Showing posts with label deepfake. Show all posts
Showing posts with label deepfake. Show all posts

Thursday, April 7, 2022

Anthromimetics - Flattening the Curve of the Uncanny Valley


Image results for anthromimetics are disappointing, so you get a beautiful Hubble picture of the universe instead, the Veil Nebula, via the ESA and NASA, 2021.

A search for anthromimetics wants me to see anthropometrics. Correcting gives me 78 results. Anthromimicry? 10 results. (Two of them are from this weblog, and 2 are from what I would call procedurally generated domain name redirects, kind of like the lists of phone numbers and why you get calls from yourself sometimes. The other sites are circa 2012.)

So I will explain - biomimicry copies biology and anthromimicry copies humans. It's what we're doing to robots to help them cross the uncanny valley. We want them to copy us good so we can believe them. 


A new machine-learning system helps robots understand and perform certain social interactions
Nov 2021, phys.org

"In a simulated environment, a robot watches its companion, guesses what task it wants to accomplish, and then helps or hinders this other robot based on its own goals."

In order to teach robots how to be social, we're just showing them footage of people helping each other, and go figure the robots gets better at figuring out what you're trying to do, and then helps you do it. Sounds simple until you realize that to be "social" is extraordinarily complex, and something that has taken most of us our entire lives to get right (and that's a lot of training data). 

via MIT's Computer Science and Artificial Intelligence Laboratory, and a member of the Center for Brains, Minds, and Machines: Ravi Tejwani et al, Social Interactions as Recursive MDPs (2021).


Making (and breaking) eye contact makes conversation more engaging
Sep 2021, phys.org

Synchrony - "In the past, it has been assumed that eye contact creates synchrony, but our findings suggest that it's not that simple," says senior author Thalia Wheatley, a professor of psychological and brain sciences at Dartmouth, and principal investigator of the Dartmouth Social Systems Laboratory. "We make eye contact when we are already in sync, and, if anything, eye contact seems to then help break that synchrony. Eye contact may usefully disrupt synchrony momentarily in order to allow for a new thought or idea."

via Dartmouth College: Sophie Wohltjen et al, Eye contact marks the rise and fall of shared attention in conversation, PNAS (2021). doi.org/10.1073/pnas.2106645118


Voice copying algorithms found able to dupe voice recognition devices
Oct 2021, phys.org

Copybots

Training data: 90 five-minute voice snippets of people talking

via University of Chicago: Emily Wenger et al, "Hello, It's Me": Deep Learning-based Speech Synthesis Attacks in the Real World. arXiv:2109.09598v1 [cs.CR], arxiv.org/abs/2109.09598


Heart rate synchronization and palm sweat found to be signs of attraction
Nov 2021, phys.org

They set up the "dating cabins" at public events such as concerts and invited young single people to participate in their study.

When two people are attracted to one another, their heart rates tend to synchronize and their palms sweat together.

The scores of attractiveness the volunteers gave each other after a first quick peek did not hold up. Nor did their behavioral clues

via Leiden University and University of Birmingham: E. Prochazkova et al, Physiological synchrony is associated with attraction in a blind date setting, Nature Human Behaviour (2021). DOI: 10.1038/s41562-021-01197-3

Hanako - aka the Uncanny Horror

Post Script:
This one can't make the rounds enough. So here you have it again. You're looking at a screenshot of a humanoid robot designed for dental students. It recoils, it gags, and it squeals in fear and discomfort. If you find the video, you will be horrified beyond repair. That's why I only give you the picture. 

Where Did That Horrifying Malfunctioning Dental School Robot Come From?
Vice, By Samantha Cole, Oct 2020

Sunday, December 8, 2019

Artificial Impressionism


Fake news via OpenAI - Eloquently incoherent?
Nov 2019, phys.org

Robots slowly taking over. Give them a sentence and they can now keep it going for a few more sentences, but after that it gets stupid.

So you can give it a fake headline, and it will generate the first line of the story, but after that things will start to fall apart.

Good thing the targets for engineered memetic propagation are not trying to read past the first sentence!

Post Script
Researchers develop a method to identify computer-generated text
July 2019, phys.org



In the above 3 images, the first is a chunk of text written by a robot (most of the words are green, with a few yellows sprinkled in), the second is a real New York Times article (only half is green, the rest is yellow, with some red, and a sprinkle of purple) and the third picture is a clip from "the most unpredictable human text ever written", James Joyce's Finnegan's Wake (the colors green, yellow, red, purple are all evenly distributed about the page).

Green words are very predictably the next word. Yellow words are less likely to show up after the word they show up after. And red and purple are for when the next word is something you absolutely did not expect.

Because text-writing algorithms today use a statistical correlation program based on a compendium of written language (so they know what words typically occur together) the output of such algos will tend to look like the topmost image with all green words. The algos can't think for themselves, they can't 'come up with' new stuff, and they can't be unpredictable. The whole point of writing an algorithm to do this is to prescribe what it's going to do in advance, i.e., it's predictable.

Anyway, soon we won't be writing our robots to write like that. They'll use less predictable programs to generate their text, with unpredictability and random association thrown in there on purpose.