Skip to content
All tags

#machine-learning

127 posts

AI Engineer Interview Daily — 2026-08-31: ML Fundamentals

ML fundamentals interviews test whether you can diagnose the gap between 'the metric looks great' and 'production is on fire.' Today covers four high-frequency topics: why AUC-ROC inflates under heavy class imbalance, why cross-entropy beats MSE for classification (it comes down to vanishing gradients), whether bagging or boosting fixes variance versus bias, and the common misconception that multicollinearity hurts prediction — it only hurts interpretability.

Harvard CS50 AI Week 4: Learning — Supervised Learning, k-NN, SVM, Reinforcement Learning Q-learning & Nim

Week 4 enters ML: supervised classification (k-NN, SVM, Perceptron), model evaluation, RL basics (MDP, Q-learning, ε-greedy). Projects: Shopping (purchase prediction with k-NN) and Nim (learning to play via Q-learning).

How Does A/B Testing Turn a Product Change Into an Estimable Effect?

A/B testing turns a product change into an estimate with uncertainty. A useful report covers effect size, confidence, guardrails, randomization, and launch risk.

Why Not Run Many t-Tests? What Is ANOVA Protecting?

ANOVA first checks whether three or more group means differ overall, so you do not inflate false-positive risk by running many pairwise t tests.

Why Do Large-Sample Approximations Work, and When Do They Fail?

Large-sample normal approximation describes the behavior of estimators, not raw data. It is useful, but dependence, boundaries, and distribution shift can make it unreliable.

How Does Bayesian Inference Connect Prior, Data, and Posterior?

Bayesian inference updates uncertainty about an unknown parameter by combining prior belief with the likelihood from observed data, producing a posterior distribution.

What Do Bias, Variance, and Consistency Check in Point Estimation?

Bias checks whether an estimator is centered correctly, variance checks sampling fluctuation, MSE combines both, and consistency asks whether the estimator approaches truth as sample size grows.

When the Formula Distribution Is Unknown, How Does Bootstrap Estimate Uncertainty?

Bootstrap estimates uncertainty by resampling from the observed sample with replacement, rebuilding many sample-like datasets, and watching the statistic fluctuate.

Causal Inference Basics: Why Prediction Accuracy Does Not Mean Real Effect

Causal inference separates prediction from effect. A model can predict who will buy without proving that an intervention will make them buy.

How Do You Tell Goodness-of-Fit From Independence in Chi-Square Problems?

Chi-square tests compare observed counts with expected counts. First decide whether the problem is goodness-of-fit for one categorical variable or independence for two categorical variables.

When Should Bernoulli, Binomial, Normal, and Poisson Appear?

Distributions are names for data-generating situations, not formula cards. Learn when Bernoulli, Binomial, Poisson, and Normal distributions fit a problem.

How Do You Write Confidence Intervals Without Only Memorizing Bounds?

A confidence interval puts a point estimate back inside sampling fluctuation. Computing bounds is only the first step; you also need to explain standard error, critical values, and coverage.

When You See a Dataset, What Statistics Should You Check First?

Data type determines the statistical tools you can use. Start with categorical, numeric, count, and time-ordered data, then choose summaries that fit the question.

How Does the Delta Method Estimate Uncertainty for F1 and Ratio Metrics?

The delta method transfers uncertainty through a smooth function: the local derivative expands or shrinks the estimator's original standard error.

What Makes an Estimator Good: Bias, Variance, or MSE?

An estimator is a rule for using samples to infer a population parameter. To judge whether it is good, look at bias, variance, and MSE together.

When a Mixed Problem Appears, How Do You Pick the Tool in 30 Seconds?

At the final review stage, train problem recognition: identify data type, unknown quantity, and decision goal before choosing a formula and writing a contextual conclusion.

What Do Expectation and Variance Mean in Exams and Model Evaluation?

Expectation describes long-run center; variance describes fluctuation. This post computes E[X], E[X^2], and Var(X), then connects them to average loss and model stability.

How Does Experimental Design Make Results Interpretable Rather Than Merely Correlated?

Experimental design decides whether a result can be interpreted. Randomization, control, blocking, replication, blinding, and pre-specified outcomes give inference a usable foundation.

How Does Fisher Information Tell You Whether a Parameter Is Stable?

Fisher information uses likelihood curvature to measure how well the data locate a parameter; larger information usually means a smaller standard error for the MLE.

Confidence Intervals Are More Than t-Tables: What Is the General Construction?

A confidence interval is built by defining the target estimate, describing its sampling error, and choosing a rule that turns uncertainty into a range.

How Does a GLM Choose Distributions and Link Functions by Data Type?

A generalized linear model starts from the response type, chooses a suitable distribution, and uses a link function to connect the mean to a linear predictor.

From H0 to p-Values, What Decision Is a Hypothesis Test Making?

A hypothesis test is a decision process under uncertainty: write H0/H1, choose alpha, compute a test statistic and p-value, then decide whether the data is strong enough to challenge H0.

How Do Estimation, Testing, Likelihood, and Bayes Fit on One Inference Map?

The inference map starts with the question type: point estimate, uncertainty interval, decision test, likelihood model comparison, Bayesian update, or resampling.

How Does the Likelihood Ratio Test Compare Nested Models?

The likelihood-ratio test compares the log likelihood of a restricted model with a full model; the usual chi-square reference only makes sense under nested-model and approximation conditions.

When OLS Assumptions Fail, How Can the Regression Line Still Be Used?

OLS is a useful baseline, but coefficient interpretation, inference, prediction, and diagnosis depend on assumptions about linearity, errors, independence, and variance.

How Does Logistic Regression Move From Probability to Thresholds and Error Costs?

Logistic regression estimates probabilities first. Classification decisions come later, when thresholds turn those probabilities into actions under real error costs.

Why Should Classification Start With Log Odds?

Logistic regression connects a linear score to a probability between 0 and 1. Understanding odds, log odds, and odds ratios prevents wrong coefficient interpretations.

Why Does MAP Turn Priors Into Regularization?

MAP maximizes the posterior. After taking logs, the prior becomes a penalty term, which connects Bayesian estimation to L1, L2, and regularized ML objectives.

How Do Matching and Weighting Make Observational Data More Experiment-Like?

Matching and weighting do not turn observational data into a true experiment. They try to make treatment and control comparable on observed variables.

Why Does MLE Ask Which Parameter Most Likely Generated the Data?

MLE fixes the observed data and compares which parameter values make that data most plausible; log likelihood turns products into sums and connects directly to negative log loss.

Why Does the Method of Moments Match Sample Moments to Population Moments?

Method of Moments matches sample moments to theoretical population moments, then solves for parameters. It is not always the most efficient method, but it builds the first intuition for parameter estimation.

Missing Data Is Not Just Blank Cells: How Does It Distort Statistics and Models?

Missing data can change representativeness, bias estimates, and mislead ML systems. The first question is why the data are missing.

How Do You Write an ML/AI Evaluation Report That Is More Than a Leaderboard Score?

A useful ML/AI evaluation report turns statistical evidence into a decision: ship, stage, roll back, or run more experiments.

What Do Residuals, Outliers, and Leverage Reveal About Model Failure?

Model diagnostics turn fitted errors into evidence: residual patterns, outliers, leverage, and influential points reveal how a model fails.

How Does Multivariate Analysis Organize Features That Move Together?

Multivariate analysis looks at features together. Covariance, correlation, and PCA reveal shared directions that univariate summaries miss.

What Kind of Optimal Test Is the Neyman-Pearson View About?

The Neyman-Pearson view treats a test as a decision rule: under a fixed Type I error rate alpha, choose the rejection region with the highest power.

What Assumptions Do Nonparametric Methods Relax, and What Do They Cost?

Nonparametric methods are not assumption-free. They relax fixed distributional forms, often gaining flexibility while paying in efficiency, interpretation, or overfitting risk.

How Should You Analyze NTU IM 114-115 Statistics Papers Without Memorizing Answers?

Past papers train question-analysis discipline, not fortune-telling. Each problem should return to data type, unknown quantity, statistical tool, calculation path, and contextual conclusion.

How Do You Avoid Missing Cells in Joint Distribution and PMF Transformations?

Joint PMF problems require listing every cell. Marginalization, conditional probability, and variable transformations are all sums or regroupings of the original cells.

Conditional Probability, Independence, and Bayes: What Viewpoint Is the Problem Switching?

Probability problems are often hard because the viewpoint changes. Define events first, then distinguish conditioning, independence, mutual exclusivity, and Bayes' rule.

How Do Samples, Statistics, and Sampling Distributions Differ?

A sample is the data, a statistic is a function of the sample, and a sampling distribution is the distribution of that statistic under repeated sampling.

How Do PMF, PDF, and CDF Turn Probability Into Computation?

Random variables turn uncertain outcomes into numbers. PMF, PDF, and CDF then let you compute discrete probabilities, continuous interval probabilities, thresholds, and model-score distributions.

How Should coef, SE, t, F, and R-Squared Be Read Together?

A regression table is not a p-value list: coef, SE, t, F, and R-squared answer effect size, uncertainty, single-coefficient tests, overall model signal, and in-sample explanation.

Why Do Ridge, Lasso, and Weight Decay Make Models More Stable?

Regularization adds a preference against extreme parameters. Ridge, Lasso, and weight decay trade some training fit for a model that generalizes more reliably.

How Can Statistics and ML Evaluation Be Rerun to Reach the Same Conclusion?

A reproducible workflow preserves the evidence chain from data to conclusion. Results need data versions, code, seeds, environment, metrics, and raw outputs.

Why Can a Sample Say Something About a Population or Model?

Sampling makes sample statistics fluctuate, and standard error describes that fluctuation. This post separates SD, SE, sampling distributions, and CLT, then connects them to benchmark uncertainty.

How Do Sampling Distributions Become Exam-Ready Reasoning?

A sampling distribution describes how a statistic fluctuates under repeated sampling. Means, proportions, and variances each connect to common distributions used in intervals and tests.

After 53 Posts, How Do You Connect Statistics to ML, Causality, and Mathematical Statistics?

The series does not finish all of statistics. It gives beginners a working map for exams, ML/AI evaluation, causality, Bayesian thinking, time series, and mathematical statistics.

How Does One Regression Line Become Prediction, Interpretation, and Error?

Simple linear regression uses one X to describe the average change in Y. Slope, intercept, residuals, and squared error form the smallest supervised learning model.

How Does Monte Carlo Use Repeated Simulation to Answer Hard Statistical Questions?

Monte Carlo repeats a data-generating process many times so sampling variation, power, coverage, and evaluation instability become visible.

Where Should You Start Statistics If You Need Exams and ML/AI?

Do not start statistics exam prep by memorizing formulas. Start with the sequence of data, probability, sampling, inference, regression, then connect those ideas to model evaluation, A/B testing, and uncertainty in ML/AI.

Why Should Time-Series Data Not Be Randomly Split?

Time-series data have order. Random splits can leak future information into training and make forecasting or monitoring results look better than they are.

Which Test Fits a Two-Group Mean or Proportion Difference?

Two-group comparisons start by classifying the outcome and the design: numeric or binary, independent or paired. That choice determines the standard error, test statistic, and conclusion.

How Does Variable Selection Avoid Memorizing the Training Data?

Variable selection is not only about choosing predictors. It is about avoiding noisy training-set wins that do not generalize.

Statistics Is Not Formula Memorization: What Is It Deciding?

The core of statistics is judgment: describe data, estimate unknowns, compare differences, inspect associations, and make decisions under uncertainty.

tech deep-dive

What Is GPUtw.ai? Taiwan GPU Cloud, Short-Rental Compute, and Researcher Workflows

GPUtw.ai is a Taiwan-based short-rental GPU cloud. Its main value is not maximum scale, but Taiwan data centers, prepaid credits, Jupyter/ComfyUI/Ollama/vLLM templates, Vault storage, and team billing. Public information is enough for a service introduction, not enough for procurement or production endorsement.

Harvard CS181 HW0: Do These 4 Problems First — They Tell You What to Patch

HW0 checks CS181 prerequisites in four problems — y=Xw solvability, optimizing an objective, reasoning about randomness, and OLS in Python. The problem that slows you down most is the gap to patch before HW1.

Harvard CS181 HW1: Ice Core Regression — Linear, Kernel, and Neural Nets in One Assignment

HW1 uses an 800k‑year ice‑core temperature dataset to implement three regression models (OLS, RBF kernel, MLP) and compare them on the same data, laying the groundwork for later classification and deep‑learning assignments.

Harvard CS181 Machine Learning: Your 2026 Roadmap Through 7 Homeworks (With a 4-Year Comparison)

CS181 2026 is A3 with hw0–6 as the weekly clock (no public recordings); 2025 adds a practical, 2024 has two midterms, 2023 was taught by Weiwei Pan. Start with HW0, then follow hw1→hw6.

Ahead of AI: How a Scholar Built 200K Subscribers by Publishing Monthly, Not Daily

Computational biology PhD turned UW-Madison professor Sebastian Raschka launched Ahead of AI on Substack in 2022, publishing monthly deep dives into LLM papers and architectures. Four years later: 200K+ subscribers, zero sponsorships, and a book-newsletter flywheel that proves low frequency and high depth can win in a crowded AI newsletter market.

2021 AI Conference Guide: Machine Learning

2021 was the year diffusion models surpassed GANs, self-supervised learning made theoretical breakthroughs, and reinforcement learning confronted weaknesses in its evaluation methodology. NeurIPS received a then-record 9,122 submissions, ICLR’s Score-Based Generative Modeling paper became a theoretical foundation for the diffusion ecosystem, and ICML delivered substantial work on optimization theory and the dynamics of self-supervised learning.

2022 AI Conference Guide: Machine Learning

2022 was the year diffusion models took center stage, Chinchilla scaling laws rewrote large-model training, and Chain-of-Thought turned reasoning into an ability that prompts could elicit. NeurIPS passed 10,000 submissions; three of its 13 Outstanding Papers directly concerned diffusion; and Chinchilla and data pruning both challenged the belief that bigger was always better. On the eve of ChatGPT’s release, every required piece fell into place at that year’s conferences.

A Guide to the Top AI Conferences of 2023: Machine Learning

In 2023, LLMs took over the machine-learning conference agenda. NeurIPS received more than 12,000 submissions; both Outstanding Papers addressed large models, while runner-up DPO became a practical alternative to RLHF within two years. DreamFusion opened the text-to-3D field, ICML spotlighted LLM watermarking and learning-rate adaptation, and the Mamba preprint emerged as the first serious architectural challenger to the Transformer.

2024 AI Conference Review: Machine Learning

ML conference submissions exploded in 2024: NeurIPS received a record 15,671 papers, while ICML and ICLR passed 9,000 and 7,000. Research shifted from training ever-larger models toward spending inference compute more intelligently, making test-time compute scaling the year's defining new direction. VAR beat diffusion with next-scale image prediction, Rectified Flow became the theoretical foundation for Stable Diffusion 3, and ICLR gave its inaugural Test of Time Award to the original VAE paper.

2025 AI Conference Review: Machine Learning

ML conferences broke every submission record in 2025 and pushed peer review to its limit. NeurIPS received 21,575 papers and used more than 20,000 reviewers; ICML passed 12,000 for the first time, and ICLR reached 11,565. Reasoning and agents were the strongest trends. One NeurIPS runner-up, the conference's only perfect-score paper, challenged whether RLVR creates new reasoning ability. Awards for Alibaba Qwen's Gated Attention and a mechanistic theory of neural scaling laws showed a community moving from scaling at all costs toward understanding why scaling works.

AI Engineer Interview Daily — 2026-08-24: ML Fundamentals

ML fundamentals interviews don't test whether you can recite definitions — they test whether you can walk through a structured diagnostic when handed a train/val accuracy gap. Today covers four high-frequency topics: bias-variance decomposition and learning curve interpretation, geometric intuition for L1/L2 regularization and when to pick which, aligning loss functions with business objectives instead of accepting defaults, and why AdamW decouples weight decay from L2 regularization.

ai guide Reading CMU 07-280

Completing CMU 07-280: What You Know, What Is Missing, and What Comes Next

Finishing 07-280 means more than reading 24 guides: produce a search engine, supervised-model comparison, CNN/GPT-2 experiments, and a small RL-plus-MCTS system before choosing 07-380, 10-301, or a specialist course.

Reading CMU 07-280: Why Search, GPT-2, and AlphaZero Belong in One Course

07-280 is CMU's new Spring 2026 AI+ML core: 24 lectures and 12 main assignments move from heuristic search and CSPs to AlexNet, GPT-2, and AlphaZero. Its public material supports self-study, but complete recordings, Canvas checkpoints, Gradescope, and staff feedback remain unavailable.

CMU 07-280 Lecture 1: The Shared Problem Behind AI, ML, and Representation Learning

Lecture 1 uses an alien autoencoder, the scope of AI and ML, and AI history to establish the course's coordinate system: an intelligent system turns inputs into representations and decisions under uncertainty.

CMU 07-280 Lecture 5: Defining Machine Learning with Loss, Risk, and ERM

Lecture 5 formulates machine learning through `X → Y`, loss, risk, and empirical risk minimization: a training set only gives average observed loss, while the real objective remains generalization over an unknown distribution.

CMU 07-280 Lecture 6: How Decision Trees Split Data with Mutual Information

Lecture 6 recursively grows a tree from decision stumps, measures label uncertainty with entropy, and selects splits by `I(Y;W)=H(Y)-H(Y|W)`; this is computationally practical greedy ERM, not a global optimal-tree guarantee.

CMU 07-280 Lecture 7: Linear Regression and the Normal Equation

Lecture 7 applies ERM to linear functions and squared loss, moves from a one-dimensional slope to `argmin ||y-Xθ||²`, and derives the normal equation when `XᵀX` is invertible.

CMU 07-280 Lecture 8: Gradient Descent, SGD, and Learning Rate

Lecture 8 moves from a one-dimensional parabola to vector gradients and compares batch GD, SGD, and mini-batches; the learning rate determines whether updates converge, oscillate, or diverge.

CMU 07-280 Lecture 9: Logistic Regression as Probability Estimation

Lecture 9 models P(y=1|x) with a sigmoid instead of directly predicting 0 or 1, learns parameters with cross-entropy and convex optimization, and extends naturally to softmax regression.

CMU 07-280 Lecture 10: Trading Expressiveness for Stability with Features and Regularization

Lecture 10 uses φ(x) to let linear models express nonlinear functions, then controls the resulting overfitting with train/validation/test separation, L1/L2 regularization, and model selection.

CMU 07-280 Lecture 11: Building a Neural Network from Logistic Regression

Lecture 11 expands a logistic unit into a multilayer network: linear layers produce z, activations produce a, and multiple neurons jointly learn a feature transform trained through a final loss.

CMU 07-280 Lecture 12: How Backpropagation Reuses the Chain Rule

Lecture 12 treats a network as a computation graph: the forward pass stores intermediates, the backward pass propagates upstream gradients, and local linear, activation, and softmax rules compute every parameter gradient efficiently.

CMU 07-280 Lecture 16: Unifying Logistic and Linear Regression with Maximum Likelihood

Lecture 16 starts from likelihood p(D|θ), uses i.i.d. to factor the joint probability and logs to turn products into sums; Bernoulli MLE yields sample proportions, conditional Bernoulli yields logistic cross-entropy, and Gaussian noise yields squared error.

CMU 07-280 Stage Review I: From Search Problems to Supervised Learning

Lectures 1–12 form one decision pipeline: define states, moves, and objectives, then use heuristics, losses, regularization, and backpropagation to control an otherwise intractable search space.

ai deep-dive

DSPy: Compiling AI Programs with Signatures, Metrics, and Optimizers

DSPy replaces handwritten prompt strings with task Signatures, execution Modules, and Optimizers that compile better instructions and examples against a dataset and metric.

ai deep-dive

Hugging Face Is More Than a Model Download Site: Hub, Datasets, Spaces, and Inference

Hugging Face Hub is a collaboration layer for versioned models, datasets, and applications. Datasets handles data, Spaces runs demos, while Inference Providers and Endpoints provide managed inference.

Linear Regression: From LMS to Locally Weighted Regression

Linear regression is more than a best-fit line: Chapter 1 connects squared loss to gradient descent, normal equations, maximum likelihood, and locally weighted regression.

Classification and Logistic Regression: Decision Boundaries and Newton's Method

Chapter 2 derives logistic loss from a sigmoid probability model, then contrasts it with the perceptron and extends it through softmax and Newton's method.

Generalized Linear Models: Unifying Regression and Classification

Chapter 3 uses exponential families, natural parameters, and link functions to place least squares and logistic regression inside one modeling template.

Generative Learning Algorithms: GDA, Naive Bayes, and Smoothing

Chapter 4 models p(x|y) and p(y), using GDA, Naive Bayes, and Laplace smoothing to expose both the power and price of generative classification.

Kernel Methods: Nonlinear Learning Without Explicit Features

Chapter 5 replaces high-dimensional feature inner products with kernels, letting inner-product-based linear algorithms learn nonlinear functions without constructing the features.

Support Vector Machines: Margins, Duality, and SMO

Chapter 6 formalizes classification confidence as geometric margin, then builds an implementable SVM through Lagrange duality, kernels, and SMO.

Deep Learning: Modules, Backpropagation, and Vectorization

Chapter 7 decomposes neural networks into composable modules and uses backpropagation and vectorization to explain how deep models can be trained efficiently.

Generalization: Bias–Variance, Double Descent, and Sample Complexity

Chapter 8 decomposes test MSE into irreducible noise, squared bias, and variance, then uses uniform convergence and VC dimension to explain when training performance transfers to new data. Double descent shows why parameter count is not a universal measure of complexity.

Regularization and Model Selection: Explicit, Implicit, and Cross-Validated

Chapter 9 presents three controls on generalization: explicit complexity penalties, optimizer-induced implicit regularization, and model selection on data excluded from training. MAP estimation then connects a Gaussian prior to an L2 penalty.

Clustering and k-Means: A First Alternating-Optimization Algorithm

Chapter 10 introduces unsupervised learning through k-means: alternating updates make distortion non-increasing and numerically convergent, but do not guarantee a global optimum.

EM Algorithms: From Gaussian Mixtures to VAEs

Chapter 11 starts from soft assignments in Gaussian mixtures, uses Jensen's inequality to construct the ELBO, interprets EM as alternating maximization over a variational distribution and model parameters, and extends the idea to VAEs through approximate posteriors and reparameterization.

Principal Components Analysis: Projection, Reconstruction, and Reduction

Chapter 12 formulates PCA as geometric optimization: maximize projected variance along a unit direction to obtain the leading eigenvector of the covariance matrix. The top k eigenvectors give both maximum retained variance and minimum linear reconstruction error.

Independent Components Analysis: Recovering Independent Sources

Chapter 13 models ICA as x=As: observations are unknown linear mixtures, and the goal is to estimate W=A^{-1} to recover independent, non-Gaussian sources. A Jacobian determinant enters the transformed density and leads to the Bell–Sejnowski likelihood update.

Diffusion Models: Forward Noise, Reverse Generation, and the ELBO

Chapter 14 starts with a fixed Gaussian noising Markov chain and learns to reverse each transition. The ELBO turns reverse-kernel matching into weighted noise prediction, while the continuous-time view explains reverse drift through the score ∇log p_t.

CS188 Decisions and Machine Learning: From VPI and Naive Bayes to Attention

Lectures 19–25 connect rational decisions and VPI to machine learning, while Project 5 uses PyTorch for regression, classification, CNNs, attention, and an optional character-GPT.

Berkeley CS189 Spring 2025 Overview: HW1–7 with Code and Data You Can Run, Plus What Fall 2026 Looks Like

Spring 2025 at people.eecs.berkeley.edu/~jrs/189s25 is the only A3 self-study edition with notes, videos, HW1–7, code/data and past exams; Fall 2026 at eecs189.org/fa26 has a 27-lecture schedule but most materials are not yet released and the rotating site can 404 old URLs.

CMU 07-380 Fall 2026 Overview: 26 Lectures from Logic and Planning to Diffusion, HW and Project Not Yet Fully Released

07-380 Fall 2026 is the first offering of CMU's new AI II, 26 lectures from logic, planning and optimization to probabilistic graphs and generative systems; Lec01 and Prop Logic are public, HW1-7, six quizzes and the final project are still TBD — an A2→A3 transition with the 07-280 bridge.

CMU 10-301 HW1: Find ML Foundation Gaps with Mathematics and Python

HW1 is written and programming work: mathematical and CS foundations followed by a majority-vote classifier.

CMU 10-301 HW2: From Information Calculations to a Complete Decision Tree

HW2 moves from hand-calculated entropy and mutual information to an end-to-end tree learner, predictor, and evaluator.

CMU 10-301 HW3: Compare K-NN, Perceptron, and Linear Regression

HW3 is written work: a decision-tree review followed by K-NN, Perceptron, and Linear Regression through inductive bias, errors, and model selection.

CMU 10-301 HW4: Turn Logistic Regression Likelihood into a Classifier

HW4 joins probabilistic interpretation, cross-entropy gradients, and implementation into one traceable training pipeline.

CMU 10-301 HW5: Expose Neural Networks and Backpropagation with NumPy

HW5 avoids automatic differentiation so learners must track forward shapes, caches, and backward gradients themselves.

CMU 10-301 HW6: Learning Theory, MLE/MAP, and Fairness Metrics

HW6 combines generalization, MLE/MAP, probabilistic learning, fairness metrics, and social impact in one written assignment about assumptions and tradeoffs.

CMU 10-301 HW7: Move from Basic Neural Networks to Deep Learning

HW7 builds on HW5 backpropagation to address deep-model architecture and training failures, emphasizing diagnosis over merely adding layers.

CMU 10-301 HW8: From MDPs to Reinforcement-Learning Updates

HW8 connects states, actions, rewards, transitions, and value updates while separating environment dynamics, policy, and estimation error.

CMU 10-301 HW9: Close the Course with Ensembles, k-Means, PCA, and Recommenders

The final written assignment combines ensembles, clustering, representation, and recommendation to test whether you can choose a learning paradigm from problem structure.

CMU 10-301/601 Spring 2026: Learn Machine Learning Through Nine Assignments

Spring 2026 publishes material for 27 lectures and nine homework bundles; outsiders can do the core work but cannot access Panopto, Piazza, Gradescope, or official homework solutions.

learning deep-dive

CMU's AI Core Redesign: From 15-281 + 10-315 to 07-280 + 07-380

In 2026, CMU recombined its separate general-AI and SCS machine-learning introductions into the 07-280 → 07-380 sequence. This is a redistribution of content and prerequisites, not a pair of simple course renames.

Harvard AI/ML Course Guide: Do CS50 AI, CS181, and CS182 Videos Match Their Assignments?

CS50 AI is Harvard's most complete public entry point, but the Summer 2026 course still uses 2020 recordings and assignment assets while the rolling OCW projects have moved to other editions. CS181 Spring 2026 exposes current homework and notes without current recordings; CS182 Fall 2026 has not yet completed an offering.

Stanford CS109 Lecture 20 | Logistic Regression: Derive the gradient from Bernoulli likelihood

Logistic regression turns a linear score into a Bernoulli probability with sigmoid; the gradient xⱼ(y-ŷ) follows directly from the log-likelihood chain rule.

Stanford CS109 Lecture 21 | Comparing Classifiers: Beyond accuracy to calibration, error costs, and fairness

Classifier comparison requires held-out data, baselines, calibration, precision/recall, and an explicit fairness criterion—not accuracy alone.

Lambda Cloud: GPU Compute from One VM to Multi-Node AI Clusters

Lambda Cloud provides on-demand GPU VMs and 1-Click Clusters; it offers direct AI compute environments rather than automatically solving training, serving, and MLOps.

Replicate: Turn Model Versions into Prediction APIs Instead of Renting GPUs

Replicate abstracts GPUs behind versioned models, predictions, Cog, and deployments; integrators still own version pinning, async workflows, webhook verification, data persistence, and spending limits.

RunPod: GPU Pods and Serverless Endpoints Are Different Products

RunPod Pods fit interactive and persistent GPU work, while Serverless fits queued or load-balanced inference; choosing incorrectly mixes persistence, cold starts, and retry semantics.

Stanford CS229: Notes Rewritten Every Year, Public Problem Sets Frozen at 2020, and an Official Self-Test From 2008

The three things you need to self-study CS229 run on three different clocks. The lecture notes are 278 pages and were recompiled in August 2026. The newest problem sets you can download are from summer 2020. The self-assessment Stanford Online tells you to attempt before enrolling is a PDF created in 2008. Seventeen lectures from spring 2026 are public, and the last three are mislabeled.

Berkeley AI/ML Course Guide: From CS61A to CS288, What Can You Actually Study Online?

Berkeley has no standalone undergraduate AI degree. A workable path builds on the CS BA or EECS BS foundation, enters through either CS188's broad AI curriculum or CS189's mathematical machine learning curriculum, then branches into deep learning, NLP, vision, or reinforcement learning. Many 2025–2026 courses are A3, but the newest class, the newest stable URL, and the best self-study edition are not always the same.

CMU AI/ML Course Guide: The New 07-280 Core and a Public Self-Study Route

CMU's current BSAI now runs through 07-280 and 07-380 before branching into an NLP/vision core and four AI clusters, but 07-380 does not debut until Fall 2026. The residual Spring 2026 materials for 07-280 and the complete 10-301/601 site already support self-study; retired 15-281 remains a useful legacy route.

MIT AI/ML Course Guide: Course 6-4 Is a Real AI Degree, but Its Public Materials Span Three Eras

MIT has offered Course 6-4, a formal BS in Artificial Intelligence and Decision Making, since 2022. For an outside learner, however, the current degree requirements, the 2025–2026 course sites, and the best OCW editions rarely line up. A workable route follows 6-4's programming, algorithms, linear algebra, and probability foundation, then selects among 6.S191, 6.3900, 6.4110, 6.7960, vision, and robotics according to what is actually public.

Stanford CS109: A Probability Course That Turned "How to Read This Lecture With an LLM" Into Official Coursework

Every lecture in CS109's Summer 2026 offering ships with an official LLM Learning Guide — six concepts, a Learn prompt and a Test me prompt for each, written week by week across the quarter for a total of 23 PDFs. The same course's honor code Rule 4 forbids asking an LLM to solve your homework, and 65% of the grade sits in proctored exam rooms. Those two facts are halves of one design.

Stanford CS161 Lecture 18: From the Algorithmic Toolbox to LP, Coding, and ML

The finale recaps the CS161 toolbox and points toward LP duality, Reed–Solomon coding, and ML-assisted algorithms. Officially, this lecture has slides but no notes.

ML Fundamentals Interview Guide: From Bias-Variance to Evaluation Metrics

ML fundamentals interviews don't test formula memorization — they test whether you can explain concepts intuitively and hold up under follow-up questions. High-frequency topics: the practical meaning of bias-variance tradeoff, the selection logic for L1/L2 regularization, why cross-entropy beats MSE for classification, SGD vs. Adam tradeoffs, and how precision/recall priorities differ by scenario.

AI Engineer Interview Overview: From Company Types to Preparation Strategy

AI Engineer interviews go beyond ML — big tech emphasizes system design and coding, startups look for end-to-end delivery, and AI-native companies test LLM engineering depth. Strategy: identify your target company types first, then allocate prep time across six dimensions (ML fundamentals, system design, LLM applications, coding, paper reading, and behavioral).

Preparing for Google PMLE After the Exam Guide Rewrite

Google's Professional ML Engineer exam guide was rewritten in 2026: Vertex AI is renamed Gemini Enterprise Agent Platform throughout, so older study material no longer matches the product names in the questions. This guide uses the official six-section weighting as its skeleton, listing what each section tests, which official materials cover it, and what to build — plus a study schedule whose reasoning is spelled out. Official specs: $200, two hours, 50–60 multiple-choice and multiple-select questions, two-year validity, 3+ years of industry experience recommended including 1+ year on Google Cloud.

AI Project Strategy: Three Hours in a Spreadsheet Buys Back Weeks

Andrew Ng demonstrates error analysis on a deep researcher: columns are the pipeline stages, rows are 10 to 100 queries, you only look at the ones that went badly, and you mark each cell where something broke. The percentages don't have to sum to 100%. He says it takes three or four hours and saves weeks of going the wrong direction — and the fraction of people who actually do it is far below 100%.

Full Cycle of a DL Project: You Get Two Days to Collect Data

Andrew Ng walks a face-recognition door system through the entire project lifecycle, and the whole lecture has one thesis: speed. He gives teams a two-day deadline, on the reasoning that 'time spent preparing data should be commensurate with the time it takes to train the model once.' It closes on a line: my job is to build something that actually works, and that is not the same as building something that works on the test set.

A Flight Controller's Autonomy Has 56 Entries, One About Opportunity

ArduPilot's ModeReason enum is the exhaustive list of on-board autonomous decisions: of 56 values, 43 are the aircraft deciding for itself, 21 of those because it detected something dangerous, and exactly one — SOARING_THERMAL_DETECTED — because it found an opportunity. As for end-to-end, PX4 mainline's mc_nn_control has a 10 KB tensor arena and a Kconfig default of n; ArduPilot mainline has none.