Post Reply  
 
 
Thread Tools Search this Thread
22-01-2021, 19:24   #436
Fathom
Moderator
 
Fathom's Avatar
Quote:
Originally Posted by Smiles35 View Post
Gently nudging away investors. Then again, surely we are coming close to a company that will start producing robots for the home.
Watched old "Runaway" film. Reminded me of "robots for the home."
Fathom is offline  
Advertisement
22-01-2021, 20:31   #437
Capt'n Midnight
00:00
 
Capt'n Midnight's Avatar
Quote:
Originally Posted by Fathom View Post
Are the continued development and proliferation of lethal autonomous weapons inevitable?
yes

UAV's means no pilots to risk. And no witnesses in the sky.

The F35 may even be the last major manned fighter program for the USAF

You can jam radio links but AI means no link.
Capt'n Midnight is offline  
Thanks from:
04-02-2021, 22:29   #438
Fathom
Moderator
 
Fathom's Avatar
Sophia. Today. Tomorrow?
Fathom is offline  
09-02-2021, 21:16   #439
Fathom
Moderator
 
Fathom's Avatar
Robot competitions.
Fathom is offline  
10-02-2021, 22:10   #440
Fathom
Moderator
 
Fathom's Avatar
Killer Robots: Survey shows opposition remains strong (Feb 2, 2021).
Fathom is offline  
Advertisement
11-02-2021, 00:24   #441
Capt'n Midnight
00:00
 
Capt'n Midnight's Avatar
Quote:
Originally Posted by Fathom View Post
Killer Robots: Survey shows opposition remains strong (Feb 2, 2021).
Quote:
Notably, a majority opposed killer robots in five countries most active in the development and testing of weapons systems with decreasing levels of human control: Russia (58 percent), UK (56 percent), US (55 percent), China (53 percent) and Israel (53 percent).
Hmmm, not so sure I trust an article that doesn't mention the actual killer robots already deployed, the sentry guns facing north on the Korean border.
Capt'n Midnight is offline  
Thanks from:
03-03-2021, 21:50   #442
Fathom
Moderator
 
Fathom's Avatar
Quote:
Originally Posted by Capt'n Midnight View Post
Hmmm, not so sure I trust an article that doesn't mention the actual killer robots already deployed, the sentry guns facing north on the Korean border.
Good point.
Fathom is offline  
06-03-2021, 22:40   #443
Capt'n Midnight
00:00
 
Capt'n Midnight's Avatar
Killer Robots mean AI

Today's AI is just like 1960's AI. It's just database lookups, pattern matching and rules. The databases are way bigger now.

Garbage in , garbage out dates back to at least 1864.


Some comments on AI

Quote:
It may be true that AI learns from every single case it sees, the problem is that you don't know what it learned, can't ask it and can't correct it. No doubt it will some times get the right answer, but that just makes it Accidental Intelligence.


...

Learns?

That's marketing speak. No AI is capable of learning at all. It's pattern recognition aided by human curated data and human initial labelling.

But indeed, you don't really know if it's matching (data comparisons) for what you intended or some other feature in the images. We see patterns, familiar objects in clouds, toast, flames in the fire, scattered stones. Because once we understand chair, we can decide to use a crate as a chair. A child that has eaten bread and sausages will assume a sausage-in-a-bun or a hot dog is edible. A two year old can do things easily that are impossible for AI.

It's called the AI paradox and it was known nearly 60 years ago.

Expert systems were the big thing in AI in the 1980s because they used text. The problem was capturing the expert. Faster cpus, bigger databases and more RAM simply made actually simpler so called image recognition possible. There is no recognition. Just matching.

It's all marketing. None use "machine learning" or "neural networks" as those don't mean what they mean outside of AI marketing.

Even machine translation has gone backwards. It now uses a brute force approach like a giant Rosetta stone and matching phrases and words.

Text to speech isn't much better than nearly 40 years ago and so called smart agents are just voice to text front ends using pattern matching to search engines and chat bots hardly better than Eliza or ALICE. Speech recognition has moved from being a program on your car radio, phone or PC to something creepy running on a 3rd party system, the so called cloud. That's a backward step in privacy and needs the Internet.

...
Anti-Tank Dog
Another serious training mistake was revealed later; the Soviets used their own diesel engine tanks to train the dogs rather than German tanks which had gasoline engines. As the dogs relied on their acute sense of smell, the dogs sought out familiar Soviet tanks instead of strange-smelling German tanks.
Capt'n Midnight is offline  
10-03-2021, 20:37   #444
Fathom
Moderator
 
Fathom's Avatar
Quote:
Originally Posted by Capt'n Midnight View Post
Killer Robots mean AI

Today's AI is just like 1960's AI. It's just database lookups, pattern matching and rules. The databases are way bigger now.
Big data is the rage today. But not new. Demographers have been collecting and crunching it for decades.
Fathom is offline  
Advertisement
16-03-2021, 10:54   #445
smacl
Moderator
 
smacl's Avatar
 
Join Date: Feb 2006
Posts: 14,624
Quote:
Originally Posted by Capt'n Midnight View Post
No AI is capable of learning at all. It's pattern recognition aided by human curated data and human initial labelling.
I think that statement is quite a bit out of date at this point. Generative adversarial networks for example don't need much in the way of training data and are capable of learning new patterns. AIs at this point can create new patterns which in turn can be used to synthesise training for other AIs.
smacl is offline  
16-03-2021, 12:30   #446
Capt'n Midnight
00:00
 
Capt'n Midnight's Avatar
Quote:
Originally Posted by smacl View Post
I think that statement is quite a bit out of date at this point. Generative adversarial networks for example don't need much in the way of training data and are capable of learning new patterns. AIs at this point can create new patterns which in turn can be used to synthesise training for other AIs.
Garbage in , garbage out.

And that includes the coding and assumptions.

Stuff like this keeps happening.
https://www.theregister.com/2021/03/...riting_attack/

Computer recognition is the only thing protecting civilians from autonomous drones like the commercially available Prowler featured in the 1985 Chuck Norris film Code Of Silence. Dumb as brick. But it had mobility, thermal sensors and machine guns.
Capt'n Midnight is offline  
16-03-2021, 12:59   #447
Capt'n Midnight
00:00
 
Capt'n Midnight's Avatar
During WWII there was project about training pigeons. to guide bombs.
Quote:
A pigeon in each compartment, trained by operant conditioning to recognise the target, would peck at it continually. Pecks to the centre of the screen caused the missile to fly straight, whilst off- centre pecks tilted the screen which would alter the missile’s course.
There's an apocryphal story story about training them.
They had a set of target photos and a set of photos without targets. Under different lighting.

And AI researchers are still making this basic mistake. Like the photos of fishermen holding tench. The AI learnt to recognise the most distinct feature - human fingers.
Capt'n Midnight is offline  
23-03-2021, 19:11   #448
Fathom
Moderator
 
Fathom's Avatar
Quote:
Originally Posted by Capt'n Midnight View Post
During WWII there was project about training pigeons. to guide bombs.
Maybe try the Skinner pigeons that played ping pong? Keep your eye on the ball bird. Trained with an intermittent reinforcement schedule.
Fathom is offline  
Thanks from:
02-04-2021, 11:02   #449
Capt'n Midnight
00:00
 
Capt'n Midnight's Avatar
More AI garbage in garbage out .

https://www.theregister.com/2021/04/01/mit_ai_accuracy/
Quote:
The error rates vary across the datasets. In ImageNet, the most popular dataset used to train models for object recognition, the rate creeps up to six per cent.
Capt'n Midnight is offline  
Thanks from:
02-04-2021, 20:04   #450
Fathom
Moderator
 
Fathom's Avatar
Article suggests cleaning data; i.e., "pruning noisy data, counting with probabilistic thresholds to estimate noise..." What's new when prepping big data? Not sure The Register "labeled" the study correctly (excuse the pun) in its headlines.

Took a quick look at abstract for Curtis G. Northcutt, Lu Jiang, Isaac L. Chuang (15 Feb 2021). Confident Learning: Estimating Uncertainty in Dataset Labels. Under Review by Journal of Artificial Intelligence Research. Perhaps I should go beyond the abstract? Then again, AI is not my area of research, so I would probably just stumble about.
Fathom is offline  
Post Reply

Quick Reply
Message:
Remove Text Formatting
Bold
Italic
Underline

Insert Image
Wrap [QUOTE] tags around selected text
 
Decrease Size
Increase Size
Please sign up or log in to join the discussion

Thread Tools Search this Thread
Search this Thread:

Advanced Search



Share Tweet