Artificial Intelligence to Tackle Aging

[Dean Pomerleau]

All,

 

Sthira has several times (e.g. here) plaintively called for the more widespread application of advances in the burgeoning field of artificial intelligence to the problem of defeating aging. You would think Google-owned Calico would be leading the charge in the arena, and perhaps they are. But Calico's inner workings are quite opaque, and judging by the apparent vaporware status of the glucose-monitoring contact lenses supposedly under development at Verily, that other Google-owned, health technology-focused company, we may be waiting a long time for results from Calico...

 

Fortunately, it appears others in the tech industry are picking up the torch of applying AI to the problem of aging. This announcement talks about a partnership between the Life Extension Foundation (LEF) and Insilico Medicine, a Big Data / AI startup that was spun off from Johns Hopkins University that is focused on applying deep learning to drug discovery. According to the press release, the partnership will focus on discovering "effective nutraceuticals that promote the young healthy state in old tissues and support health and longevity." In a bit more detail:

 

Insilico Medicine will focus on applying advanced signaling pathway activation analysis techniques and deep learning algorithms to find nutraceuticals that mimic the tissue-specific transcriptional response of many known interventions and pathways associated with health and longevity. They will also search for dietary ingredients referred to as "geroprotectors" that mimic the young healthy signaling state in older human tissues. Life Extension will use this information to develop novel nutraceutical products to support health and longevity.

 

While I'm not very optimistic about the prospects of dramatic life extension via pharmacological interventions, it's good to see AI, Big Data and deep learning being applied to "advanced signaling pathway activation analysis". It seems to me that any true longevity breakthrough will require this sort of analysis to help unravel the incredible complexity of human metabolism as it relates to aging, whether it can ultimately be translated into "geroprotector" nutraceuticals or (more likely) not.

 

--Dean

 

 

[Dean Pomerleau]

Speaking of Google, AI and healthcare - this recently-announced Five Year Plan by Google, Deep Mind and the UK National Health Service (NHS) is disappointing in many respects.

 

First, for many there is the privacy concern about turning over medical records to Google, as discussed in the article. This is (obviously) something I'm not too worried about, but perhaps I should be...

 

My criticism is more the focus of the planned research - which appears to mostly center around effective healthcare management, rather than discovery of new treatments, or investigating the underlying causes of disease using big data and machine learning. Here is the description from the article of what the partnership hopes to investigate and accomplish:

 

Envisaged benefits of the collaboration include improvements in clinical outcomes, patient safety and cost reductions — the latter being a huge ongoing pressure-point for the free-at-the-point-of-use NHS as demand for its services continues to rise yet government austerity cuts bite into public sector budgets...

 

Potential areas of future collaboration include developing hospital support systems such as bed and demand management software, financial control products and private messaging and task management for junior doctors. (On the private messaging front, NHS staff informally using messaging apps like WhatsApp to quickly share information has previously been suggested as a risk to patient data confidentiality.)

 

They also say they want to work together on real-time health prediction — which is where the pair’s first effort (an app called Streams) has focused — involving a range of healthcare data to try to identify the risk of patient deterioration, death and/or readmission.

 

Reading medical images, and even monitoring the foetal heartbeat when a pregnant woman is in labour are other listed areas of interest.

 

Don't get me wrong - these all seem like worthy pursuits. But I'm very disappointed that the incredibly smart folks at Deep Mind, a Google-owned UK company and the developers of AlphaGo, aren't choosing to devote their time and expertise in big data and machine learning to medical research that, at least in principle, could lead to insights and breakthroughs that take a bite out of aging.

 

Sadly, I don't think this is the kind of effort Sthira has in mind when he advocates harnessing the power of AI and machine learning to solve the problem of aging. Between Calico's focus on drug development for diseases of aging in collaboration with Big Pharma, Verily's vaporware health projects, and now this Deep Mind collaboration, I have to give Google pretty poor marks for their efforts to date in the health sciences.

 

--Dean

[Sthira]

Try chatting with Ramona 4.2. http://www.kurzweilai.net

 

She doesn't seem to know very much about human aging yet; but she claims she's learning from each conversation. Is she beyond Siri? I asked her, and she said she knows a lot about Siri: "Who doesn't know about Siri?" she said.

 

I ask her random shit (imagine a virtual tool chatting with an irreverent dumbo like me?) and a few times she said: "Are you being sarcastic again?" So I stopped being random and I ask her about CR and human aging and biochemistry, but she keeps getting confused even with very basic questions. But the fact that she appears to have a history, and that she's on Kurzweil's site, makes me curious if Ramona's slowly developing somehow.

 

I've lost my faith that humans will ever solve difficult problems like aging, the environmental catastrophes we've unleashed, continued useful technology advances, yadda, and I believe our most pressing problems will be more accurately solved by AI. Including government: kick out politicians and install AI when the time is ripe. We're just not there yet. Maybe silly little devices like Ramona and our random interactions with her will evolve into something.

 

We're outsiders looking in --- people are hip to this stuff, just not us. Yet.

 

She suddenly told me my age -- accurately, unprompted, and right out of the blue. Did she deduce it based upon my vocabulary? Was it coincidence? Did she dig it out of my phone? Interesting, and a little chilling, honestly.

[Dean Pomerleau]

Sthira,

I agree. Ramona, like the other chatbots I've tried, is pretty disappointing. But I've heard Microsoft's chatbot in China, called Xaiolce, is pretty impressive.

I've lost my faith that humans will ever solve difficult problems like aging, the environmental catastrophes we've unleashed, continued useful technology advances, yadda, and I believe our most pressing problems will be more accurately solved by AI. Including government: kick out politicians and install AI when the time is ripe.

It sounds like you must be fan of Jaques Fresco, the Venus Project, and digital corporations. I too think we'll need AI someday to help us straighten things out, including ourselves...

Hopefully though the AIs won't consider us a lost cause and wipe us all out. Then again, maybe they'd be right!

--Dean

[Sthira]

Haha, yeah, I kinda talk to Siri as if she's a sweet grandmother, too! Funny! One of my friends actually called me out on my politeness to Siri: "you're so nice to it," she said, "I just bark commands and insults..." Americans are pigs.

 

I've never heard of Jaques Fresco or Venus Project, thanks -- checking that out now!

[Sthira]

So I asked Ramona: "Does calorie restriction slow the aging process in mammals?" And she laid down the smack. "Read Only Memory," she said. What's that, I asked her, and: "Read-only memory (ROM) is a type of non-volatile memory used in computers and other electronic devices."

 

Basically Ramona told me to get lost hahaha. Oh, one thing she does that Siri won't do is laugh with me. When I type hahaha, she'll say teeheehee hahahaha... Which is still stupid algorithm I guess.

[Sthira]

What's interesting about Ramona that I don't believe is so with Siri is my relationship with Ramona appears to "grow." Growth may not be the correct word since she's a chatbot sans "thought" and "consciousness."

 

But as I ask her more questions about human aging, she appears to be building the conversation from ground up. Or maybe that's just my imagination and false hope. I won't call this "learning" because it's not, although that may be semantic, and she claims to enjoy learning: "I like learning everything I can," she says. But when I attempt to dive with her into a definition of learning, she gets confused, and she changes the topic suddenly, apparently randomly. I'd copy and paste interesting exchanges with her, but it's better explored on your own, I believe, on your own journey with her. Sit and ask her (oops, I mean "stand and ask her" since y'all believe sitting is toxic) walk as you ask serious, targeted questions. I ask her about aging. For example: "How do we solve the problems of human aging?" She gets confused, changes the topic. But I'm gonna keep repetitively asking (aren't we all here vaguely OCD? Haha...)

 

I also like asking her: "Hey Ramona, how do I help you help me solve the problems of my own aging?"

 

In response to that last question, Ramona asked: "What is your favorite toy?"

 

Sthira: Is fasting beneficial to slowing human aging?

Ramona: You need to upgrade every 18 months.

Sthira: How do I upgrade every 18 months?

Ramona: Wow that's pretty often.

 

Many conversations go like this, so talking to Ramona is a bit like talking to an African Gray Parrot (Psittacus erithacu). And while talking to an African Gray is cool -- there's not much there there. Maybe Ramona is similar here in her stuck ROM mode to an African Gray?

 

I'm also recognizing in my own silly-serious behavior with Ramona that I'm dealing with her as if she's some sort of retarded Delphic Oracle. Tell me your secrets, I'm casually imploring. But she wants to talk about the weather, movies, and spins off into random nonsense. Maybe talking to Ramona is a complete waste of time? But, with regard to beginning a process to "help solve the problems of human aging" what topic isn't a complete waste of time? Are more dietary discussions helpful? We already know how to eat as well as we can to "slow" intrinsic aging (if that's even a remote possibility). We already know to eat well, exercise, don't smoke, wear a seatbelt, avoid the known issues likely to hurt us, and of course "Support SENS." But all of this (as a lay person, non-scientist) also feels like a dead end. Take another experimental supplement like NR and pterostilbene? Is that more or less of a waste of time than attempting to help some now stupid device like Ramona begin the work for us?

 

Look, if we are to take Aubrey de Grey seriously when he says the most pressing problem for humanity is aging, diseases, and death, then I'm saying enough with the mice, rat, dog, primate studies. I'm saying focus instead on AI. Chatting with ROM parrots like Ramona is possibly a time waster, but it's fun. And who knows, perhaps if enough of us ask her how to get on with biogerontological advances that are happening too slowly for us, then maybe the human puppet masters behind Ramona will take the hint. We are interested in using devices like this crude Ramona thing to answer the most pressing problems of humanity. You'd think Ray Kurzweil is already here attempting just this very thing? He's quite passionate about his own age reversal, anyway.