https://learnrahulrai-ui.github.io/ml_blog/ Machine learning worked out by hand: every rule stripped of jargon, drawn as a picture, solved with a pencil, then written in code. A short book in ten chapters -- no hype, no frameworks, no JavaScript. en Sun, 14 Jun 2026 15:42:09 +0000 https://learnrahulrai-ui.github.io/ml_blog/posts/lstm-from-pencil.html https://learnrahulrai-ui.github.io/ml_blog/posts/lstm-from-pencil.html Sun, 14 Jun 2026 00:00:00 +0000 A KATA: one clean, self-contained walk through the RNN and the LSTM by pencil, made to be redone on a blank sheet again and again until your hand knows it. Words into 32-number rows, grid-times-row by hand, the RNN's one-memory recipe and why it FADES (a theorem), then the... https://learnrahulrai-ui.github.io/ml_blog/posts/the-deep-factory.html https://learnrahulrai-ui.github.io/ml_blog/posts/the-deep-factory.html Sat, 13 Jun 2026 00:00:00 +0000 The simple factory lands ~70%. Three layers of armour push it further: a humbler steadies every inspector's 65,536 numbers to middle 0, scatter 1 (arithmetic done by hand, full clerk count); send-home zeroes a random 25% of sheet entries to break secret teams (one coin per... https://learnrahulrai-ui.github.io/ml_blog/posts/the-two-memory-worker.html https://learnrahulrai-ui.github.io/ml_blog/posts/the-two-memory-worker.html Sat, 13 Jun 2026 00:00:00 +0000 The plain worker crushes its one memory every word, so word 1 fades to nothing by word 90. The fix, derived by hand: carry TWO memories -- a long keep (A, never crushed, free to grow past 1) and a spoken recent (B) -- and let three 0..1 voters decide what to keep, admit, and... https://learnrahulrai-ui.github.io/ml_blog/posts/the-walking-worker.html https://learnrahulrai-ui.github.io/ml_blog/posts/the-walking-worker.html Sat, 13 Jun 2026 00:00:00 +0000 A photo was already numbers; a review is WORDS, and a factory can only multiply numbers. Number the 10,000 commonest words by frequency, pad every review to 100, then a NOTEPAD swaps each word-number for a learned 32-number note (so "boring" and "dull" drift close, and "not" can... https://learnrahulrai-ui.github.io/ml_blog/posts/five-machines-against-memorising.html https://learnrahulrai-ui.github.io/ml_blog/posts/five-machines-against-memorising.html Fri, 12 Jun 2026 00:00:00 +0000 A network that learns too well memorises the study pile's freckles and flunks new patients. One plain machine catches the disease on a sheet of clothing photos; four cures each fight it differently -- a tax on big dials (L2), sending clerks home (dropout), a fire alarm that... https://learnrahulrai-ui.github.io/ml_blog/posts/the-magic-paper-factory.html https://learnrahulrai-ui.github.io/ml_blog/posts/the-magic-paper-factory.html Fri, 12 Jun 2026 00:00:00 +0000 Flatten a photo and you destroy every "next to" -- so this machine keeps the picture folded and slides a tiny 3x3 window (a magic paper of nine dials) across it, lighting up where its shape sits. Built one worker at a time with every wrong picture corrected: the edge-finder that... https://learnrahulrai-ui.github.io/ml_blog/posts/how-the-dials-learn.html https://learnrahulrai-ui.github.io/ml_blog/posts/how-the-dials-learn.html Thu, 11 Jun 2026 00:00:00 +0000 The dials finally learn. Backpropagation worked by hand on a one-input, one-hidden, one-output network: the chain rule link by link, the (guess - truth) shortcut derived, every slope checked against a brute-force wiggle, then the error sent one room further back through the... https://learnrahulrai-ui.github.io/ml_blog/posts/the-web-of-adders.html https://learnrahulrai-ui.github.io/ml_blog/posts/the-web-of-adders.html Thu, 11 Jun 2026 00:00:00 +0000 A fully self-contained start -- no prior posts needed. One room of clerks draws straight lines; stacked rooms draw anything, but only if a non-linear bend (the zero-out rule, ReLU) breaks the collapse between them. The forward pass walked by hand: three-cut split, humbling the... https://learnrahulrai-ui.github.io/ml_blog/posts/clustering-reference.html https://learnrahulrai-ui.github.io/ml_blog/posts/clustering-reference.html Tue, 09 Jun 2026 00:00:00 +0000 The loose ends Chapter 6 does not stop for: the k you pick, two more rulers (Hamming for words, Mahalanobis for stretched many-walled rooms), the missing-data traps, why crush a room at all, and the part the lab skips -- the ethics of sorting PEOPLE into piles (bias, privacy,... https://learnrahulrai-ui.github.io/ml_blog/posts/eda-and-distance.html https://learnrahulrai-ui.github.io/ml_blog/posts/eda-and-distance.html Tue, 09 Jun 2026 00:00:00 +0000 No answer column, no right or wrong -- just 50 states and 3 measurements. The mean and spread of each column, the ruler problem (Assault in hundreds drowns UrbanPop in tens), two gap rulers (straight-line Euclidean and city-block Manhattan), the sheet of gaps, and the fix: put... https://learnrahulrai-ui.github.io/ml_blog/posts/filling-the-blanks.html https://learnrahulrai-ui.github.io/ml_blog/posts/filling-the-blanks.html Tue, 09 Jun 2026 00:00:00 +0000 A sheet that is mostly holes: users x movies, a rating only where someone watched. Split the holey grid into two skinny full sheets of hidden tastes and flavours -- PCA's scores-times-recipe shape again. Then the lab end to end: humble each movie column, drop lazy zeros, and run... https://learnrahulrai-ui.github.io/ml_blog/posts/grouping-by-nearest-centre.html https://learnrahulrai-ui.github.io/ml_blog/posts/grouping-by-nearest-centre.html Tue, 09 Jun 2026 00:00:00 +0000 The first machine that actually GROUPS: pick K, then loop two moves -- assign every dot to its nearest centre, move each centre to its pile's mean. Tightness as the score, the unlucky-start trap, and picking K by the elbow. https://learnrahulrai-ui.github.io/ml_blog/posts/nci60-revisited.html https://learnrahulrai-ui.github.io/ml_blog/posts/nci60-revisited.html Tue, 09 Jun 2026 00:00:00 +0000 All three tools on one hard sheet: 64 samples, 6,830 genes -- more columns than rows. Crush with PCA, carve with K-means and the family tree, then unseal the cancer types and grade the blind groups. Peeking at labels is a scorecard, never an input. https://learnrahulrai-ui.github.io/ml_blog/posts/the-family-tree-clustering.html https://learnrahulrai-ui.github.io/ml_blog/posts/the-family-tree-clustering.html Tue, 09 Jun 2026 00:00:00 +0000 No K up front: start with everyone alone and marry the two closest groups over and over, recording each wedding's height. Read the family tree, cut it wherever you like, and the cut decides how many groups. Single, complete, average, and Ward linkage compared. https://learnrahulrai-ui.github.io/ml_blog/posts/the-strongest-direction-pca.html https://learnrahulrai-ui.github.io/ml_blog/posts/the-strongest-direction-pca.html Tue, 09 Jun 2026 00:00:00 +0000 13 chemical measurements = 13 walls. You cannot draw a 13-wall room. Shine a flashlight, trace the longest shadow (PC1), then the next at a right angle (PC2). A 2-column by-hand worked example, the recipe (loadings), how much each shadow carries (PVE), choosing how many to keep... https://learnrahulrai-ui.github.io/ml_blog/posts/committees-bag-forest-boost.html https://learnrahulrai-ui.github.io/ml_blog/posts/committees-bag-forest-boost.html Sun, 07 Jun 2026 00:00:00 +0000 One chart memorises and panics; 200 averaged charts are eerily steady. Bagging re-deals with repeats and averages 200 trees (OOB free exam built in). Random forest hides a handful of columns at each cut so the trees truly differ. Boosting chains stumps in a line, each fixing the... https://learnrahulrai-ui.github.io/ml_blog/posts/humble-dials-ridge-lasso.html https://learnrahulrai-ui.github.io/ml_blog/posts/humble-dials-ridge-lasso.html Sun, 07 Jun 2026 00:00:00 +0000 Free dials grow huge and memorise; put them on a leash. Ridge (the square fine) shrinks all dials and zeroes none, solving in one step. Lasso (the absolute fine) snaps weak dials to exactly zero and hands you a shortlist of columns. Pick the knob by the valley. https://learnrahulrai-ui.github.io/ml_blog/posts/mixing-ruler-gini-pruning.html https://learnrahulrai-ui.github.io/ml_blog/posts/mixing-ruler-gini-pruning.html Sun, 07 Jun 2026 00:00:00 +0000 Flip to the cancer sick/well sheet: the badness ruler swaps from squared misses to Gini (chance two random grabs disagree) -- derived by hand from four counts. Then pruning: a tax per leaf snips weak branches after the tree is fully grown; the alpha menu comes from the tree... https://learnrahulrai-ui.github.io/ml_blog/posts/question-charts-by-hand.html https://learnrahulrai-ui.github.io/ml_blog/posts/question-charts-by-hand.html Sun, 07 Jun 2026 00:00:00 +0000 No dials, no formula: a machine that asks yes/no questions. Built from scratch -- sort each column, try every midpoint, pick the cleanest cut, recurse. Five wrong pictures corrected: one column per split, depth is people not columns, build and use are separate phases, same... https://learnrahulrai-ui.github.io/ml_blog/posts/the-dial-by-hand.html https://learnrahulrai-ui.github.io/ml_blog/posts/the-dial-by-hand.html Sun, 07 Jun 2026 00:00:00 +0000 The whole machine on a blank sheet: the dials are (X^T X)^-1 X^T y, the untangler that shares credit between overlapping columns. The one-column slope is a LADDER that matches only when columns don't overlap -- name it out loud before you build anything on it. https://learnrahulrai-ui.github.io/ml_blog/posts/wobble-bands-bootstrap.html https://learnrahulrai-ui.github.io/ml_blog/posts/wobble-bands-bootstrap.html Sun, 07 Jun 2026 00:00:00 +0000 A single dial value hides how shaky it is. Fake 200 hospital visits by re-dealing the pile with repeats, and read off a 95% wobble band per dial (chop 2.5% each end). The ~37% left out of each deal become a free, honest exam. https://learnrahulrai-ui.github.io/ml_blog/posts/classification-lab.html https://learnrahulrai-ui.github.io/ml_blog/posts/classification-lab.html Sat, 06 Jun 2026 00:00:00 +0000 All classification concepts from Chapter 3 in one flip-to reference: cross-entropy vs MSE, the C parameter, LDA with priors, hyperparameters vs dials, GridSearchCV fold safety, min-max scaling, precision vs recall in business, ROC vs PR curves, skewed piles, and... https://learnrahulrai-ui.github.io/ml_blog/posts/leash-and-cloud.html https://learnrahulrai-ui.github.io/ml_blog/posts/leash-and-cloud.html Fri, 05 Jun 2026 00:00:00 +0000 Two ways to humble an over-confident machine. L2: add a squared-dial price to the leftover so no single dial dominates (C = 1/lambda -- small C is a heavy squeeze). LDA: forget rolling downhill; read the two cloud centres and their shared spread, then solve for the wall in one... https://learnrahulrai-ui.github.io/ml_blog/posts/picking-settings-skewed-classes.html https://learnrahulrai-ui.github.io/ml_blog/posts/picking-settings-skewed-classes.html Fri, 05 Jun 2026 00:00:00 +0000 Grid-hunting settings by rotating folds; why pinch-to-fit scaling breaks around outliers; the precision-vs-recall trade in fraud detection; precision-recall curves for when one class is 100x rarer; and macro vs micro averaging for 3+ bins. https://learnrahulrai-ui.github.io/ml_blog/posts/sorting-into-bins.html https://learnrahulrai-ui.github.io/ml_blog/posts/sorting-into-bins.html Fri, 05 Jun 2026 00:00:00 +0000 The shift from "guess a number" to "guess a bin": the S-curve squash, the cross-entropy leftover, and how any dial sum becomes a chance. Then the four-box table of CAUGHT/ALARM/MISSED/CLEAR, four scores that read different corners of it, and why a machine that catches nobody can... https://learnrahulrai-ui.github.io/ml_blog/posts/the-trade-curve.html https://learnrahulrai-ui.github.io/ml_blog/posts/the-trade-curve.html Fri, 05 Jun 2026 00:00:00 +0000 One cutoff judges the machine at one spot; slide it through every position and you get the trade curve (ROC). The area beneath it -- AUC -- compresses the whole curve to one number: pick a random sick lump and a random well lump; AUC is the chance the machine scored the sick one... https://learnrahulrai-ui.github.io/ml_blog/posts/grading-the-guess-mse-r2.html https://learnrahulrai-ui.github.io/ml_blog/posts/grading-the-guess-mse-r2.html Thu, 04 Jun 2026 00:00:00 +0000 The same straight-stick rule on a sheet of cars, graded two ways: MSE (RMSE with the root left off) and R^2 -- a race against a fool who only ever shouts the average. Why you divide by the whole wobble, why test R^2 can dip below zero, and the stiff-vs-jumpy split of the... https://learnrahulrai-ui.github.io/ml_blog/posts/reading-the-dials.html https://learnrahulrai-ui.github.io/ml_blog/posts/reading-the-dials.html Thu, 04 Jun 2026 00:00:00 +0000 Each column keeps one dial -- plus pushes up, minus drags down. Zip the names back on, find the most-negative (a MIN, not a MAX), and watch the trap: raw dials wear each column's own units, so they are not comparable until the columns sit on one shared ruler. https://learnrahulrai-ui.github.io/ml_blog/posts/chapter1-what-ml-actually-does.html https://learnrahulrai-ui.github.io/ml_blog/posts/chapter1-what-ml-actually-does.html Mon, 01 Jun 2026 00:00:00 +0000 The whole picture on the California housing pile: why 20% of rows are locked away first, the ask-closest rule and the straight-stick rule introduced plainly, RMSE against the always-average baseline, and picking settings by rotating folds. https://learnrahulrai-ui.github.io/ml_blog/posts/neighbor-machine-knn.html https://learnrahulrai-ui.github.io/ml_blog/posts/neighbor-machine-knn.html Mon, 01 Jun 2026 00:00:00 +0000 The ask-closest rule from the gap formula up: why a column measured in thousands runs the whole show and must be put on the same ruler, the cost per guess and why close stops meaning close as columns pile up, k as the stiff-vs-jumpy dial. https://learnrahulrai-ui.github.io/ml_blog/posts/straight-stick-linear-regression.html https://learnrahulrai-ui.github.io/ml_blog/posts/straight-stick-linear-regression.html Mon, 01 Jun 2026 00:00:00 +0000 From the bowl-shaped squared leftover to the exact dials: a three-row pencil solve of the flat-point equations, right-angle leftovers and the flat-shadow thrower, the Gauss-Markov conditions, and why you cut the sheet (QR/SVD) rather than flip it.