Focus on Learning Machine Learning That Directly Connects to Dating

Focus on Learning Machine Learning That Directly Connects to Dating​

It’s certainly possible to learn computer science on your own, without attending a university.

However...

First, Learn Mathematics​

If you are thinking of studying computer science, I want to emphasize one crucial point: you should first focus on learning mathematics thoroughly. Mathematics is more than just a set of calculation techniques; it’s a vital tool for developing logical thinking and problem-solving skills, which form the foundation of computer science. Even if someone tells you, "You don’t need math," or "Math is too hard, so you should avoid it," you should not listen to such advice. In fact, following that kind of thinking might limit your learning path in the long run.

If you look at university computer science curricula, you’ll see that a significant portion of the courses—often the ones with the most hours—are math-related. Understanding core computer science concepts such as programming, data structures, and algorithms requires a solid grasp of mathematics. For example, evaluating the efficiency of an algorithm demands mathematical analysis skills, and fields like database design or machine learning are fundamentally built upon mathematical theories.

That’s why taking the time to study mathematics deeply is actually the most effective way to learn computer science both efficiently and profoundly. A strong foundation in mathematics equips you with abstract thinking skills and the ability to approach complex problems calmly and logically. These abilities will help you navigate and overcome the challenges you’ll encounter in computer science.

Ultimately, learning mathematics is not just about acquiring technical skills—it’s a major step toward becoming a more creative and capable engineer or researcher by broadening your thinking and perspective.

That said, there is one crucial thing you should never lose sight of as you study mathematics and computer science.

And that is...

What Really Matters Is Connecting Your Learning to Dating​

There’s a big gap between simply learning Machine Learning and truly connecting that learning to dating.

Learning Computer Science on your own is often enough—but trying to figure out how to connect that learning to dating all by yourself is, in fact, a waste of effort when you consider the essence of dating itself. We hope you’ll realize this as early in life as possible.

For now, we’ve gathered some resources for you below. But if you’re interested in focusing on learning that directly connects to dating, please check the guidance at the bottom of this page.

Learning Resource​

Machine Learning (機械学習) を網羅的に体系的に効率的に学ぶためのカリキュラム​

Definition and examples of a broad variety of machine learning tasks​

• a. Supervised learning
  • i. Classification
  • ii. Regression
• b. Reinforcement learning
• c. Unsupervised learning
  • i. Clustering

Fundamental ideas:​

• a. No free lunch theorem: no one learner can solve all problems; representational design decisions have consequences.
• b. Sources of error and undecidability in machine learning

A simple statistical-based supervised learning such as linear regression or decision trees​

• a. Focus on how they work without going into mathematical or optimization details; enough to understand and use existing implementations correctly

The overfitting problem/controlling solution complexity (regularization, pruning – intuition only)​

• a. The bias (underfitting) – variance (overfitting) tradeoff

Working with Data​

• a. Data preprocessing
  • i. Importance and pitfalls of preprocessing choices
• b. Handling missing values (imputing, flag-as-missing)
  • i. Implications of imputing vs flag-as-missing
• c. Encoding categorical variables, encoding real-valued data
• d. Normalization/standardization
• e. Emphasis on real data, not textbook examples

Representations​

• a. Hypothesis spaces and complexity
• b. Simple basis feature expansion, such as squaring univariate features
• c. Learned feature representations

Machine learning evaluation​

• a. Separation of train, validation, and test sets
• b. Performance metrics for classifiers
• c. Estimation of test performance on held-out data
• d. Tuning the parameters of a machine learning model with a validation set
• e. Importance of understanding what a model is doing, where its pitfalls/shortcomings are, and the implications of its decisions

Basic neural networks​

• a. Fundamentals of understanding how neural networks work and their training process, without details of the calculations
• b. Basic introduction to generative neural networks (e.g., large language models)

Ethics for Machine Learning​

• a. Focus on real data, real scenarios, and case studies
• b. Dataset/algorithmic/evaluation bias and unintended consequences

Formulation of simple machine learning as an optimization problem, such as least squares linear regression or logistic regression​

• a. Objective function
• b. Gradient descent
• c. Regularization to avoid overfitting (mathematical formulation)

Ensembles of models​

• a. Simple weighted majority combination

Deep learning​

• a. Deep feed-forward networks (intuition only, no mathematics)
• b. Convolutional neural networks (intuition only, no mathematics)
• c. Visualization of learned feature representations from deep nets
• d. Other architectures (generative NN, recurrent NN, transformers, etc.)

Performance evaluation​

• a. Other metrics for classification (e.g., error, precision, recall)
• b. Performance metrics for regressors
• c. Confusion matrix
• d. Cross-validation
  • i. Parameter tuning (grid/random search, via cross-validation)

Overview of reinforcement learning methods​

Two or more applications of machine learning algorithms​

• a. E.g., medicine and health, economics, vision, natural language, robotics, game play

General statistical-based learning, parameter estimation (maximum likelihood)​

Supervised learning​

• a. Decision trees
• b. Nearest-neighbor classification and regression
• c. Learning simple neural networks / multi-layer perceptrons
• d. Linear regression
• e. Logistic regression
• f. Support vector machines (SVMs) and kernels
• g. Gaussian Processes

Overfitting​

• a. The curse of dimensionality
• b. Regularization (mathematical computations, L2 and L1 regularization)

Experimental design​

• a. Data preparation (e.g., standardization, representation, one-hot encoding)
• b. Hypothesis space
• c. Biases (e.g., algorithmic, search)
• d. Partitioning data: stratification, training set, validation set, test set
• e. Parameter tuning (grid/random search, via cross-validation)
• f. Performance evaluation
  • i. Cross-validation
  • ii. Metric: error, precision, recall, confusion matrix
  • iii. Receiver operating characteristic (ROC) curve and area under ROC curve

Bayesian learning (Cross-Reference AI/Reasoning Under Uncertainty)​

• a. Naive Bayes and its relationship to linear models
• b. Bayesian networks
• c. Prior/posterior
• d. Generative models

Deep learning​

• a. Deep feed-forward networks
• b. Neural tangent kernel and understanding neural network training
• c. Convolutional neural networks
• d. Autoencoders
• e. Recurrent networks
• f. Representations and knowledge transfer
• g. Adversarial training and generative adversarial networks
• h. Attention mechanisms

Representations​

• a. Manually crafted representations
• b. Basis expansion
• c. Learned representations (e.g., deep neural networks)

Unsupervised learning and clustering​

• a. K-means
• b. Gaussian mixture models
• c. Expectation maximization (EM)
• d. Self-organizing maps

Graph analysis (e.g., PageRank)​

Semi-supervised learning​

Graphical models (See also: AI-Probability)​

Ensembles​

• a. Weighted majority
• b. Boosting/bagging
• c. Random forest
• d. Gated ensemble

Learning theory​

• a. General overview of learning theory / why learning works
• b. VC dimension
• c. Generalization bounds

Reinforcement learning​

• a. Exploration vs exploitation tradeoff
• b. Markov decision processes
• c. Value and policy iteration
• d. Policy gradient methods
• e. Deep reinforcement learning
• f. Learning from demonstration and inverse RL

Explainable / interpretable machine learning​

• a. Understanding feature importance (e.g., LIME, Shapley values)
• b. Interpretable models and representations

Recommender systems​

Hardware for machine learning​

• a. GPUs / TPUs

Application of machine learning algorithms to:​

• a. Medicine and health
• b. Economics
• c. Education
• d. Vision
• e. Natural language
• f. Robotics
• g. Game play
• h. Data mining (Cross-reference DM/Data Analytics)

【Important】 For Those Who Want to Enjoy a Life Without Struggling to Find Dating Partners​

There are many people out there who, despite learning Machine Learning, waste their time by studying in a way that fails to connect their knowledge to creating meaningful shared experiences through dating.

We, on the other hand, have dedicated ourselves to learning in ways that truly connect to creating meaningful shared experiences through dating, and as a result, we’ve been able to enjoy a life where we never struggle to find dating partners.

Would you like to join us at the Dating University and focus on learning that directly connects to creating meaningful shared experiences through dating—and enjoy a life where dating partners are never in short supply?

Begin Your Journey Today​

"I want to live a life where finding a date is never a struggle."
If that’s how you feel — yet your daily life still feels stuck or unfulfilled — you’re not alone.
When even a hint of doubt or unease lingers in your heart, that’s the perfect time to reach out.
We invite you to take that first step.
Let us help you explore how the right kind of learning can turn dating into something natural — and transform your everyday life.