Quotes from the book "AI for thinking humans - Melanie Mitchelle"

Here are some interesting things I read and learned from the book "Artificial Intelligence for Thinking Humans" by Melanie Mitchelle.

Solve intelligence and solve everything else with that. 

Neural networks are more important than the atom bomb. 

Writing rules is difficult, we embed commonsense which is hard to code. 

A pile of narrow intelligence will never achieve general intelligence. Integration is key. 

When a machine feels things, it's conscious.  Simulation of thinking and thinking are not the same.  

 

We should be more worried about machine stupidity than Machine Intelligence, Where machines are trusted with something they don't fully understand.  

When we will have AGI? Take your answer double it, Triple it, and quadruple it, that might be the answer. 

We humans overestimate AI and underestimate our own knowledge.  

AI labeled men standing in the kitchen as women, as there were more images of women in the kitchen, and that's how the model Is trained. 

AI goes in 2 phases since 1955:

• AI Spring: Innovation excites, attracts funding, and hypes up expectations. 

• AI Winter: Hypes bust.

Explainable AI, transparent AI is the way to go. EU passed a law that the use of AI should be explainable. Hope it won't be as confusing as our Explanation of benefits statement from your insurance provider.

Long Bets is a website that hosts a platform for making long-term predictions about the future. On Long Bets, individuals can propose "Long Bets" by making predictions about events that they believe will occur in the distant future. These predictions are then challenged by others who disagree, creating a public record of the debate.  The website serves as a forum for intellectual discourse and encourages participants to think critically about the future and the potential implications of current trends and decisions. It's an interesting platform for exploring foresight, risk assessment, and the long-term consequences of contemporary actions. 

Singularity University is a Silicon Valley-based think tank and educational institution started by Ray Kurzweil that focuses on exponential technologies and their potential to address humanity's grand challenges.

Bongard puzzles are a type of cognitive puzzle that involves identifying patterns and rules based on sets of examples. They are often used in AI and cognitive science research to test pattern recognition and problem-solving abilities. Check them on Google, you will enjoy it.

Symbolic AI relies on logic and symbols, perceptrons are basic units of early neural networks, and neural networks are sophisticated models inspired by the human brain, capable of learning complex patterns from data.

Imagenet, alexnet , CNNs, and Amazon mechanical Turks:

• ImageNet: ImageNet is a large-scale dataset of labeled images that has been instrumental in advancing the field of computer vision. It contains millions of images across thousands of categories or classes. ImageNet has served as a benchmark dataset for evaluating the performance of image classification algorithms and models.

• AlexNet: AlexNet is a convolutional neural network architecture designed by Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton. It won the ImageNet Large Scale Visual Recognition Challenge (ILSVRC) in 2012 by a significant margin, marking a milestone in the use of deep learning for image classification tasks. AlexNet consists of eight layers, including five convolutional layers followed by max-pooling layers, and three fully connected layers.

• Convolutional Neural Networks (CNNs): CNNs are a class of deep neural networks particularly well-suited for analyzing visual imagery. They are composed of multiple layers of neurons that have learnable weights and biases. CNNs use a specialized architecture that takes advantage of the spatial structure of the input data. Convolutional layers apply filters to the input data, capturing different features at various spatial locations.

• Amazon Mechanical Turk (MTurk) is a crowdsourcing marketplace that enables individuals and businesses (known as "requesters") to outsource tasks called Human Intelligence Tasks (HITs) to a large number of workers (known as "turkers") over the internet.

Happy Learning !!!

P.S.: I utilized AI assistance in crafting this blog.