Hints on how A.E. Graves spends her imaginary spare time
After reading about all of the biases that can be introduced into AI systems which can have life-and-death influences on real humans in real life, I remembered reading a GREAT interview with the founder of the Algorithmic Justice League, who is helping people understand, and hopefully avoid, the dangers of blindly trusting software that may have profound flaws.
Prevent incorrectly trained software from mathwashing bad ideas!
AJL’s website is worth a visit:
You Look Like a Thing and I Love You
by Janelle Shane, Ph.D.
published by Voracious (Little, Brown and Company)
2019
I’ve enjoyed Janelle Shane’s site, aiweirdness.com, for some time, and when she mentioned that she had published a book on the same themes, I couldn’t resist it.
What are her themes? Machine learning, mostly, and how difficult it is to train a neural network to do what you really want it to do. You THINK you are training your software to recognize cancerous lumps, and it does well with your training data, but it doesn’t work so well in real life. In retrospect, you trained it with images of cancerous lumps that have rulers next to them to show the size of the lump, while no one cares (or measures) what size benign lumps are. Your program relied on the rulers to know whether or not a lump is cancerous: ruler = yes, no ruler = no. You invented… a RULER-DETECTOR.
Why I am reading about this geeky, specialist topic? I have to deal with the limitations of “AI”s of various designs all the time. Voicemail hell? That’s a not-very-intelligent program imitating an AI, possibly with AI voice recognition. Applying for a job? Software is screening my resume. Getting a laboratory test? Software may be screening that for me, too!
If you’ve ever gotten into an argument with your phone, you know that these programs are… not perfect. Depending on whether you have a high or low voice, they may not seem to work at all. My father is still amused that one of his friends couldn’t get her voice assistant on her phone to understand ANYTHING she said, but my father (who sounds like Darth Vader) could ALWAYS be understood. Why? Because it was trained this way.
Janelle Shane finds amusing ways to talk about how neural networks and other near-AI programs work, what they are good at, why they fail at so many tasks, and how the data sets they train on can make them vulnerable to manipulation.
You will laugh, as I did, as an AI trained to generate metal band names learns to generate ice cream flavors! You’ll laugh often, really: Ms. Shane has some good stories, and good quotes from people who fought to teach their AI something specific, and their AI interpreted them literally and won. The challenges she sets up for the simple neural nets she build are VERY FUNNY.
It isn’t just jokes and witty examples: you won’t laugh at the idea of a navigation-bot telling you to drive TOWARD a fire (because there is less traffic in that direction!), nor at racial and gender biases that oblivious employees train software with, nor at the fact that image recognition programs that train on the same free (manipulatable) data sets can be mis-trained to see things that aren’t visible / obvious / correct to humans.
Maybe there’s a rare but catastrophic bug that develops, like the one that affected Siri for a brief period of time, causing her to respond to users saying “Call me an ambulance” with “Okay, I’ll call you ‘an ambulance’ from now on.”
Excerpt From: Janelle Shane. “You Look Like a Thing and I Love You.” Apple Books. https://books.apple.com/us/book/you-look-like-a-thing-and-i-love-you/id1455076486
It is good (and refreshing) to truly think about the serious implications of our rush to be dependent upon machines, and the hazy way we think that machines are neutral decision makers, when nearly every application we have developed for them is not neutral in inputs, programming, or impact.
There is a lovely photo of Harpa in the book I’m reading, and that made me go looking for additional photos of it… even though I am privileged enough to have seen it in person, I know that people who were wearing better coats than I was during my brief visit likely took better photos of it with real cameras. 🙂
The photo on their website is stunning (and is not the same as in the preview below, at least at the moment), and so I’m sharing that link so you can see it yourself.
While I’m enjoying the catalog-textbook-deep-dive, I realized that the web page for the exhibit that it is based on is excellent. It is worth sharing, so I will.
It is a foggy day, and I am filled with French press coffee, vegan sandwiches, birdsong (without automobile traffic, you can hear the birds!), and things I want to research that I’m either reading about, or that came up while reading about something else.
Reading is a rabbit hole with many tunnels.
I’m writing out reviews about books I enjoy(ed) here, because writing about them is part of my ‘processing.’ Conveying feelings and thoughts in language requires a lot of translation, and that translation teaches me things. If I really enjoy something, I should be able to describe it concisely and well in words, and choosing the right words thoughtfully can take some time. (For example, it wouldn’t be enough to say that the protagonists of William Gibson’s recent fiction are “cool,” because that doesn’t convey what I mean: I mean that they are capable, curious, observant, cool-headed, and have some special niche expertise that they enjoy. I admire people like that, and want to be like that. “Cool” wouldn’t cover it. )
While I am writing reviews of recently read books in a linear fashion, I’m also various stages of reading other books. In my youth, I would force myself to finish each book (whether or not I was enjoying it) before I could go onto the next; but now I drop any I’m not in the mood for (I have so little time – why spend it on a book that isn’t working for me?), and to jump in and out of longer books, based on where they are taking me. Truly stimulating books send me on research projects or reveries that take me a while to return from!
I am in various stages of reading:
My reading of the Eliasson book today has made me pause to:
It is FUN to let my mind wander so freely, and to have such stimulating material to read.
Working long hours prevents me from spending more time thinking like this – while I am working, I am VERY dedicated to solving work problems, and sometimes dream about possible solutions, or wake up dwelling on work issues to solve – but a weekend like this is like a brain vacation / day spa / slice of free-thinking heaven.
Humor is what I’m using to get through this situation, because I’d prefer to laugh than to dwell on the lack of understanding across segments of my country of basic survival concepts.
Opinion | America, please don’t put bleach inside yourself like the president says
This should go without saying, but: Please, America, don’t inject or put disinfectant inside yourself like the president talked about. It is not good for you. Actually, it is bad. “Sunlight is the best disinfectant” is just a metaphorical saying. Also, do not put a “light” “inside” yourself until we can figure out what the president is trying to say.
I think my favorite of many good quotes is:
I know it seems counterintuitive, but given the choice to listen to a plastic bottle or the president of the United States, I beg you: Listen to the plastic bottle.
Alexandra Petri’s writing for WaPo is topical and very, very funny. (She does Twitter well, too!)
During this ultra-serious time, this news item made me laugh out loud, so I want to share it.
We regret to inform you that your inflatable T. rex costume is not virus-proof
Earlier this month, Felix Brodie parked his car in front of the Vons supermarket in Rancho Cucamonga, Calif., and suited up: full-length black cloak, balaclava, black leather gloves, and the final touch – a reproduction of a 17th-century beaked plague doctor’s mask (complete with top hat) that he bought on Etsy.
Mail / Art / Book
curated by Jennie Hinchcliff and John Held, Jr.
published by San Francisco Center for the Book
2014
This is another stellar catalog from the SFCB’s shows, and from longstanding teacher and author Jennie Hinchcliff (co-author of the fun book, Good Mail Day).
Mail art – art that is specifically intended to be sent through the mail, and enjoyed by other mail artists – is small and can be difficult to display, but this book does a solid job of arranging and photographing this exhibit in a way that preserves the many 3-D items, as well as the relationships between envelope and contents (when there is a distinction).
The submissions were fun and varied widely, and it was great to see all the complex, thoughtful, colorful, creating works that made it through the mail!
Calligraphies In Conversation, 6th Annual Exhibition
curated by Arash Shirinbab
published by San Francisco Center for the Book and Ziya Art Center
2019
This is a beautiful, fully illustrated catalog of an exhibit of calligraphic writing from multiple traditions, and it is really gorgeous. Work from fourteen artists shows a lovely stylistic and creative range. I had been expecting Chinese calligraphy for its local (SF Bay Area) popularity and long tradition, plus some western-language calligraphy, and was delighted to see those PLUS work in Urdu, Hebrew, Arabic, and more. My favorite piece is in a Devanagari / Sanskrit script over gorgeous shades of blue – the composition and color are WONDERFUL.
This is a beautiful and inspiring catalog.
I haven’t mentioned that the past few months have been all about obsessively reading the news. Having a highly contagious global pandemic break out, one so severe that China closed internal borders and quarantined millions of people during its early peak, is A BIG DEAL – it’s a worldwide concern that everyone sensible wants to know something about.
Once it reached my country (& my coast!), and the voluntary precautions kicked in, my news-reading increased further. And once the MANDATORY precautions took effect, I could devote time I used to spend commuting, enjoying the outdoors, or running errands ENTIRELY to news reading.
Which isn’t entirely healthy: no one really wants to see a global death count on the front page of their device the moment their alarms go off. But that appears to be what it takes to get some people to take this seriously. (Though I suspect the people who aren’t taking it seriously DO NOT READ, which would explain many things.)
With infection counts and deaths rising, and routine business activities temporarily halted, the economy has been upended, and states – which rely on business running for revenue – have been struggling. And then things got weird, because a bunch of senators starting talking about having US States go bankrupt. Which… is not a thing US States DO. Also, I couldn’t see the point. I was missing something.
This article in the Atlantic explains what I was missing:
The gist: States would give up their sovereignty over their finances if they declared bankruptcy, and hand control over to the current federal legislators, who are majority Republican in the Senate; the wealthy blue states could come under the control of senators from the least wealthy red states, and have their larger budgets/populations forced to match red state priorities.
Ohhhhhhhhhh! NOW it makes sense! It is not at all democratic, and that is the way red state senators (many of whom are funded by business interests beyond their own borders, and remain in power through gerrymandering and voter suppression) prefer things. Which is terrible, but never surprising.
This is a good, clear read.
