以下哪种关于 CNN 中池化的说法不正确？

Which of the following is NOT true for pooling in CNNs?

Question

以下哪种关于 CNN 中池化的说法不正确？

Which of the following is NOT true for pooling in CNNs?

BlackTom AI · Accepted Answer

Pooling performs local spatial downsampling (for example max or average pooling), summarizing neighboring activations and reducing spatial resolution.

1) Relevant concepts
- Pooling types: max-pooling, average-pooling. Both aggregate (summarize) values in a local neighborhood (e.g., 2×2 window).
- Purpose: reduce spatial dimensions, provide translation tolerance, decrease computation and number of activations passed to subsequent layers.
- Pooling is a form of feature summarization/extraction (it extracts a summary statistic of local features), although it is usually non-learnable.

2) Step-by-step reasoning applied to each option
- Option A: "It is a method of summarizing neighboring feature detectors." This matches pooling’s definition (true), because pooling computes a summary (max or average) over neighboring activations.
- Option B: "It is a method for increasing the number of features for the next layer." Pooling reduces spatial size, so it does not increase the number of features for the next layer (false).
- Option C: "It is a method of reducing the number of features for the next layer." This is consistent with downsampling behavior (true).
- Option D: "It is not a method of feature extraction." Pooling does perform feature summarization/extraction (true that pooling extracts local summary features), so this statement is false.

3) Example calculation (intermediate steps)
- Suppose an input feature map is 32×32×C. Applying 2×2 pooling with stride 2: each non-overlapping 2×2 block becomes one value per channel → output size becomes 16×16×C. The spatial count went from 1024 to 256 per channel, a 4× reduction (32×32 → 16×16). This shows pooling reduces, not increases, the number of spatial activations/features.

4) Why the correct answers are correct
- The two statements that are NOT true are: (1) "It is a method for increasing the number of features for the next layer"—pooling reduces spatial activations; and (2) "It is not a method of feature extraction"—pooling does perform local feature summarization. Thus both are false descriptions of pooling.

5) Why the other options are incorrect choices for "NOT true"
- Option A and Option C describe true aspects of pooling: pooling summarizes neighboring detectors and reduces the number of features, so they should not be selected as “NOT true.”

6) Conclusion
The statements that are NOT true for pooling are: "It is a method for increasing the number of features for the next layer." and "It is not a method of feature extraction." This matches the provided answer.

以下哪种关于 CNN 中池化的说法不正确？ Which of the following is NOT true for pooling in CNNs?多项选择题

类似问题

此問題已經被重新評分。以下哪种关于 CNN 中池化的说法不正确？ Which of the following is NOT true for pooling in CNNs?

What is the main purpose of pooling in text CNNs (Convolutional Neural Networks)? (Select one option from below.)

Which of the following trees corresponds to a potential parse of the ambiguous sentence below, with correct syntactic categories? Some diagnostics are provided.

Which of the following sentences contain two non-finite verbs? Select all that apply.

Analyze the following sentence: The sentence contains [ Select ] 8 6 5 7 3 4 functional words, [ Select ] 2 3 1 4 5 count nouns, [ Select ] 4 5 2 1 3 mass nouns, and [ Select ] 1 5 2 3 4 adjectives.

更多留学生实用工具

智能学习助手

风格化写作助手

论文查重助手

文献引用助手

课堂转译助手

课堂笔记助手

Quiz搜索助手

学校历年真题

智能刷题助手

智能匹配练习

希望你的学习变得更简单