느린 학습
A lower learning rate may need more epochs.
브라우저 실제 학습외부 API 없는 로컬 교육용 실험입니다. 합성 데이터는 운영 환경 측정이 아닙니다.
Study guide & lesson outline
LEARNING CONTRACTUnderstand → implement → investigate
Goal: explain the mechanism in this lesson, reproduce one example and test a failure case. Reading alone does not mark mastery.
- Understand: state assumptions, shapes/units and a counterexample.
- Implement: reproduce a numerical example and inspect boundaries.
- Investigate: compare one alternative under a matched budget.
02 · 예측 및 설정
기술 설명, 매개변수 이름과 참고 답변은 현재 영어입니다.
03 · 근거 확인
방법과 참고 문헌
Actual full-batch gradient descent in JavaScript. 2 inputs → tanh hidden layer → sigmoid. Binary cross-entropy; 100 synthetic training and 100 held-out samples from the same generator. Held-out accuracy is not real-world generalization.
Text / 原始公式
h=tanh(Wx+b); p=σ(vᵀh+c); L=−mean[y log p+(1−y)log(1−p)]
Training convergence on tiny synthetic data does not imply an LLM can train in the browser.
Dive into Deep Learning · multilayer perceptrons ↗연습 문제
neural-3 / 01What does “Slow learning” test?
A lower learning rate may need more epochs.
Original teaching answer / local derivation. Method reference: Dive into Deep Learning · multilayer perceptrons ↗
실험 열기 →neural-3 / 02Which computation produces the result in “Slow learning”?
Original teaching answer / local derivation. Method reference: Dive into Deep Learning · multilayer perceptrons ↗
실험 열기 →neural-3 / 03What is the baseline configuration for “Slow learning”?
{"rate":0.01,"hidden":8,"epochs":1000,"task":0,"seed":42}
Original teaching answer / local derivation. Method reference: Dive into Deep Learning · multilayer perceptrons ↗
실험 열기 →neural-3 / 04What numerical result should you check for “Slow learning”?
Run the seeded training lab and inspect the actual loss and accuracy; convergence is not assumed.
Original teaching answer / local derivation. Method reference: Dive into Deep Learning · multilayer perceptrons ↗
실험 열기 →neural-3 / 05How would you run a controlled comparison for “Slow learning”?
Keep the supplied fixture and other settings fixed; change rate. Watch train loss, held-out accuracy and decision boundary over epochs.
Original teaching answer / local derivation. Method reference: Dive into Deep Learning · multilayer perceptrons ↗
실험 열기 →neural-3 / 06What failure could invalidate a conclusion from “Slow learning”?
Training convergence on tiny synthetic data does not imply an LLM can train in the browser.
Original teaching answer / local derivation. Method reference: Dive into Deep Learning · multilayer perceptrons ↗
실험 열기 →neural-3 / 07Is “Slow learning” a benchmark of a deployed language model?
No. Actual full-batch gradient descent in JavaScript. 2 inputs → tanh hidden layer → sigmoid. Binary cross-entropy; 100 synthetic training and 100 held-out samples from the same generator. Held-out accuracy is not real-world generalization.
Original teaching answer / local derivation. Method reference: Dive into Deep Learning · multilayer perceptrons ↗
실험 열기 →neural-3 / 08Which evidence should you report for “Slow learning”?
Watch train loss, held-out accuracy and decision boundary over epochs. Export inputs, engine, result and method; distinguish synthetic fixtures from observed production data.
Original teaching answer / local derivation. Method reference: Dive into Deep Learning · multilayer perceptrons ↗
실험 열기 →neural-3 / 09How can another researcher reproduce “Slow learning”?
Open neural-3, reset the preset, retain the exact inputs and run. Retain the seed, training epochs and all weights.
Original teaching answer / local derivation. Method reference: Dive into Deep Learning · multilayer perceptrons ↗
실험 열기 →neural-3 / 10What is the next validation step after “Slow learning”?
Training convergence on tiny synthetic data does not imply an LLM can train in the browser. Test the hypothesis on a representative held-out workload and document a baseline, success metric and stopping rule before claiming deployment readiness.
Original teaching answer / local derivation. Method reference: Dive into Deep Learning · multilayer perceptrons ↗
실험 열기 →Assumptions, implementation & acceptance
XOR labels are 1 exactly when the signs differ. Stable logits BCE is max(z,0)−zy+log1p(exp(−|z|)); its logit gradient is sigmoid(z)−y. Training and held-out sets are separately seeded samples, not real-world validation.
Engine formulas, scenario presets and numerical safeguards reviewed. Seeded synthetic inputs are not a production benchmark.
Before running: predict a boundary case
A lower learning rate may need more epochs.
Hint: isolate the assumption
Keep the baseline input fixed, change one parameter or boundary, and compare both the numerical result and the claimed units. A passing example verifies this fixture only.
Minimal executable example
Save the example and linked modules in one folder. Use Node.js 22 or newer, then run node neural-3.mjs. No third-party runtime dependencies.
import {calculate,network} from './academy-core.mjs';
const params={
"rate": 0.01,
"hidden": 8,
"epochs": 1000,
"task": 0,
"seed": 42
};
const result=network(params).step(params.epochs);
console.log(result.metrics);Test and transfer
A lower learning rate may need more epochs.
Record input, code version, expected result and observed result in the notebook below. Then construct a counterexample that violates the assumption above and explain why the original conclusion no longer follows.
Download complete kit (.tar.gz) · Complete reference kit and regression checksPredict, test and keep evidence
Before revealing an answer, predict one result. After the exercise, record the discrepancy and a transfer test. Saved locally across language editions.
Open notebook / export / clear →The mathematical review scope is recorded per lesson. It does not establish empirical paper reproduction; report suspected errors in your notes. Review scope