PinnedInTowards AIbyTilak Varma·Nov 2Word2Vec Finally Explained: How “King − Man + Woman” Became “Queen”Research Paper Breakdown #6. CBOW, Skip-gram, hierarchical softmax, and the 2013 paper that taught computers what words mean, explained…A response icon1A response icon1
PinnedInTowards AIbyTilak Varma·Oct 26ResNet Finally Explained: Why Deeper Networks Got Worse, and the Shortcut That Fixed ItResearch Paper Breakdown #5. Residual learning, shortcut connections, and the 152-layer network that won ImageNet 2015, explained with the…A response icon2A response icon2
PinnedInTowards AIbyTilak Varma·Oct 19Batch Normalization Finally Explained: Why Deep Networks Stop Chasing a Moving TargetResearch Paper Breakdown #4. Batch statistics, learnable gamma and beta, and the train/eval switch that breaks models in production…A response icon1A response icon1
PinnedInTowards AIbyTilak Varma·Oct 16Adam Finally Explained: Why One Optimizer Trains So Many Neural NetworksResearch Paper Breakdown #3 .Momentum, RMSProp, and the bias correction that makes the first few steps work, explained with the math.A response icon2A response icon2
PinnedInTowards AIbyTilak Varma·Oct 13DropOut Finally Explained : Why Breaking Half Your Neural Networks Makes It SmarterResearch Paper Breakdown #2. Random masking, inverted scaling, and the implicit ensemble that regularizes without a penalty term, explained…A response icon2A response icon2
InTowards AIbyTilak Varma·5d agoBahdanau Attention Finally Explained: How Neural Translation Learned Where to LookResearch Paper Breakdown #7. The fixed-vector bottleneck, bidirectional encoders, and soft alignment — explained with the math, a worked…
InTowards AIbyTilak Varma·Nov 5A Prototype Is Not a ProductAn open letter to product managers, directors and non-technical heads who watched an AI build an MVP in an afternoon and quietly moved the…
InTowards AIbyTilak Varma·Sep 29AlexNet Finally Explained: The 2012 Breakthrough That Changed Computer VisionResearch Paper Breakdown #1. ReLU, dropout, overlapping pooling, and the two gaming GPUs that won ImageNet 2012, explained with the math.A response icon2A response icon2
InTowards AIbyTilak Varma·Sep 27Cloudflare Saved 100TB of RAM With Math. I Thought It Was Ragebait.I read the whole boring, math-heavy blog so you don’t have to. Here’s what they actually did, explained simply.A response icon2A response icon2
Tilak Varma·Sep 182 A.M. and Not a Single Ounce of SleepA day I didn’t plan, written down before I forget it.