Friday, November 27, 2020

Global AI Narratives Latin America workshop

 

This morning I attended the Zoom meeting of the first Global AI Narratives Latin America workshop, which was broadcast from Santiago, Chile. It's the first of three virtual workshops hosted by The Leverhulme Centre for the Future of Intelligence (University of Cambridge) and the The Pontificia Universidad Católica de Chile.

This workshop was entitled "AI at the edges of Latin America," and covered AI as represented in literature, art and other media.

It took me a bit to join this meeting, since I first tried YouTube, where I didn't see the translation function, so I joined the Zoom meeting where it worked fine.

I did catch the opening statements by Stephen Cave (Executive Director of the Leverhulme Centre for the Future of Intelligence), who mentioned this workshop would explore AI as more than merely an engineering problem.

Next up was Kanta Dihal (Leverhulme Senior Research Fellow) who introduced the AI Narratives Project, "a joint endeavour by the Leverhulme Centre and the Royal Society." You can download their report "Portrayals and perceptions of AI and why they matter." (.pdf) Among the topics they were examining include the racialization of robots, published as an article "The Whiteness of AI."

The first speaker was writer Alberto Chimal with a talk entitled “Captured Intelligences.”

He said that speculative science fiction in Latin America tends to frame people as users or even victims of technology rather than originators. For me, what immediately came to mind was the work of Jorge Luis Borges and I wondered how he would have handled AI as part of the magic realist movement.

He mentioned the missteps he's observed, including one Black Mirror episode "Be Right Back" where a young wife has her dead husband brought back to "life" in a android simulation. (Interesting to note, that one of the most popular episodes of this series, was the optimistic and positive emotional take on transhumanism called "San Junipero." I've often heard that the hardest thing to do in fiction, I think it was a quote from short story writer Tobias Wolff, is to make goodness interesting, which might explain the paucity of stories about beneficial AI. Iain M. Banks, in his Cultural series of sci-fi stories, has AI minds that work well in governing the post-scarcity, quasi-utopian world he's created.)

A version of this was actually attempted: "South Korean TV broadcaster MBC recently aired a Korean language documentary that centers on a family’s loss of their young daughter, seven-year-old Nayeon. Using the power of photogrammetry, motion capture, and virtual reality, the team recreated Nayeon for one last goodbye with the family’s mother, Ji-sung."

Literature pieces mentioned by Alberto: Los superjuguetes duran todo el verano (Brian W. Aldiss), 2001: A Space Odyssey (Arthur C. Clarke), El estruendo del silencio (Bernardo Fernández), Kentukis (Samanta Schweblin), La segunda celeste en "Manos de lumbre" (Alberto Chimal)


In his talk "AI in the art of Latin America," Raul Cruz, (his painting served as the poster for this workshop), said that images about AI should not be seen as just supplements for other media such as book covers. His works can be as large as 1.5 by 2 meters, and can elicit a response by the viewers approaching them from a distance in a gallery. His art is quite striking along with a distressed metallic palette combined with cultural and mythological figures.

He offered a visual proposal where you might imagine a parallel timeline where Mayan kings still existed but had access to advanced technology.


Soledad Véliz of the Pontificia Universidad Católica de Chile, finished with “Good night stories for Artificial Intelligence.” She asked us to view AI as a child that needs to be train in moral values by responsible adults. She cited a paper by Riedl & Harrison to show how this might be done by "Using Stories to Teach Human Values to Artificial Agents." (.pdf) Also mentioned as an example was "The Goldilocks Principle" (Hill, Bordes, Chopra & Weston, 2016).


I have a similar inter-disciplinary approach to the study of AI. Before medical school my background was in graphic arts and creative writing, and I see these experiences as a complement to my scientific training.

Developing the creative narratives, mythologies you might consider them, will be essential to make the public aware of the expansive nature of AI, and not just the realm of red-eyed, killer robots.

 





Sunday, November 15, 2020

AI News & Trends, Nov. 14, 2020

 Health Care

Wired: The US Government Will Pay Doctors to Use These AI Algorithms: The artificial intelligence programs can diagnose eye disease in diabetics and complications in stroke patients

Viz.ai has a program that will diagnosis hemorrhaging of the brain on CT scan.  Digital Diagnostics has software that helps diagnosed diabetic retinopathy.

Google Cloud: Healthcare gets more productive with new industry-specific AI tools


Companion Androids

Psychology Today: The Age of Sex Robots: the pros and cons in this emerging sexual age

A "a potentially satisfying relationship with a robotic partner," or for couples "an act of infidelity"?


Automated Vehicles

Wired: Split-Second ‘Phantom’ Images Can Fool Tesla’s Autopilot: researchers found they could stop a Tesla by flashing a few frames of a stop sign for less than half a second on an internet-connected billboard


Deep Learning

ZDNet: What's next for AI: Gary Marcus talks about the journey toward robust artificial intelligence

ZDNet: C3.ai, machine learning startup backed by software pioneer Tom Siebel, files for IPO: the software-as-a-service company that has been using masses of GPUs to run deep learning programs plans to list under the ticker “AI.”


Military

AXIOS: The military is calling in AI for support

This week, Armenia accepted a cease-fire with its neighbor Azerbaijan over Nagorno-Karabakh. "Azerbaijan dominated the conflict in part thanks to the ability of its fleets of cheap, armed drones to destroy Armenia's tanks."

Anduril's website describes its military drone: "Ghost 4 is an autonomous VTOL [vertical take-off and landing] sUAS [small unmanned aircraft system, typically under 55 lbs.] that operates on the Lattice AI platform. Ghost 4 is modular, man-portable, waterproof, and combines long endurance, high payload capacity and a near-silent acoustic signature for a wide variety of mission capabilities."
Axios: Anduril is a "young Silicon Valley startup backed by Peter Thiel and co-founded by Palmer Luckey that focuses on defense — showed off its Lattice software system, which processes sensor data through machine-learning algorithms to automatically identify and track targets like an incoming cruise missile."

Work Automation/Ethics

Wired: 6 Sci-Fi Writers Imagine the Beguiling, Troubling Future of WorkThe Future of Work: ‘Work Ethics,’ by Yudhanjaya Wijeratne


Tuesday, November 10, 2020

News & Trends: Cognitive Science and AI

Health Care

WSJ: AI May Help Identify Patients With Early-Stage Dementia: Research suggests that how we type, sleep and speak may offer clues of cognitive decline

"Much of the current machine-learning research is focused on sifting through patients’ electronic health records to determine what combination of risk factors most accurately reflects cognitive decline."

J Am Geriatr Soc. 2020 Jan;68(1):103-111. doi: 10.1111/jgs.16182. Epub 2019 Oct 14. Development and Validation of eRADAR: A Tool Using EHR Data to Detect Unrecognized Dementia

npj Digital Medicine volume 1, Article number: 8 (2018)  Detecting neurodegenerative disorders from web search signals


Deep Learning

MIT Technology Review: AI pioneer Geoff Hinton: “Deep learning is going to be able to do everything”: Thirty years ago, Hinton’s belief in neural networks was contrarian. Now it’s hard to find anyone who disagrees, he says.


Covid-19

WSJ: Covid-19 Pandemic Highlights AI’s Short-Term Faults, Long-Term Potential: The pandemic has proven to be a reality check to the hype around AI


Workforce

WaPo: Biden Needs to Keep an Eye on Jobs — in China

"Over the next decade, more than 20% of China’s workforce is expected to be re-employed in high-end manufacturing. Workers will need clearer direction in this new environment on jobs, incomes and benefits. Already, state media is publicizing a host of new jobs like artificial intelligence trainers as officially recognized occupations that can bring big increases in pay."


Axios: AI and automation are creating a hybrid workforce


Existential Threats

NYTimes: Elon Musk: ‘A.I. Doesn’t Need to Hate Us to Destroy Us" interview with Kara Swisher

Kara Swisher: Right. So many years ago when we met you said A.I. would treat us like house cats. That they’re too smart to hate us. And you said, we’ll be like house cats. That’s how they think of us. And then later, when I met with you at your office, you switched it to anthills, which was your analogy that when you see an anthill you don’t kick it over unless you’re kind of a jerk. But when you’re making a highway you just roll over it. Can you give a metaphor of where we are with AI right now?

Elon Musk: I was just pointing out with the anthill analogy that A.I. does not need to hate us to destroy us. In a sense, that if it decides that it needs to go in a particular direction and we’re in the way then it would without no hard feelings it would just roll over us. We would roll over an anthill that’s in the way of a road. You don’t hate ants. You’re just building a road. It’s a risk not a prediction. So, yeah. I think that we really need to think of intelligence as really not being uniquely confined to humans. And that the potential for intelligence in computers is far greater than in biology. Just far, far greater. There’s a great, quite a funny, essay called, I think it’s called, “They’re Made of Meat“. Which, like if some super advanced civilization coming across Earth and they’re obviously all computers and they just can’t believe we’re made of meat.


Military

DefenseOne: Pentagon's 'AI Factory' Looks To Distribute Tools Across DOD

"The Joint Artificial Intelligence Center is looking to become an AI service hub for the Defense Department rather than a product building unit, according to the JAIC’s new director."

In August Deloitte Consulting was awarded a $106 million contract to design and build the Joint Common Foundation (JCF). The JCF will "operate as a kind of factory" providing DOD with “an AI development environment to test, validate and field AI capabilities.” 



Saturday, September 19, 2020

Heartland Developers Conference: AI Presentations

 The Heartland Developers Conference is a virtual event organized by the AIM Institute of Omaha, and presented by Microsoft starting at 9am, CDT on September 24, 2020. These are the AI presentations:


Evan Hennis, MS | 10 - 10:45 AM
'Computer Vision with Raspberry Pi'

Sr. Software Engineer, Eckron Software


Gordon Summers | 11 - 11:45 AM
'Train your AI to not be Racist'

Principal, Cabri Group


Atif Farid Mohammad | 2:15 - 3:00 PM
'Use of AI in Healthcare'

AI Professor, UNCC | Sr. Research Scientist, OnTrak Inc.


Joshua Armah | 3:15 - 4:00 PM
'Cognitive AI for Social Impact'

President, Wiase Research Institute

Approaching AI from different dimensions

When I began to examine topics in Artificial Intelligence, I realized that a higher level approach was needed, namely the broader category of cognitive science. Primarily, I'm talking about human cognition as a starting point to further explore how this will translate into machine intelligence.

Here's the outline I've constructed for myself:

COGNITIVE SCIENCE

Anthropology - man's evolution as a sentient species and the mythologies that have emerged. Further progress may be embraced with techniques for transhumanism

Art and the Humanities - the personal and creative aspect that we tend to think is uniquely human but is being challenged by AI minds that can write and create images, to an extent that needs to be explored

Artificial Intelligence - with the advancements into machine learning and deep learning (neural networks)

Linguistics - involved with natural language processing

Philosophy - ethical considerations

Psychology - it may not be restricted to the human experience when significantly advanced AI bots are realized

Neurobiology - wet ware and how it reveals the workings of the brain

Sociology - this encompasses how life will change especially the working life of individual who will be replaced by automation. One important topic in economics is the use of a Universal Basic Income