AI/ML Everywhere: A Beginner’s Guide

Source: https://unsplash.com/s/photos/ai-and-machine-learning

Let’s take a break from the buzz around LLMs and Diffusion Models and refocus our attention on the fundamental principles that have made this remarkable innovation possible. AI/ML tools have never been more accessible, as people from all walks of life are utilizing these tools. These technologies, which would have felt like magic only a few decades ago, have become so common and unsurprising nowadays. This fact encourages us to take a step back and analyze how we got here in the first place and what lies ahead in the future.

The term “Artificial Intelligence” was coined in 1956 by John McCarthy for machines “capable of doing tasks that require human intelligence”. One of these first AI programs was Expert Systems which tried to emulate the decision-making process of humans. These systems quite literally used if-then rules to make decisions. Yes, it’s not just a meme.

https://www.reddit.com/r/ProgrammerHumor/comments/15t7vie/youknowitstrue/

So now if you want the machine to do some task, you code it. This is the traditional programming paradigm. But coding is not easy, right :) For some tasks, you can program these rules but some tasks are complex and it’s impossible to figure out all the possible rules. Thus, we want the machines to do these intelligent tasks without being explicitly programmed. But from where do we get these rules then? As we want the machines to do the tasks that we can do, then why not make them learn like we do? Humans learn from experience, when a human child sees a picture of a cat, it can learn to identify all the cats in the world. This idea gave birth to “Machine Learning” — a subfield of AI that enables computers to learn from data. After showing a large number of labeled images of a cat to a computer, it can identify the set of patterns and learn the rules that enables it to start identifying cat images. This type of algorithm was known as Perceptron developed by Frank Rosenblatt in 1957 which was designed to mimic the functioning of biological neuron. This marked the beginning of a long and fruitful journey in the field of machine learning.

Source: https://www.analyticsvidhya.com/blog/2021/10/perceptron-building-block-of-artificial-neural-network/

After the Perceptron, many other important machine learning algorithms were developed, and all of them have contributed to the growth of the machine learning field which we see today. Here is a list of the most significant breakthroughs.

1960s: k-Nearest Neighbors (kNN)

1967: Support Vector Machine (SVM)

1970s: The Decision Tree Algorithm

1980s: The Backpropagation Algorithm (used for training neural networks)

1990: Kernel Methods

1995: Random Forests

1996: Deep Belief Networks

1997: Long Short-Term Memory (LSTM)

1999: AdaBoost

2012: The Convolutional Neural Network (AlexNet)

2014: The Generative Adversarial Network (GAN)

2014: Word2Vec

2015: ResNet (Residual Neural Network)

2016: The Recurrent Neural Network (RNN)

2016: AlphaGo (Reinforcement Learning)

2017: The Transformer Model

2018: Bidirectional Encoder Representations from Transformers (BERT)

I have only included major breakthroughs up to 2018 because after that the AI space boomed and is still a booming industry. The list above is extensive yet still not exhaustive because AI/ML is a really hot research topic and so many amazing research papers are published every week. If you want to stay up to date with the recent trends in AI/ML, I recommend you to stay active on X/Twitter and make sure to follow these accounts,

Researchers and Academics

1. Andrew Ng (@AndrewYNg) — AI pioneer and founder of AI Fund
2. Yann LeCun (@ylecun) — Director of AI Research at Facebook and Turing Award winner
3. Fei-Fei Li (@drfeifei) — Director of the Stanford AI Lab and former Chief Scientist of AI at Google Cloud
4. Ian Goodfellow (@goodfellow_ian) — Research Scientist at Google Brain and GAN inventor
5. Sebastian Thrun (@SebastianThrun) — Founder of Udacity and AI researcher
6. Geoffrey Hinton (@geoffreyhinton) — Turing Award winner and AI pioneer
7. Demis Hassabis (@demishassabis) — Co-founder and CEO of DeepMind
8. David Silver (@DavidSilverUK) — Co-founder and CTO of DeepMind

AI/ML Influencers and Bloggers

1. KDNuggets (@kdnuggets) — Popular AI/ML blog and news site
2. AI Alignment (@AI_Alignment) — AI safety and ethics discussions
3. AI Impacts (@AI_Impacts) — AI safety and ethics discussions
4. Machine Learning Mastery (@MachineLearnin) — AI/ML tutorials and guides
5. AI in Industry (@AI_in_Industry) — AI applications and industry trends
6. AI Business News (@AIBusinessNews) — AI business news and trends
7. AI Trends (@AITrends) — AI trends and insights
8. The AI Alignment Podcast (@AI_Alignment_P) — AI safety and ethics discussions

AI/ML Companies and Organizations

1. Google AI (@GoogleAI) — Official Google AI Twitter account
2. DeepMind (@DeepMindAI) — AI research organization
3. Facebook AI (@FacebookAI) — Official Facebook AI Twitter account
4. Microsoft AI (@MicrosoftAI) — Official Microsoft AI Twitter account
5. NVIDIA AI (@NVIDIAAI) — Official NVIDIA AI Twitter account
6. Amazon AI (@AmazonAI) — Official Amazon AI Twitter account
7. IBM Watson (@IBMWatson) — Official IBM Watson Twitter account
8. AI for Everyone (@AI_for_Everyone) — AI education and awareness

Conferences and Events

1. NeurIPS (@NeurIPSConf) — Conference on Neural Information Processing Systems
2. ICML (@ICMLconf) — International Conference on Machine Learning
3. ICCV (@ICCVconf) — International Conference on Computer Vision
4. AAAI (@AAAIConference) — Association for the Advancement of Artificial Intelligence
5. AI Summit (@AI_Summit) — AI conference and event

Remember to also follow AI/ML-related hashtags, such as
#AI, #MachineLearning, #DeepLearning, #NLP, and #ComputerVision,
to stay updated with the latest news and discussions in the field.