File size: 15,263 Bytes
777489b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a69f3ba
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
<!DOCTYPE html>
<html lang="en">
<head>
  <meta charset="UTF-8" />
  <meta name="viewport" content="width=device-width, initial-scale=1.0" />
  <title>Data Labeling Tic Tac Toe</title>
  <script src="https://cdn.tailwindcss.com"></script>
</head>
<body class="bg-gradient-to-br from-indigo-50 via-sky-50 to-teal-100 min-h-screen flex items-center justify-center p-4">
  <div class="max-w-6xl w-full bg-white rounded-2xl shadow-xl p-6">
    <h1 class="text-3xl font-extrabold mb-3 text-center text-indigo-800">Data Labeling Tic Tac Toe</h1>
    <p class="text-center mb-2 text-sm text-gray-600">
      Reference: <a href="https://www.linkedin.com/pulse/data-labelling-michael-lively-ofhle/" class="text-indigo-600 hover:underline" target="_blank" rel="noopener noreferrer">Data Labelling: The Ghost in the Machine</a>
    </p>
    <p class="text-center mb-4 text-sm text-gray-500">
      Answer correctly to claim a square. Miss the question, and your turn passes.
    </p>

    <div id="banner" class="hidden bg-indigo-100 border border-indigo-300 text-indigo-800 font-bold text-center py-2 rounded mb-4"></div>

    <div class="flex flex-col lg:flex-row items-center lg:items-start justify-center gap-5">
      <!-- Game Board -->
      <div id="board" class="grid grid-cols-3 gap-2 border-4 border-indigo-300 rounded-xl p-3 bg-indigo-50"></div>

      <!-- Question Panel -->
      <div id="questionPanel" class="w-full lg:w-1/2 bg-gray-50 rounded-xl shadow p-5 min-h-80">
        <h2 id="panelQuestion" class="text-xl font-semibold text-gray-800">Select a square to view the question</h2>
        <ul id="panelChoices" class="mt-4 space-y-2"></ul>
        <p id="panelHint" class="mt-3 text-sm text-gray-600 hidden"></p>
        <button id="hintBtn" class="mt-4 px-3 py-1 bg-gray-200 text-gray-700 rounded hover:bg-gray-300">Show Hint</button>
      </div>
    </div>

    <div class="flex flex-col sm:flex-row justify-between items-center gap-3 mt-6">
      <p id="status" class="text-lg font-medium text-gray-800"></p>
      <button id="restartBtn" class="px-4 py-2 bg-indigo-600 text-white rounded hover:bg-indigo-700">Restart</button>
    </div>
  </div>

  <script>
    const questions = [
      {
        question: 'What is the main purpose of data labeling in supervised machine learning?',
        choices: [
          'To compress datasets before training',
          'To give models labeled examples they can learn from',
          'To remove all categorical data from a dataset',
          'To increase the number of model parameters'
        ],
        answer: 2,
        hint: 'Labels provide the ground truth used during training.'
      },
      {
        question: 'Which type of image labeling assigns a single class to an entire image?',
        choices: [
          'Image classification',
          'Object detection',
          'Semantic segmentation',
          'Speaker identification'
        ],
        answer: 1,
        hint: 'The whole image receives one label.'
      },
      {
        question: 'Which image labeling method uses bounding boxes around objects?',
        choices: [
          'Image classification',
          'Object detection',
          'Sentiment analysis',
          'Label encoding'
        ],
        answer: 2,
        hint: 'The model must learn both what the object is and where it is.'
      },
      {
        question: 'Semantic segmentation labels what part of an image?',
        choices: [
          'Only the image filename',
          'Only the largest object',
          'Every pixel with a class',
          'Only the image metadata'
        ],
        answer: 3,
        hint: 'This is more detailed than a bounding box.'
      },
      {
        question: 'Named Entity Recognition is used to label which kind of information?',
        choices: [
          'Names, dates, organizations, and locations in text',
          'Pixels in an image',
          'Temperature values in a spreadsheet',
          'Audio volume levels'
        ],
        answer: 1,
        hint: 'NER is a text labeling task.'
      },
      {
        question: 'What does Inter-Annotator Agreement measure?',
        choices: [
          'How fast a model trains',
          'How much different annotators agree on labels',
          'How many GPUs are used',
          'How large a dataset is after compression'
        ],
        answer: 2,
        hint: 'It checks consistency between human labelers.'
      },
      {
        question: 'Why is high Inter-Annotator Agreement important?',
        choices: [
          'It means the labels are likely more consistent and dependable',
          'It guarantees the model will never fail',
          'It removes the need for preprocessing',
          'It makes numerical scaling unnecessary'
        ],
        answer: 1,
        hint: 'Models learn better from consistent labels.'
      },
      {
        question: 'Which agreement metric is especially useful because it can handle missing labels and different numbers of annotators?',
        choices: [
          'Percent agreement',
          'Krippendorff\'s Alpha',
          'Min-max scaling',
          'One-hot encoding'
        ],
        answer: 2,
        hint: 'It is one of the most robust agreement measures.'
      },
      {
        question: 'Why can percent agreement be misleading as a labeling quality metric?',
        choices: [
          'It ignores agreement that may happen by chance',
          'It only works on image files',
          'It requires ratio data',
          'It removes missing labels automatically'
        ],
        answer: 1,
        hint: 'Some annotators can agree randomly.'
      },
      {
        question: 'Which data type has ordered values but no true zero point?',
        choices: [
          'Nominal data',
          'Interval data',
          'Binary data',
          'Object data'
        ],
        answer: 2,
        hint: 'Temperature in Celsius is a common example.'
      },
      {
        question: 'Which data type has a meaningful zero and supports meaningful ratios?',
        choices: [
          'Ratio data',
          'Nominal data',
          'Ordinal data',
          'Interval data'
        ],
        answer: 1,
        hint: 'Height, weight, and revenue are examples.'
      },
      {
        question: 'Movie ratings such as poor, fair, good, and excellent are examples of what data type?',
        choices: [
          'Nominal data',
          'Ordinal data',
          'Ratio data',
          'Continuous data'
        ],
        answer: 2,
        hint: 'The categories have a meaningful order.'
      },
      {
        question: 'One-hot encoding converts categorical data into what format?',
        choices: [
          'A binary vector',
          'A text paragraph',
          'A bounding box',
          'A continuous audio signal'
        ],
        answer: 1,
        hint: 'Each category becomes a separate 0/1 indicator.'
      },
      {
        question: 'Why is scaling numerical data important before modeling?',
        choices: [
          'It prevents large-range features from dominating the model',
          'It removes all labels from the dataset',
          'It changes text into images',
          'It guarantees perfect accuracy'
        ],
        answer: 1,
        hint: 'Think about a feature ranging from 0 to 1 versus one ranging from 0 to 1000.'
      },
      {
        question: 'Standardization usually transforms numerical data to have what?',
        choices: [
          'A mean of 0 and standard deviation of 1',
          'Only values of true and false',
          'Only text labels',
          'A maximum value of exactly 100'
        ],
        answer: 1,
        hint: 'This is a common scaling technique.'
      },
      {
        question: 'Which human factor can directly reduce labeling consistency over time?',
        choices: [
          'Fatigue and boredom',
          'Higher screen resolution',
          'More storage space',
          'Faster network speed'
        ],
        answer: 1,
        hint: 'Human attention is limited.'
      },
      {
        question: 'What is a major risk when labeling guidelines are unclear?',
        choices: [
          'Annotators may apply labels inconsistently',
          'The dataset becomes too small to store',
          'All labels become numerical automatically',
          'The model trains without data'
        ],
        answer: 1,
        hint: 'Clear rules improve shared interpretation.'
      },
      {
        question: 'What is the goal of training and calibration sessions for annotators?',
        choices: [
          'To align annotators on the same labeling standard',
          'To increase file size',
          'To remove quality control',
          'To avoid checking disagreements'
        ],
        answer: 1,
        hint: 'Calibration improves consistency.'
      },
      {
        question: 'What is the main tradeoff in the paradox of big data for labeling?',
        choices: [
          'More data can help generalization, but poor consistency can hurt precision',
          'More data always guarantees better labels',
          'Smaller data always removes bias',
          'Bigger datasets do not need quality control'
        ],
        answer: 1,
        hint: 'Size does not automatically mean quality.'
      },
      {
        question: 'Why do inconsistent labels confuse machine learning models?',
        choices: [
          'They give the model conflicting patterns to learn',
          'They reduce monitor brightness',
          'They make all features ratio data',
          'They eliminate the need for testing'
        ],
        answer: 1,
        hint: 'The model learns from the labels it is given.'
      }
    ];
    let currentPlayer, boardState, cellQuestions, selectedCell;
    const boardEl = document.getElementById('board');
    const statusEl = document.getElementById('status');
    const bannerEl = document.getElementById('banner');
    const hintBtn = document.getElementById('hintBtn');
    const panelQuestion = document.getElementById('panelQuestion');
    const panelChoices = document.getElementById('panelChoices');
    const panelHint = document.getElementById('panelHint');
    const restartBtn = document.getElementById('restartBtn');
    function shuffle(arr) {
      for (let i = arr.length - 1; i > 0; i--) {
        const j = Math.floor(Math.random() * (i + 1));
        [arr[i], arr[j]] = [arr[j], arr[i]];
      }
      return arr;
    }
    function buildBalancedPositions(nQuestions, kChoices) {
      const base = Math.floor(nQuestions / kChoices);
      const remainder = nQuestions % kChoices;
      const slots = [];
      for (let pos = 1; pos <= kChoices; pos++) {
        for (let c = 0; c < base; c++) slots.push(pos);
      }
      const extra = shuffle(Array.from({ length: kChoices }, (_, i) => i + 1)).slice(0, remainder);
      slots.push(...extra);
      return shuffle(slots);
    }
    function initGame() {
      currentPlayer = 'X';
      boardState = Array(9).fill('');
      selectedCell = null;
      bannerEl.classList.add('hidden');
      statusEl.innerText = '';
      const shuffledQs = shuffle([...questions]).slice(0, 9);
      const kChoices = 4;
      const targetPositions = buildBalancedPositions(shuffledQs.length, kChoices);
      cellQuestions = shuffledQs.map((q, idx) => {
        const originalCorrectIdx0 = q.answer - 1;
        const correctText = q.choices[originalCorrectIdx0];
        const wrongs = q.choices.filter((_, i) => i !== originalCorrectIdx0);
        shuffle(wrongs);
        const targetPos1 = targetPositions[idx];
        const targetPos0 = targetPos1 - 1;
        const newChoices = Array(kChoices).fill(null);
        newChoices[targetPos0] = correctText;
        let w = 0;
        for (let i = 0; i < kChoices; i++) {
          if (newChoices[i] === null) newChoices[i] = wrongs[w++];
        }
        return {
          question: q.question,
          choices: newChoices,
          answer: targetPos1,
          hint: q.hint
        };
      });
      panelQuestion.innerText = 'Select a square to view the question';
      panelChoices.innerHTML = '';
      panelHint.classList.add('hidden');
      panelHint.innerText = '';
      renderBoard();
    }
    function renderBoard() {
      boardEl.innerHTML = '';
      boardState.forEach((mark, idx) => {
        const btn = document.createElement('button');
        let cls = 'bg-white h-24 w-24 sm:h-28 sm:w-28 flex items-center justify-center text-4xl font-extrabold rounded-lg shadow hover:bg-indigo-100 transition';
        if (mark === 'X') cls += ' text-blue-600';
        if (mark === 'O') cls += ' text-red-600';
        btn.className = cls;
        btn.innerText = mark;
        btn.disabled = mark !== '';
        btn.addEventListener('click', () => openQuestion(idx));
        boardEl.appendChild(btn);
      });
      statusEl.innerText = `Current: ${currentPlayer}`;
    }
    function openQuestion(idx) {
      if (boardState[idx] !== '') return;
      selectedCell = idx;
      const q = cellQuestions[idx];
      panelQuestion.innerText = q.question;
      panelChoices.innerHTML = '';
      panelHint.classList.add('hidden');
      panelHint.innerText = q.hint;
      q.choices.forEach((choiceText, i) => {
        const li = document.createElement('li');
        const btn = document.createElement('button');
        btn.innerText = choiceText;
        btn.className = 'w-full text-left px-4 py-2 bg-indigo-50 rounded hover:bg-indigo-100 border border-indigo-100';
        btn.addEventListener('click', () => handleAnswer(i + 1));
        li.appendChild(btn);
        panelChoices.appendChild(li);
      });
    }
    function handleAnswer(choice) {
      if (selectedCell === null) return;
      const q = cellQuestions[selectedCell];
      if (choice === q.answer) {
        boardState[selectedCell] = currentPlayer;
        panelChoices.innerHTML = '';
        panelHint.classList.add('hidden');
        panelQuestion.innerText = 'Correct! Select another open square.';
        renderBoard();
        if (checkWin()) return endGame(`${currentPlayer} wins!`);
        if (boardState.every(c => c)) return endGame('Stalemate!');
      } else {
        alert('Incorrect! Turn missed.');
      }
      currentPlayer = currentPlayer === 'X' ? 'O' : 'X';
      statusEl.innerText = `Current: ${currentPlayer}`;
      selectedCell = null;
    }
    function checkWin() {
      const wins = [
        [0,1,2], [3,4,5], [6,7,8],
        [0,3,6], [1,4,7], [2,5,8],
        [0,4,8], [2,4,6]
      ];
      return wins.some(combo => combo.every(i => boardState[i] === currentPlayer));
    }
    function endGame(msg) {
      bannerEl.innerText = `🎉 ${msg}`;
      bannerEl.classList.remove('hidden');
      document.querySelectorAll('#board button').forEach(b => b.disabled = true);
      panelQuestion.innerText = msg;
      panelChoices.innerHTML = '';
      panelHint.classList.add('hidden');
    }
    hintBtn.addEventListener('click', () => {
      if (selectedCell !== null) panelHint.classList.toggle('hidden');
    });
    restartBtn.addEventListener('click', initGame);
    initGame();
  </script>
</body>
</html>