Wait, Aren’t They Just Three Identical Classes? No, They Are Not.
Date:
Machine Learning, Deep Learning, Machine Perception
Table of Contents
Introduction
As you’ve probably noticed by now, I love cramming thematically similar courses into the same semester. It’s my version of academic cross-training: if I saturate myself in a topic from multiple directions, I can actually see how the pieces fit together. This trio—Machine Learning (ML), Deep Learning (DL), and Machine Perception (MP)—was my latest attempt. And no, despite how the course catalog makes them sound, they are not the same class.
ML: Deep Learning (ML:DL)
This course was exactly what you’d expect if someone asked, “Can you please give me the Keras/TensorFlow crash course, but make it rigorous?” We marched through convolutional networks, recurrent networks, attention mechanisms, and the practical tricks you need to keep them from exploding into NaNs. The emphasis: how to build modern deep learning models, and the math that makes them run. [placeholder for architecture diagrams here]
Machine Learning (ML)
The “older sibling” course. Less about GPUs, more about the fundamentals—probabilistic models, bias/variance trade-offs, optimization, and classical algorithms like SVMs, decision trees, and ensemble methods. It was the place where you finally understood why logistic regression is both simple and deceptively powerful. And every equation came with a reminder that this stuff predates the deep learning hype by decades. [insert equation block: e.g., loss function minimization]
Machine Perception (MP)
This was the “applied” sibling—same family, different career path. MP leaned hard into vision: how to make sense of pixels, shapes, and signals. It’s where we zoomed in (literally) on convolution, feature extraction, and multimodal input. If ML:DL gave me the raw tools, MP was about where to point the camera (pun intended) and what to do with the signal. [placeholder for example image/feature map]
So—three classes, three perspectives:
- ML gave me the foundations.
- DL showed me the bleeding-edge implementations.
- MP grounded it in vision and perception.
Together, they felt like a complete ecosystem. And I loved every bit of it.