Insight The Reality of 'Continual Learning' ― Can AI Truly Evolve on Its Own?
As many AI services claim 'continual learning' capabilities, here's how to distinguish genuine self-learning from glorified note-taking.
Read →
Machine Learning
14 articles tagged with "Machine Learning"
Insight 
Series Part 1: From Photo Booths to Virtual Try-On — The 20-Year Quest
The origin story of META FIT: how a decades-old vision of seeing yourself in clothes before buying evolved from hardware kiosks to GAN-powered virtual try-on, plus a comprehensive survey of 15+ VTON research models.
Read →
Series Part 2: Understanding GANs — The Engine Behind Virtual Try-On
A deep dive into Generative Adversarial Networks: how the generator-discriminator dynamic works, why GANs dominated image generation before diffusion models, and how they power virtual try-on systems.
Read →
Series Part 4: Pose Estimation, Body Measurement, and 3D Reconstruction
How OpenPose skeletal detection, Graphonomy human parsing, and custom body measurement algorithms work together to enable accurate virtual fitting — plus an exploration of PiFu for 2D-to-3D reconstruction.
Read →
Series Part 3: Inside PF-AFN — The Try-On Engine in Code
A code-level walkthrough of the Parser-Free Appearance Flow Network: Feature Pyramid encoding, CUDA-accelerated correlation kernels, optical flow warping, and the ResUnet generator that composites garments onto people.
Read →
Series Part 5: Results, Failure Modes, and the Path to Modern Image Generation
What the GAN-based virtual try-on system achieved, where it failed (and why), the smartphone app design, and how diffusion models are changing everything for the next generation of META FIT.
Read →
Series Part 2: Finding 'Same Nutrition, Different Meal' with Cosine Similarity
Using cosine similarity on nutritional vectors to find recipes that match a target meal's nutrition profile but offer completely different flavors — at both the recipe and menu level.
Read →
Series Part 1: Solving the Daily Meal Planning Problem with Data
How I tackled the universal 'what's for dinner' problem over a decade ago using classical data science — cleansing 20,000 recipes, 200,000 ingredient records, and nutritional data into a unified ML-ready dataset.
Read →
Series Part 3: Predicting 'Non-Boring' Menus with LSTM Time Series
Reframing meal planning as a text generation problem — using LSTM neural networks with temperature-controlled sampling to predict diverse, non-repetitive menus from historical meal sequences.
Read →