How to Evolve Beyond Glorified Curve-Fitter

From zero to modeling hero...

These days, a machine learning workflow typically looks like this:

For many, the craft of data science and machine learning boils to importing from a machine learning bag of tricks,

ML Bag of Tricks

and tweaking nuances of this training workflow. With enough practice, a practitioner gets paid handsomely to repeatedly build, debug, and rebuild this curve-fitting workflow across various problems.

But as the market for glorified curve-fitters saturates, how will you and your team add value?

After completing this workshop, practitioners will have had a complete refactoring of their mental frameworks for modeling.

Specifically, they will have

Seen how to iteratively build a generative machine learning model from scratch using probabilistic programming.
Used deep learning tools without buying into deep learning hype.
Gone beyond accuracy and connect model performance directly to a business's key performance indicators.
The ability to talk about the inductive bias of a model using language from the domain they are modeling.
Applied mental models from Bayesian decision theory, information theory, Shannon's communication theory, and dynamic systems to model building.

Enrollment is currently closed

Your Instructor

Robert Osazuwa Ness

Robert didn't start in machine learning. He started his career by becoming fluent in Mandarin Chinese and moving to Tibet to do developmental economics fieldwork. He later obtained a graduate degree from Johns Hopkins School of Advanced International Studies.
After switching to the tech industry, Robert's interests shifted to modeling data. He attained his Ph.D. in mathematical statistics from Purdue University, and then he worked as a research engineer in various AI startups. He has published in journals and venues across these spaces, including RECOMB and NeurIPS, on topics including causal inference, probabilistic modeling, sequential decision processes, and dynamic models of complex systems. In addition to startup work, he is a machine learning professor at Northeastern University.

Frequently Asked Questions

When does the course start and finish?

Course enrollment opens periodically. You have access to the instructor for four months after the course opens. You will have indefinite access to the course materials thereafter.

How long do I have access to the course materials?

How does lifetime access sound? After enrolling, you have unlimited access to this course for as long as you like - across any and all devices you own.

What if I am unhappy with the course?

We would never want you to be unhappy! If you are unsatisfied with your purchase, contact us in the first 30 days and we will give you a full refund.

What is the technical background required for this course?

This is a high-level course. Anyone who is comfortable thinking quantitatively about data is suited for this course. The "Do Causality like a Bayesian" section will unavoidably introduce some mathematical notation. For those seeking just a high level understanding of the material, they can safely skip over the notation and focus on the common English explanations.

I am a manager/investor after a high-level overview that can inform stategic decisions. Is this for me?

Yes. In contrast with other causal modeling courses at this school, this one stays at a high level. This course provides mental models for thinking about causality in the context of machine learning. For strategic decision-makers, the models will contextualize causal reasoning within an organization's machine learning objectives.

I already know all about Bayes and generative machine learning. Will this course benefit me?

Very probably yes. The course goes beyond low-level Bayesian math to high-level mental frameworks for modeling. If you have some exposure to the math, you will still make connections that people don't typically make. The goal of the course is to give you a unique way of thinking about problems, rather than teach you math. You don't have to take our word for it. Try it, and get a refund if it doesn't work for you.

Where's the causal inference? This doesn't look like causal inference...

The goal of this course is to connect causal modeling to machine learning in a practical way. Specific causal inference topics such as causal effect estimation, confounder adjustment, propensity scores, instrumental variables, potential outcomes, etc. are covered in other AltDeep courses.

The ML community is in a mental rut. It's time for you to opt-out.

Break free from trend-following with AltDeep's radical approach to machine learning.

Those who complete this course break the mold of engineer/researcher/data scientist. They walk away with unconventional causality-based mental models for tackling practical problems in practice and research.