Cristian Leo

Google, Microsoft, LinkedIn, and many more tech companies are using Graph RAG. Why? Let’s understand it by building one from scratch. Cristian Leo · Follow Published in Towards Data Science · 24 min read · 7 hours ago — Image illustrating a knowledge graph with interconnected nodes and edges against a tech-inspired gradient background — Image generated by the author using DALL-E You may not realize it, but you’ve been interacting with Knowledge Graphs (KGs)

Optimizing Sparse Neural Networks: Understanding Gradient Flow for Faster Training, Improved Efficiency, and Better Performance in Deep Learning Models Cristian Leo · Follow Published in Towards Data Science · 22 min read · 9 hours ago — AI Image generated depicting the gradients flowing in Neural Networks In recent years, the AI field has been obsessed with building larger and larger neural networks, believing that more complexity leads to better performance. Indeed, this approach has

How AutoGluon Dominated Kaggle Competitions and How You Can Beat It. The algorithm that beats 99% of Data Scientists with 4 lines of code. Cristian Leo · Follow Published in Towards Data Science · 19 min read · 8 hours ago — Image generated by DALL-E In two popular Kaggle competitions, AutoGluon beat 99% of the participating data scientists after merely 4h of training on the raw data (AutoGluon Team. “AutoGluon: AutoML for Text, Image,

Nadam is one of the most capable optimizers in Deep Learning. Let’s delve into its math, and build the algorithm from scratch. Cristian Leo · Follow Published in Towards Data Science · 18 min read · 7 hours ago — Image generated by DALL-E In our previous discussion on the Adam optimizer, we explored how Adam has transformed the optimization landscape in machine learning with its adept handling of adaptive learning rates. Known for its…

Dive into AlexNet, the first modern CNN, understand its mathematics, implement it from scratch, and explore its applications. Cristian Leo · Follow Published in Towards Data Science · 32 min read · 4 hours ago — Image generated by DALL-E Convolutional Neural Networks (CNNs) are a specialized kind of deep neural networks designed primarily for processing structured array data such as images. CNNs operate by recognizing patterns directly from pixel data of images, eliminating the

Dive into CNN, the backbone of Computer Vision, understand its mathematics, implement it from scratch, and explore its applications Cristian Leo · Follow Published in Towards Data Science · 26 min read · 9 hours ago — Image by DALL-E Index · 1: Introduction · 2: The Math Behind CNN Architecture ∘ 2.1: Convolutional Layers ∘ 2.2: Stride ∘ 2.3: Padding ∘ 2.4: Multiple Filters and Depth ∘ 2.5: Weight Sharing ∘ 2.6: Feature Map

Dive into the techniques to fine-tune Neural Networks, understand their mathematics, build them from scratch, and explore their applications Cristian Leo · Follow Published in Towards Data Science · 31 min read · 10 hours ago — Image by DALL-E While you might get by in machine learning by trying out a few models, picking the best performer, and tweaking some settings, deep learning doesn’t play by the same rules. If you’ve ever experimented with