate Crossbar’s embedded ReRAM with Microsemi products manufactured at the 1x nm process node.

Military and aerospace, eh?

Being smart about using artificial intelligence in the field of medicine

Since my August 20, 2018 post featured an opinion piece about the possibly imminent replacement of radiologists with artificial intelligence systems and the latest research about employing them for diagnosing eye diseases, it seems like a good time to examine some of the mythology embedded in the discussion about AI and medicine.

Imperfections in medical AI systems

An August 15, 2018 article for Slate.com by W. Nicholson Price II (who teaches at the University of Michigan School of Law; in addition to his law degree he has a PhD in Biological Sciences from Columbia University) begins with the peppy, optimistic view before veering into more critical territory (Note: Links have been removed),

For millions of people suffering from diabetes, new technology enabled by artificial intelligence promises to make management much easier. Medtronic’s Guardian Connect system promises to alert users 10 to 60 minutes before they hit high or low blood sugar level thresholds, thanks to IBM Watson, “the same supercomputer technology that can predict global weather patterns.” Startup Beta Bionics goes even further: In May, it received Food and Drug Administration approval to start clinical trials on what it calls a “bionic pancreas system” powered by artificial intelligence, capable of “automatically and autonomously managing blood sugar levels 24/7.”

An artificial pancreas powered by artificial intelligence represents a huge step forward for the treatment of diabetes—but getting it right will be hard. Artificial intelligence (also known in various iterations as deep learning and machine learning) promises to automatically learn from patterns in medical data to help us do everything from managing diabetes to finding tumors in an MRI to predicting how long patients will live. But the artificial intelligence techniques involved are typically opaque. We often don’t know how the algorithm makes the eventual decision. And they may change and learn from new data—indeed, that’s a big part of the promise. But when the technology is complicated, opaque, changing, and absolutely vital to the health of a patient, how do we make sure it works as promised?

Price describes how a ‘closed loop’ artificial pancreas with AI would automate insulin levels for diabetic patients, flaws in the automated system, and how companies like to maintain a competitive advantage (Note: Links have been removed),

[…] a “closed loop” artificial pancreas, where software handles the whole issue, receiving and interpreting signals from the monitor, deciding when and how much insulin is needed, and directing the insulin pump to provide the right amount. The first closed-loop system was approved in late 2016. The system should take as much of the issue off the mind of the patient as possible (though, of course, that has limits). Running a close-loop artificial pancreas is challenging. The way people respond to changing levels of carbohydrates is complicated, as is their response to insulin; it’s hard to model accurately. Making it even more complicated, each individual’s body reacts a little differently.

Here’s where artificial intelligence comes into play. Rather than trying explicitly to figure out the exact model for how bodies react to insulin and to carbohydrates, machine learning methods, given a lot of data, can find patterns and make predictions. And existing continuous glucose monitors (and insulin pumps) are excellent at generating a lot of data. The idea is to train artificial intelligence algorithms on vast amounts of data from diabetic patients, and to use the resulting trained algorithms to run a closed-loop artificial pancreas. Even more exciting, because the system will keep measuring blood glucose, it can learn from the new data and each patient’s artificial pancreas can customize itself over time as it acquires new data from that patient’s particular reactions.

Here’s the tough question: How will we know how well the system works? Diabetes software doesn’t exactly have the best track record when it comes to accuracy. A 2015 study found that among smartphone apps for calculating insulin doses, two-thirds of the apps risked giving incorrect results, often substantially so. … And companies like to keep their algorithms proprietary for a competitive advantage, which makes it hard to know how they work and what flaws might have gone unnoticed in the development process.

There’s more,

These issues aren’t unique to diabetes care—other A.I. algorithms will also be complicated, opaque, and maybe kept secret by their developers. The potential for problems multiplies when an algorithm is learning from data from an entire hospital, or hospital system, or the collected data from an entire state or nation, not just a single patient. …

The [US Food and Drug Administraiont] FDA is working on this problem. The head of the agency has expressed his enthusiasm for bringing A.I. safely into medical practice, and the agency has a new Digital Health Innovation Action Plan to try to tackle some of these issues. But they’re not easy, and one thing making it harder is a general desire to keep the algorithmic sauce secret. The example of IBM Watson for Oncology has given the field a bit of a recent black eye—it turns out that the company knew the algorithm gave poor recommendations for cancer treatment but kept that secret for more than a year. …

While Price focuses on problems with algorithms and with developers and their business interests, he also hints at some of the body’s complexities.

Can AI systems be like people?

Susan Baxter, a medical writer with over 20 years experience, a PhD in health economics, and author of countless magazine articles and several books, offers a more person-centered approach to the discussion in her July 6, 2018 posting on susanbaxter.com,

The fascination with AI continues to irk, given that every second thing I read seems to be extolling the magic of AI and medicine and how It Will Change Everything. Which it will not, trust me. The essential issue of illness remains perennial and revolves around an individual for whom no amount of technology will solve anything without human contact. …

But in this world, or so we are told by AI proponents, radiologists will soon be obsolete. [my August 20, 2018 post] The adaptational learning capacities of AI mean that reading a scan or x-ray will soon be more ably done by machines than humans. The presupposition here is that we, the original programmers of this artificial intelligence, understand the vagaries of real life (and real disease) so wonderfully that we can deconstruct these much as we do the game of chess (where, let’s face it, Big Blue ate our lunch) and that analyzing a two-dimensional image of a three-dimensional body, already problematic, can be reduced to a series of algorithms.

Attempting to extrapolate what some “shadow” on a scan might mean in a flesh and blood human isn’t really quite the same as bishop to knight seven. Never mind the false positive/negatives that are considered an acceptable risk or the very real human misery they create.

Moravec called it

It’s called Moravec’s paradox, the inability of humans to realize just how complex basic physical tasks are – and the corresponding inability of AI to mimic it. As you walk across the room, carrying a glass of water, talking to your spouse/friend/cat/child; place the glass on the counter and open the dishwasher door with your foot as you open a jar of pickles at the same time, take a moment to consider just how many concurrent tasks you are doing and just how enormous the computational power these ostensibly simple moves would require.

Researchers in Singapore taught industrial robots to assemble an Ikea chair. Essentially, screw in the legs. A person could probably do this in a minute. Maybe two. The preprogrammed robots took nearly half an hour. And I suspect programming those robots took considerably longer than that.

Ironically, even Elon Musk, who has had major production problems with the Tesla cars rolling out of his high tech factory, has conceded (in a tweet) that “Humans are underrated.”

I wouldn’t necessarily go that far given the political shenanigans of Trump & Co. but in the grand scheme of things I tend to agree. …

Is AI going the way of gene therapy?

Susan draws a parallel between the AI and medicine discussion with the discussion about genetics and medicine (Note: Links have been removed),

On a somewhat similar note – given the extent to which genetics discourse has that same linear, mechanistic  tone [as AI and medicine] – it turns out all this fine talk of using genetics to determine health risk and whatnot is based on nothing more than clever marketing, since a lot of companies are making a lot of money off our belief in DNA. Truth is half the time we don’t even know what a gene is never mind what it actually does;  geneticists still can’t agree on how many genes there are in a human genome, as this article in Nature points out.

Along the same lines, I was most amused to read about something called the Super Seniors Study, research following a group of individuals in their 80’s, 90’s and 100’s who seem to be doing really well. Launched in 2002 and headed by Angela Brooks Wilson, a geneticist at the BC [British Columbia] Cancer Agency and SFU [Simon Fraser University] Chair of biomedical physiology and kinesiology, this longitudinal work is examining possible factors involved in healthy ageing.

Turns out genes had nothing to do with it, the title of the Globe and Mail article notwithstanding. (“Could the DNA of these super seniors hold the secret to healthy aging?” The answer, a resounding “no”, well hidden at the very [end], the part most people wouldn’t even get to.) All of these individuals who were racing about exercising and working part time and living the kind of life that makes one tired just reading about it all had the same “multiple (genetic) factors linked to a high probability of disease”. You know, the gene markers they tell us are “linked” to cancer, heart disease, etc., etc. But these super seniors had all those markers but none of the diseases, demonstrating (pretty strongly) that the so-called genetic links to disease are a load of bunkum. Which (she said modestly) I have been saying for more years than I care to remember. You’re welcome.

The fundamental error in this type of linear thinking is in allowing our metaphors (genes are the “blueprint” of life) and propensity towards social ideas of determinism to overtake common sense. Biological and physiological systems are not static; they respond to and change to life in its entirety, whether it’s diet and nutrition to toxic or traumatic insults. Immunity alters, endocrinology changes, – even how we think and feel affects the efficiency and effectiveness of physiology. Which explains why as we age we become increasingly dissimilar.

If you have the time, I encourage to read Susan’s comments in their entirety.

Scientific certainties

Following on with genetics, gene therapy dreams, and the complexity of biology, the June 19, 2018 Nature article by Cassandra Willyard (mentioned in Susan’s posting) highlights an aspect of scientific research not often mentioned in public,

One of the earliest attempts to estimate the number of genes in the human genome involved tipsy geneticists, a bar in Cold Spring Harbor, New York, and pure guesswork.

That was in 2000, when a draft human genome sequence was still in the works; geneticists were running a sweepstake on how many genes humans have, and wagers ranged from tens of thousands to hundreds of thousands. Almost two decades later, scientists armed with real data still can’t agree on the number — a knowledge gap that they say hampers efforts to spot disease-related mutations.

In 2000, with the genomics community abuzz over the question of how many human genes would be found, Ewan Birney launched the GeneSweep contest. Birney, now co-director of the European Bioinformatics Institute (EBI) in Hinxton, UK, took the first bets at a bar during an annual genetics meeting, and the contest eventually attracted more than 1,000 entries and a US$3,000 jackpot. Bets on the number of genes ranged from more than 312,000 to just under 26,000, with an average of around 40,000. These days, the span of estimates has shrunk — with most now between 19,000 and 22,000 — but there is still disagreement (See ‘Gene Tally’).

… the inconsistencies in the number of genes from database to database are problematic for researchers, Pruitt says. “People want one answer,” she [Kim Pruitt, a genome researcher at the US National Center for Biotechnology Information {NCB}] in Bethesda, Maryland] adds, “but biology is complex.”

I wanted to note that scientists do make guesses and not just with genetics. For example, Gina Mallet’s 2005 book ‘Last Chance to Eat: The Fate of Taste in a Fast Food World’ recounts the story of how good and bad levels of cholesterol were established—the experts made some guesses based on their experience. That said, Willyard’s article details the continuing effort to nail down the number of genes almost 20 years after the human genome project was completed and delves into the problems the scientists have uncovered.

Final comments

In addition to opaque processes with developers/entrepreneurs wanting to maintain their secrets for competitive advantages and in addition to our own poor understanding of the human body (how many genes are there anyway?), there are same major gaps (reflected in AI) in our understanding of various diseases. Angela Lashbrook’s August 16, 2018 article for The Atlantic highlights some issues with skin cancer and shade of your skin (Note: Links have been removed),

… While fair-skinned people are at the highest risk for contracting skin cancer, the mortality rate for African Americans is considerably higher: Their five-year survival rate is 73 percent, compared with 90 percent for white Americans, according to the American Academy of Dermatology.

As the rates of melanoma for all Americans continue a 30-year climb, dermatologists have begun exploring new technologies to try to reverse this deadly trend—including artificial intelligence. There’s been a growing hope in the field that using machine-learning algorithms to diagnose skin cancers and other skin issues could make for more efficient doctor visits and increased, reliable diagnoses. The earliest results are promising—but also potentially dangerous for darker-skinned patients.

… Avery Smith, … a software engineer in Baltimore, Maryland, co-authored a paper in JAMA [Journal of the American Medical Association] Dermatology that warns of the potential racial disparities that could come from relying on machine learning for skin-cancer screenings. Smith’s co-author, Adewole Adamson of the University of Texas at Austin, has conducted multiple studies on demographic imbalances in dermatology. “African Americans have the highest mortality rate [for skin cancer], and doctors aren’t trained on that particular skin type,” Smith told me over the phone. “When I came across the machine-learning software, one of the first things I thought was how it will perform on black people.”

Recently, a study that tested machine-learning software in dermatology, conducted by a group of researchers primarily out of Germany, found that “deep-learning convolutional neural networks,” or CNN, detected potentially cancerous skin lesions better than the 58 dermatologists included in the study group. The data used for the study come from the International Skin Imaging Collaboration, or ISIC, an open-source repository of skin images to be used by machine-learning algorithms. Given the rise in melanoma cases in the United States, a machine-learning algorithm that assists dermatologists in diagnosing skin cancer earlier could conceivably save thousands of lives each year.

… Chief among the prohibitive issues, according to Smith and Adamson, is that the data the CNN relies on come from primarily fair-skinned populations in the United States, Australia, and Europe. If the algorithm is basing most of its knowledge on how skin lesions appear on fair skin, then theoretically, lesions on patients of color are less likely to be diagnosed. “If you don’t teach the algorithm with a diverse set of images, then that algorithm won’t work out in the public that is diverse,” says Adamson. “So there’s risk, then, for people with skin of color to fall through the cracks.”

As Adamson and Smith’s paper points out, racial disparities in artificial intelligence and machine learning are not a new issue. Algorithms have mistaken images of black people for gorillas, misunderstood Asians to be blinking when they weren’t, and “judged” only white people to be attractive. An even more dangerous issue, according to the paper, is that decades of clinical research have focused primarily on people with light skin, leaving out marginalized communities whose symptoms may present differently.

The reasons for this exclusion are complex. According to Andrew Alexis, a dermatologist at Mount Sinai, in New York City, and the director of the Skin of Color Center, compounding factors include a lack of medical professionals from marginalized communities, inadequate information about those communities, and socioeconomic barriers to participating in research. “In the absence of a diverse study population that reflects that of the U.S. population, potential safety or efficacy considerations could be missed,” he says.

Adamson agrees, elaborating that with inadequate data, machine learning could misdiagnose people of color with nonexistent skin cancers—or miss them entirely. But he understands why the field of dermatology would surge ahead without demographically complete data. “Part of the problem is that people are in such a rush. This happens with any new tech, whether it’s a new drug or test. Folks see how it can be useful and they go full steam ahead without thinking of potential clinical consequences. …

Improving machine-learning algorithms is far from the only method to ensure that people with darker skin tones are protected against the sun and receive diagnoses earlier, when many cancers are more survivable. According to the Skin Cancer Foundation, 63 percent of African Americans don’t wear sunscreen; both they and many dermatologists are more likely to delay diagnosis and treatment because of the belief that dark skin is adequate protection from the sun’s harmful rays. And due to racial disparities in access to health care in America, African Americans are less likely to get treatment in time.

Happy endings

I’ll add one thing to Price’s article, Susan’s posting, and Lashbrook’s article about the issues with AI , certainty, gene therapy, and medicine—the desire for a happy ending prefaced with an easy solution. If the easy solution isn’t possible accommodations will be made but that happy ending is a must. All disease will disappear and there will be peace on earth. (Nod to Susan Baxter and her many discussions with me about disease processes and happy endings.)

The solutions, for the most part, are seen as technological despite the mountain of evidence suggesting that technology reflects our own imperfect understanding of health and disease therefore providing what is at best an imperfect solution.

Also, we tend to underestimate just how complex humans are not only in terms of disease and health but also with regard to our skills, understanding, and, perhaps not often enough, our ability to respond appropriately in the moment.

There is much to celebrate in what has been accomplished: no more black death, no more smallpox, hip replacements, pacemakers, organ transplants, and much more. Yes, we should try to improve our medicine. But, maybe alongside the celebration we can welcome AI and other technologies with a lot less hype and a lot more skepticism.

AI x 2: the Amnesty International and Artificial Intelligence story

Amnesty International and artificial intelligence seem like an unexpected combination but it all makes sense when you read a June 13, 2018 article by Steven Melendez for Fast Company (Note: Links have been removed),

If companies working on artificial intelligence don’t take steps to safeguard human rights, “nightmare scenarios” could unfold, warns Rasha Abdul Rahim, an arms control and artificial intelligence researcher at Amnesty International in a blog post. Those scenarios could involve armed, autonomous systems choosing military targets with little human oversight, or discrimination caused by biased algorithms, she warns.

Rahim pointed at recent reports of Google’s involvement in the Pentagon’s Project Maven, which involves harnessing AI image recognition technology to rapidly process photos taken by drones. Google recently unveiled new AI ethics policies and has said it won’t continue with the project once its current contract expires next year after high-profile employee dissent over the project. …

“Compliance with the laws of war requires human judgement [sic] –the ability to analyze the intentions behind actions and make complex decisions about the proportionality or necessity of an attack,” Rahim writes. “Machines and algorithms cannot recreate these human skills, and nor can they negotiate, produce empathy, or respond to unpredictable situations. In light of these risks, Amnesty International and its partners in the Campaign to Stop Killer Robots are calling for a total ban on the development, deployment, and use of fully autonomous weapon systems.”

Rasha Abdul Rahim’s June 14, 2018 posting (I’m putting the discrepancy in publication dates down to timezone differences) on the Amnesty International website (Note: Links have been removed),

Last week [June 7, 2018] Google released a set of principles to govern its development of AI technologies. They include a broad commitment not to design or deploy AI in weaponry, and come in the wake of the company’s announcement that it will not renew its existing contract for Project Maven, the US Department of Defense’s AI initiative, when it expires in 2019.

The fact that Google maintains its existing Project Maven contract for now raises an important question. Does Google consider that continuing to provide AI technology to the US government’s drone programme is in line with its new principles? Project Maven is a litmus test that allows us to see what Google’s new principles mean in practice.

As details of the US drone programme are shrouded in secrecy, it is unclear precisely what role Google plays in Project Maven. What we do know is that US drone programme, under successive administrations, has been beset by credible allegations of unlawful killings and civilian casualties. The cooperation of Google, in any capacity, is extremely troubling and could potentially implicate it in unlawful strikes.

As AI technology advances, the question of who will be held accountable for associated human rights abuses is becoming increasingly urgent. Machine learning, and AI more broadly, impact a range of human rights including privacy, freedom of expression and the right to life. It is partly in the hands of companies like Google to safeguard these rights in relation to their operations – for us and for future generations. If they don’t, some nightmare scenarios could unfold.

Warfare has already changed dramatically in recent years – a couple of decades ago the idea of remote controlled bomber planes would have seemed like science fiction. While the drones currently in use are still controlled by humans, China, France, Israel, Russia, South Korea, the UK and the US are all known to be developing military robots which are getting smaller and more autonomous.

For example, the UK is developing a number of autonomous systems, including the BAE [Systems] Taranis, an unmanned combat aircraft system which can fly in autonomous mode and automatically identify a target within a programmed area. Kalashnikov, the Russian arms manufacturer, is developing a fully automated, high-calibre gun that uses artificial neural networks to choose targets. The US Army Research Laboratory in Maryland, in collaboration with BAE Systems and several academic institutions, has been developing micro drones which weigh less than 30 grams, as well as pocket-sized robots that can hop or crawl.

Of course, it’s not just in conflict zones that AI is threatening human rights. Machine learning is already being used by governments in a wide range of contexts that directly impact people’s lives, including policing [emphasis mine], welfare systems, criminal justice and healthcare. Some US courts use algorithms to predict future behaviour of defendants and determine their sentence lengths accordingly. The potential for this approach to reinforce power structures, discrimination or inequalities is huge.

In july 2017, the Vancouver Police Department announced its use of predictive policing software, the first such jurisdiction in Canada to make use of the technology. My Nov. 23, 2017 posting featured the announcement.

The almost too aptly named Campaign to Stop Killer Robots can be found here. Their About Us page provides a brief history,

Formed by the following non-governmental organizations (NGOs) at a meeting in New York on 19 October 2012 and launched in London in April 2013, the Campaign to Stop Killer Robots is an international coalition working to preemptively ban fully autonomous weapons. See the Chronology charting our major actions and achievements to date.

Steering Committee

The Steering Committee is the campaign’s principal leadership and decision-making body. It is comprised of five international NGOs, a regional NGO network, and four national NGOs that work internationally:

Human Rights Watch
Article 36
Association for Aid and Relief Japan
International Committee for Robot Arms Control
Mines Action Canada
Nobel Women’s Initiative
PAX (formerly known as IKV Pax Christi)
Pugwash Conferences on Science & World Affairs
Seguridad Humana en América Latina y el Caribe (SEHLAC)
Women’s International League for Peace and Freedom

For more information, see this Overview. A Terms of Reference is also available on request, detailing the committee’s selection process, mandate, decision-making, meetings and communication, and expected commitments.

For anyone who may be interested in joining Amnesty International, go here.

University of Waterloo (Canada) team combines wearable tech with artificial intelligence (AI) for health

A May 16, 2018 University of Waterloo news release (also on EurekAlert) trumpets the research,

A team of Waterloo researchers found that applying artificial intelligence to the right combination of data retrieved from wearable technology may detect whether your health is failing.

The study, which involved researchers from Waterloo’s Faculties of Applied Health Sciences and Engineering, found that the data from wearable sensors and artificial intelligence that assesses changes in aerobic responses could one day predict whether a person is experiencing the onset of a respiratory or cardiovascular disease.

“The onset of a lot of chronic diseases, including type 2 diabetes and chronic obstructive pulmonary disease, has a direct impact on our aerobic fitness,” said Thomas Beltrame, who led the research while at the University of Waterloo, and is now at the Institute of Computing in University of Campinas in Brazil. “In the near future, we believe it will be possible to continuously check your health, even before you realize that you need medical help.”

The study monitored active, healthy men in their twenties who wore a shirt for four days that incorporated sensors for heart rate, breathing and acceleration. They then compared the readings with laboratory responses and found that it was possible to accurately predict health-related benchmarks during daily activities using only the smart shirt.

“The research found a way to process biological signals and generate a meaningful single number to track fitness,” said Richard Hughson, co-author and kinesiology professor at the Schlegel-University of Waterloo Research Institute for Aging.

Beltrame and Hughson co-authored the study with Alexander Wong, Canada Research Chair in artificial intelligence and medical imaging and an engineering professor at Waterloo. He is affiliated with both the Waterloo Artificial Intelligence Institute and the Schlegel-University of Waterloo Research Institute for Aging. Robert Amelard, of the Schlegel-University of Waterloo Research Institute for Aging, is also a co-author. The study appears in the Journal of Applied Physiology.

“This multi-disciplinary research is a great example of how artificial intelligence can be a potential game-changer for healthcare by turning data into predictive knowledge to help healthcare professionals better understand an individual’s health,” said Wong. “It can have a significant impact on improving quality of life and well-being.”

Carré Technologies developed the smart shirts, called Hexoskin, used in the research.

The team plans to test these systems on mixed ages and genders, and people with health issues to see how people might wear the sensors to gauge whether their health is failing.

I wonder if this is the 2nd try for publicity about this work. Take a look at the publication date,

Extracting aerobic system dynamics during unsupervised activities of daily living using wearable sensor machine learning models by Thomas Beltrame, Robert Amelard, Alexander Wong, and Richard L. Hughson. Journal of Applied Physiology 124 (2)
Volume 124Issue 2February 2018Pages 473-48 https://doi.org/10.1152/japplphysiol.00299.2017 [Published] 23 Feb 2018

This paper is behind a paywall.

Interested parties can find Carré Technologies here.

May 28, 2018 release date for inorganic material database ‘AtomWork-Adv’

Announced in a May 23, 2018 news item on Nanowerk,

[Japan National Institute for Materials Science] NIMS will make its inorganic materials database, AtomWork-Adv (pronounced “atom work advanced”), available to the general public as a fee-based service starting Monday, May 28, 2018. This service will be provided by the Data Platform from the Center for Materials Research by Information Integration (CMI2), Research and Services Division of the Materials Data and Integrated System (MaDIS), NIMS.

A May 23, 2018 NIMS press release, which originated the news item, fills out the details (Note: Paragraph 1 is largely repetitive but there is contact information in there),

  1. NIMS will make its inorganic materials database, “AtomWork-Adv” (pronounced “atom work advanced”), available to the general public as a fee-based service starting Monday, May 28, 2018. This service will be provided by the Data Platform (Yibin Xu, Director) from the Center for Materials Research by Information Integration (CMI2), Research and Services Division of the Materials Data and Integrated System (MaDIS), NIMS. AtomWork-Adv markedly improves upon the amount of data available and the usability of the current web-based “AtomWork” database. We hope that this service will promote data-driven materials development using AI and machine learning.
  2. The increasingly popular use of AI and machine learning in materials development requires a high-quality materials database. NIMS publicized the AtomWork inorganic materials database—constructed using literature published up to 2002—on its MatNavi webpage (http://mits.nims.go.jp/index.html). While AtomWork is available free of charge, the data it contains is not updated and its functions, such as allowing users to copy, download and search data, are limited.
  3. The AtomWork-Adv database compiles crystal structure data (approx. 274,000 datasets), X-ray diffraction data (approx. 496,000 datasets), material properties data (approx. 298,000 datasets) and phase diagram data (approx. 40,000 datasets) collected from literature published up to 2014. The number of datasets in these categories are about three to five times greater than those in the Atom Work database. In addition, AtomWork-Adv is equipped with user-friendly functions, such as a versatile search tool which enables searches by element, composition, crystal structure and material property, a matrix function which identifies the number of datasets available for a specific binary material combination and an automatic charting function which allows the plotting of a graph between two material property variables and which displays material names in relation to these variables. Users can download data for use in data science-driven materials research and development (the number of downloads is restricted).
  4. Data will be continuously added to and updated in the fee-based AtomWork-Adv database. We are planning to add new data collected from literature between 2015 and 2016 to the database during FY2018.
  5. This database project was supported by the “Materials Research by Information Integration” Initiative (MI2I) sponsored by the Japan Science and Technology Agency (JST)’s Support Program for Starting up Innovation Hub.

Happy atom hunting!

AI (artificial intelligence) for Good Global Summit from May 15 – 17, 2018 in Geneva, Switzerland: details and an interview with Frederic Werner

With all the talk about artificial intelligence (AI), a lot more attention seems to be paid to apocalyptic scenarios: loss of jobs, financial hardship, loss of personal agency and privacy, and more with all of these impacts being described as global. Still, there are some folks who are considering and working on ‘AI for good’.

If you’d asked me, the International Telecommunications Union (ITU) would not have been my first guess (my choice would have been United Nations Educational, Scientific and Cultural Organization [UNESCO]) as an agency likely to host the 2018 AI for Good Global Summit. But, it turns out the ITU is a UN (United Nations agency) and, according to its Wikipedia entry, it’s an intergovernmental public-private partnership, which may explain the nature of the participants in the upcoming summit.

The news

First, there’s a May 4, 2018 ITU media advisory (received via email or you can find the full media advisory here) about the upcoming summit,

Artificial Intelligence (AI) is now widely identified as being able to address the greatest challenges facing humanity – supporting innovation in fields ranging from crisis management and healthcare to smart cities and communications networking.

The second annual ‘AI for Good Global Summit’ will take place 15-17 May [2018] in Geneva, and seeks to leverage AI to accelerate progress towards the United Nations’ Sustainable Development Goals and ultimately benefit humanity.

WHAT: Global event to advance ‘AI for Good’ with the participation of internationally recognized AI experts. The programme will include interactive high-level panels, while ‘AI Breakthrough Teams’ will propose AI strategies able to create impact in the near term, guided by an expert audience of mentors representing government, industry, academia and civil society – through interactive sessions. The summit will connect AI innovators with public and private-sector decision-makers, building collaboration to take promising strategies forward.

A special demo & exhibit track will feature innovative applications of AI designed to: protect women from sexual violence, avoid infant crib deaths, end child abuse, predict oral cancer, and improve mental health treatments for depression – as well as interactive robots including: Alice, a Dutch invention designed to support the aged; iCub, an open-source robot; and Sophia, the humanoid AI robot.

WHEN: 15-17 May 2018, beginning daily at 9 AM

WHERE: ITU Headquarters, 2 Rue de Varembé, Geneva, Switzerland (Please note: entrance to ITU is now limited for all visitors to the Montbrillant building entrance only on rue Varembé).

WHO: Confirmed participants to date include expert representatives from: Association for Computing Machinery, Bill and Melinda Gates Foundation, Cambridge University, Carnegie Mellon, Chan Zuckerberg Initiative, Consumer Trade Association, Facebook, Fraunhofer, Google, Harvard University, IBM Watson, IEEE, Intellectual Ventures, ITU, Microsoft, Massachusetts Institute of Technology (MIT), Partnership on AI, Planet Labs, Shenzhen Open Innovation Lab, University of California at Berkeley, University of Tokyo, XPRIZE Foundation, Yale University – and the participation of “Sophia” the humanoid robot and “iCub” the EU open source robotcub.

The interview

Frederic Werner, Senior Communications Officer at the International Telecommunication Union and** one of the organizers of the AI for Good Global Summit 2018 kindly took the time to speak to me and provide a few more details about the upcoming event.

Werner noted that the 2018 event grew out of a much smaller 2017 ‘workshop’ and first of its kind, about beneficial AI which this year has ballooned in size to 91 countries (about 15 participants are expected from Canada), 32 UN agencies, and substantive representation from the private sector. The 2017 event featured Dr. Yoshua Bengio of the University of Montreal  (Université de Montréal) was a featured speaker.

“This year, we’re focused on action-oriented projects that will help us reach our Sustainable Development Goals (SDGs) by 2030. We’re looking at near-term practical AI applications,” says Werner. “We’re matchmaking problem-owners and solution-owners.”

Academics, industry professionals, government officials, and representatives from UN agencies are gathering  to work on four tracks/themes:

In advance of this meeting, the group launched an AI repository (an action item from the 2017 meeting) on April 25, 2018 inviting people to list their AI projects (from the ITU’s April 25, 2018? AI repository news announcement),

ITU has just launched an AI Repository where anyone working in the field of artificial intelligence (AI) can contribute key information about how to leverage AI to help solve humanity’s greatest challenges.

This is the only global repository that identifies AI-related projects, research initiatives, think-tanks and organizations that aim to accelerate progress on the 17 United Nations’ Sustainable Development Goals (SDGs).

To submit a project, just press ‘Submit’ on the AI Repository site and fill in the online questionnaire, providing all relevant details of your project. You will also be asked to map your project to the relevant World Summit on the Information Society (WSIS) action lines and the SDGs. Approved projects will be officially registered in the repository database.

Benefits of participation on the AI Repository include:

WSIS Prizes recognize individuals, governments, civil society, local, regional and international agencies, research institutions and private-sector companies for outstanding success in implementing development oriented strategies that leverage the power of AI and ICTs.

Creating the AI Repository was one of the action items of last year’s AI for Good Global Summit.

We are looking forward to your submissions.

If you have any questions, please send an email to: ai@itu.int

“Your project won’t be visible immediately as we have to vet the submissions to weed out spam-type material and projects that are not in line with our goals,” says Werner. That said, there are already 29 projects in the repository. As you might expect, the UK, China, and US are in the repository but also represented are Egypt, Uganda, Belarus, Serbia, Peru, Italy, and other countries not commonly cited when discussing AI research.

Werner also pointed out in response to my surprise over the ITU’s role with regard to this AI initiative that the ITU is the only UN agency which has 192* member states (countries), 150 universities, and over 700 industry members as well as other member entities, which gives them tremendous breadth of reach. As well, the organization, founded originally in 1865 as the International Telegraph Convention, has extensive experience with global standardization in the information technology and telecommunications industries. (See more in their Wikipedia entry.)

Finally

There is a bit more about the summit on the ITU’s AI for Good Global Summit 2018 webpage,

The 2nd edition of the AI for Good Global Summit will be organized by ITU in Geneva on 15-17 May 2018, in partnership with XPRIZE Foundation, the global leader in incentivized prize competitions, the Association for Computing Machinery (ACM) and sister United Nations agencies including UNESCO, UNICEF, UNCTAD, UNIDO, Global Pulse, UNICRI, UNODA, UNIDIR, UNODC, WFP, IFAD, UNAIDS, WIPO, ILO, UNITAR, UNOPS, OHCHR, UN UniversityWHO, UNEP, ICAO, UNDP, The World Bank, UN DESA, CTBTOUNISDRUNOG, UNOOSAUNFPAUNECE, UNDPA, and UNHCR.

The AI for Good series is the leading United Nations platform for dialogue on AI. The action​​-oriented 2018 summit will identify practical applications of AI and supporting strategies to improve the quality and sustainability of life on our planet. The summit will continue to formulate strategies to ensure trusted, safe and inclusive development of AI technologies and equitable access to their benefits.

While the 2017 summit sparked the first ever inclusive global dialogue on beneficial AI, the action-oriented 2018 summit will focus on impactful AI solutions able to yield long-term benefits and help achieve the Sustainable Development Goals. ‘Breakthrough teams’ will demonstrate the potential of AI to map poverty and aid with natural disasters using satellite imagery, how AI could assist the delivery of citizen-centric services in smart cities, and new opportunities for AI to help achieve Universal Health Coverage, and finally to help achieve transparency and explainability in AI algorithms.

Teams will propose impactful AI strategies able to be enacted in the near term, guided by an expert audience of mentors representing government, industry, academia and civil society. Strategies will be evaluated by the mentors according to their feasibility and scalability, potential to address truly global challenges, degree of supporting advocacy, and applicability to market failures beyond the scope of government and industry. The exercise will connect AI innovators with public and private-sector decision-makers, building collaboration to take promising strategies forward.

“As the UN specialized agency for information and communication technologies, ITU is well placed to guide AI innovation towards the achievement of the UN Sustainable Development ​Goals. We are providing a neutral close quotation markplatform for international dialogue aimed at ​building a ​common understanding of the capabilities of emerging AI technologies.​​” Houlin Zhao, Secretary General ​of ITU​

Should you be close to Geneva, it seems that registration is still open. Just go to the ITU’s AI for Good Global Summit 2018 webpage, scroll the page down to ‘Documentation’ and you will find a link to the invitation and a link to online registration. Participation is free but I expect that you are responsible for your travel and accommodation costs.

For anyone unable to attend in person, the summit will be livestreamed (webcast in real time) and you can watch the sessions by following the link below,

https://www.itu.int/en/ITU-T/AI/2018/Pages/webcast.aspx

For those of us on the West Coast of Canada and other parts distant to Geneva, you will want to take the nine hour difference between Geneva (Switzerland) and here into account when viewing the proceedings. If you can’t manage the time difference, the sessions are being recorded and will be posted at a later date.

*’132 member states’ corrected to ‘192 member states’ on May 11, 2018 at 1500 hours PDT.

*Redundant ‘and’ removed on July 19, 2018.