Welcome to Cortex, we are a full-stack end-to-end machine learning development company aimed at creating state of the art models. Based in the beautiful Republic of Moldova with our hearts open to the entire world, we feel like a part of the global AI community, connected through the internet and bonded by our common ideas.
We started our company to offer other organizations around the world the capability of AI. Our journey starts here with this post and as a welcoming gift (as are customs in our country), we want to share machine learning resources for curious minds to learn and advance in this field.
Books, being the best methods to learn new things, books offer an unprecedented level of diving deep into the unknown and explaining thoroughly all the details of Deep Learning.
- “Deep Learning” by Ian Goodfellow and Yoshua Bengio and Aaron Courville (probably the best available book on deep learning ever written)
- “Hands-on Machine Learning” with Scikit-learn, Keras, and TensorFlow, 2nd edition content by Aurélien Géron (amazing of ML and DL algorithms in action using python libraries and frameworks)
- “Deep Learning with Python” by François Chollet, the author of the book makes building neural networks as simple as building with lego blocks
- Free book: “Machine Learning Yearning” by Andrew NG (the most comprehensive book to get a brief feel of ML and to quickly jump in while learning best practices from the start)
- Free book: “Understanding Machine Learning” from Theory to Algorithms by Shai Shalev-Shwartz and Shai Ben-David (a more advanced approach to uncovering the mysteries of ML)
MOOC’s represent the majority of online resources taken by students online to get into machine learning while being mostly introductory level courses they provide a good understanding on the subject and the tools that leaders in the industry use, a good walkthrough by these instructors and you’re good to go start building your projects:
- Andrew Yang’s Deep Learning Specialization is probably the most popular online course to take when starting DL
- If you want to get a solid grasp on theoretical concepts and the necessary mathematics needed for machine learning ColumbiaX offers this EDX Course
- This one is a wonderful program if you like learning all the niche ideas in an area taught by Jeremy Howard (Kaggle’s #1 competitor 2 years running, and founder of Enlitic)
- Elements of AI from the University of Helsinki is aimed at presenting the capabilities of AI
- Get the feeling of being in a top University with Oxford’s Deep Learning lectures
Youtube videos are free and can save you a lot of time while displaying more of an introduction these tutorials can familiarize you with key concepts and notions:
- Daniel Bourke (the nicest guy on the planet and the best video to dive straight into ml essentials)
- Sentdex has one of the best videos available on the most popular frameworks Tensorflow and Pytorch covering CNN’s and RNN’s — TensorFlow, PyTorch
- Tech With Tim— a python expert explaining theoretical ml with easy to grasp visuals and presenting them with code live, what more could you wish for
- Lex Fridman — the best podcasts on AI with industry leaders
- Statquest — high-quality visuals about ml algorithms
AI is our passion, machine learning is the field of activity but deep learning is our specialty and what we do best, our engineers are DL enthusiasts who believe in the power of neural nets to revolutionize industries across every category.
Deep Learning is defined usually as a multi-layer perceptron (a perceptron is a neural network unit, an artificial neuron, that does certain computations to detect features or business intelligence in the input data) it was introduced by Frank Rosenblatt in 1957 but there appeared a problem with a single perceptron, it can’t represent the boolean XOR function. So multi-layer perceptrons or Deep Learning as we know it today was born. Since its creation, the perceptron model went through significant modifications.
We discovered different activation functions, learning rules, and even weight initialization methods. It is multi-layered which is the biggest difference from “shallow” machine learning algorithms (Linear Regression, SVM, Naive Bayes) shallow meaning they have only one layer that computes operations on data. Deep Learning is expected to boom soon and has great potential to disrupt industries such as insurance, healthcare and life sciences, retail, manufacturing, travel and hospitality, financial services, and energy.
In the future, you will read blogs about interesting AI topics like the Philosophy of an AI system, ethics, psychology, art, and a special one on humor, on which we wish to hear your opinion. Well try to write blogs in regards to what humanity has discovered so far on this topics and their version in the AI world with respects to current technological discoveries and in the scope of what is achievable, of course non of the aforementioned points are fully understood in the human world, but well try our best to find tangent points and similarities between them. The term Artificial Intelligence has been associated with over the top expectations and hype.
Lately, public perception has been covered with lots of incomprehensible information and impossible feats circulate its image and even worse destroys common opinion — movies like “Terminator” where Skynet wants to destroy humanity and HAL 9000 from “2001: A Space Odyssey” add to this effect.
We strive to clear its name and explain in layman’s terms what is AI to the general public while contemplating the future of this technology and inventing useful applications for this wonder-tech. In the end, I hope that people will better understand it and appreciate the true value it can offer humanity.
Over the years the world has had wrong predictions about technologies that they don’t get, for example, the printing press was seen as magic, and carts that went without horses were something amazing and unseen until then, of course, innovations opened the door to speculation but then after the world embraced technology with open arms and began to better understand them at the level that every man today has a clear mind about how a car works and how the newspaper is printed.
The same way if explained carefully AI can be understood and democratized so that public perception can still be shifted in a peaceful direction. In short, the way a hummer is a tool in a builders toolbox a Support Vector Machine is the essential tool of a data scientist.
While AI as an umbrella term that covers a lot of other fields, it is usually associated with Reinforcement Learning (is an area of machine learning concerned with how software agents ought to take actions in an environment in order to maximize the notion of cumulative reward) because of companies like DeepMind that beat the world champion at GO, Lee Sedol and Elon Musk’s OpenAI that beat world champions in Dota 2.
It involves algorithms that are rewarded in an environment as they “learn” over time, of course, their accomplishments are outstanding and cannot be undervalued, but they are still far from reaching AGI (Artificial General Intelligence is the hypothetical intelligence of a machine that has the capacity to understand or learn any intellectual task that a human being can) and even further from achieving digital consciousness.
We first have to define what human consciousness is (maybe computers will help us understand ourselves better in this regard). Educating about AI is fascinating because it involves multiple other disciplines outside of computer science. It requires a strong grasp of mathematics and statistical analysis. Besides AI involves tackling questions regarding ethics, psychological inquiries, and moral issues that have to be addressed before creating the system, you wouldn’t want to build a rocket ship after launching it.
At Cortex, we put ourselves in front of these challenges every day to find ethical answers and to keep us on track. Our technology stack consists of powerful frameworks like TensorFlow, Keras, and PyTorch, which are supported by giants like Google and Facebook. We are always researching what other possible algorithms and methods can be used. We do this to always keep an edge over the competition, to advance innovation in our community, and for the global well-being in general.
Since this field is so young many algorithms don’t see the light of day after being published in research papers or outside of their jupyter notebooks, which is unfortunate considering how advanced and powerful they are. Because of this, we need more engineers and free thinkers alike to find the best applications for these amazing tools and the uses cases that best fit them. Also, there is a huge need for modern ethicists and philosophers to start addressing important fundamental questions that await humanity over the horizon. Knowing how we persevere during hard times makes me hopeful for the future plus our ingenuity has always been our most valuable asset in these terms.
We are personally very excited to start this journey together with you and everybody at Cortex and to innovate together with the Global Machine Learning community. If you would like to hear more from us you can follow on social media for more news in the world of AI.