How can I start learning about artificial intelligence and machine learning? originally appeared on Quora: the place to gain and share knowledge, empowering people to learn from others and better understand the world.
Answer by Alexandre Robicquet, Co-Founder of Crossing Minds, on Quora:
For those of you who are interested in careers in AI and machine learning, we recommend the following:
- Learn how to CODE: Coding is an incredible exercise of discipline and logic, which – when done the right way – can help your mind grasp problems and solutions you wouldn’t have originally considered. A great (way) to start would be Python, which is a high-level and sophisticated programming language, yet very practical for machine learning.
- OWN what you’re coding: Some people claim to be ML engineers or AI engineers because they’re capable cloning a git repository (borrowing a chunk of code that someone wrote and made public) for a specific task or follow a tutorial line-by-line. It is a great start, however, there’s nothing more harmful (technically speaking) for an AI company than an engineer that does not understand what (s)he is doing, coding and deploying. Understanding and owning your code (as small as you may think it is) will give you an incredible advantage and control over your AI project. It doesn’t matter if it is not the most “optimized” code at first, as long as you understand it. One good exercise would be to participate to Kaggle competitions or actively contribute to a popular github repository. Both will give you a validation from the community, that are very valuable for companies hiring ML-focused engineers.
- UNDERSTAND what you’re coding. Machine Learning is a complex and vast field, which is based on specific mathematical concepts and statistical approaches. Understanding the mathematics behind the code, will give you an incredible advantage when it comes to optimizing your algorithm, fixing a bug, or simply recognizing a problem and translating it into an AI problem. This also will help you gather the right dataset and own your code.
(this last section concerns ML & Startups – not research)
- Don’t Invent Problems to Solve. It’s not uncommon to see startups, especially in Silicon Valley, launched because the founders have a solution (an algorithm, a dataset, a pipeline, etc.) and decide to then invent a problem. Please, don’t do that. The best way to successfully build and grow an AI startup is to identify a REAL problem in people’s everyday lives and then find a solution that you CODE, OWN, and UNDERSTAND.
Finally, PERSEVERE. AI and ML are complicated fields that require a lot of discipline and work. This is a long journey, so hold on. Be humble, never hesitate to ask questions and help your community.
This question originally appeared on Quora – the place to gain and share knowledge, empowering people to learn from others and better understand the world. You can follow Quora on Twitter and Facebook. More questions:
Credit: Google News