That is why we do not have AI. We have learning algorithms but they only do 1 thing.
-
Join the MyBroadband community
South Africa’s biggest forum. Discuss, discover, and connect with thousands of members.
-
Question of the Day: Will you be subscribing to the new Disney+ R49 streaming package?
Artificial Intelligence Takes On Earthquake Prediction
- Thread starter Binary_Bark
- Start date
He is correct. You generally decide on the type of algorithm, and sometimes details like the network topology, tree depths, etc. for the problem. Usually it is a trade off between training time and predictability. In some situations some simpler algorithms can’t predict well enough and other algorithms end up interpolating instead of prediction (ie, overfit). The parameters are all learned by the training process, the rest of the mechanics don’t care if you’re processing images, market data, voice, etc.
It is easy to say someone is wrong. Proof it. You are a programmer. I integrate programming instruction sets with hardware. I can do your job, Programming was 1 of my subjects. Programmers can't program for hardware. ( most of them. )
LD50
Expert Member
- Joined
- Mar 25, 2019
- Messages
- 1,986
- Reaction score
- 625
Here's a good article explaining
machinelearningmastery.com
How Machine Learning Algorithms Work (they learn a mapping of input to output)
How do machine learning algorithms work? There is a common principle that underlies all supervised machine learning algorithms for predictive […]
machinelearningmastery.com
What are you talking about - what be this 3 machine fundamental rules?
You confusing science fiction with actual/true/real programming/algorithms?
Stop talking out of your a$$.
Moosedrool
Honorary Master
@Ponderer You seem to have a very aggressive emotional response to the idea of machine learning. Apart from creating programs which generate a immense complex network of instructions based on its interpretation of data received what exactly is blocking code from generating complex pieces of code? And why the emotional response?
Do you think he is in the process of self awareness?
"It is easy to say someone is wrong."
Its even easier if someone makes an a$$ of themselves.
"I integrate programming instruction sets with hardware."
What do you mean when you say that you "integrate" "instruction sets with hardware"?
"I can do your job, Programming was 1 of my subjects."
Anyone can program - some better than others.
Maybe you should have paid a bit more attention in your Programming subject/course.
In your case, ask for a refund.
"Programmers can't program for hardware."
Ask for a complete refund.
Ok Mr. programmer. Please supply me with C++ code for using the Back EMF on a 1200V BLDC motor to 5V 8Bit ADC to get the right firing sequence for the Coils. and while you are at it forward reverse also please.
Only partial AI exist. IBM trying to build a neural processor for true AI.
I'll believe it the day Google Assistant can understand me when I say "Call that guy that I gave a lift yesterday"
And then proceed to gossip about him with Google Assistant over a cup to tea.
Google and Microsoft far behind. The japs did voice recocnition many years ago in toys already.
I want to say anything and it should respond correctly in an intelligent/sane manner, otherwise it's simply not intelligent.
I spent 10 minutes with Bixby before telling her to F-Off.
Johnatan56
Honorary Master
Weak or narrow is the correct term, not partial. So far every AI system is only an expert in the field designed, one hypothesis is that we use an AI like Google assistant to interface with all other domain AI and thereby attempt to build a "complete" one.
The problem here is that everyone believes AI to be like the sci-fi movies, we haven't gotten there yet and I doubt that we would for a while yet.
All responses are emotional - just overly tired of all the BS floating around.
As to machine learning.
A machine is unable to learn or think - its the algorithm that is running that does the "thinking" and "learning".
<edit>
BTW - the "learning" part of an algorithm is part of the "thinking" of an algorithm.
Last edited:
LOL - true AI - very funny.
agentrfr
Executive Member
What passes for "AI" these days is just neural nets. Basically PID based input/output units in a web shape. Hardly intelligent... Theoretically if you had enough neurons you could simulate the operation of a brain - or at least the current understanding of how brains operate
Moosedrool
Honorary Master
If a biological brain is fundamentally based on the same governing laws of physics and chemistry. What makes you think a deeply complex neural network with the same complexities can't exist in a computer?
A few years ago I asked the question about the actual reasoning for self awareness and inevitability. When you make the decision to move you pinky finger, is this not a vast network of predetermined data and your observations at the time governing exactly what you'll do in the next few minutes?
Self awareness can in this harsh mechanical outlook be boiled down to nothing but an illusion.
You have to make the argument that a soul or some supernatural entity exists in order to think otherwise.