CNN & Image Classification
Convolutions, pooling, architectures (VGG, ResNet), transfer learning, data augmentation, fine-tuning
1What is a convolution operation in a CNN?
What is a convolution operation in a CNN?
Answer
A convolution is a mathematical operation that applies a filter (kernel) over an image by sliding this filter across the input and computing the dot product at each position. This allows extracting local features such as edges, textures, or patterns. Unlike dense networks where each neuron is connected to all inputs, convolution exploits the spatial structure of images by sharing filter weights across the entire image.
2What is the role of stride in a convolution layer?
What is the role of stride in a convolution layer?
Answer
Stride defines the step size when moving the filter across the image. A stride of 1 moves the filter by one pixel at each step, while a stride of 2 moves it by 2 pixels, thus reducing the output size. Increasing stride allows reducing the spatial dimension of the feature map and computational cost, but may also cause information loss if the stride is too large.
3What is the purpose of padding in a convolution layer?
What is the purpose of padding in a convolution layer?
Answer
Padding consists of adding pixels (usually zeros) around the input image before applying convolution. This allows controlling the output size and preserving information at image edges. With 'same' padding, the output has the same size as input, while with 'valid' (no padding), the size decreases. Padding also prevents edge pixels from being underrepresented in feature maps.
What is the difference between Max Pooling and Average Pooling?
What is a feature map in a CNN?
+21 interview questions
Other Data Science & ML interview topics
Python Basics
Python Object-Oriented Programming
Python Data Structures
Git Fundamentals
SQL Basics
NumPy Fundamentals
Pandas Basics
Jupyter & Google Colab
SQL Joins & Advanced Queries
Advanced Pandas
Visualization with Matplotlib & Seaborn
Interactive Visualizations with Plotly
Descriptive Statistics
Inferential Statistics
Web Scraping
BigQuery & Cloud Data
Feature Engineering
Supervised ML: Regression
Supervised ML: Classification
Decision Trees & Ensembles
Unsupervised ML
ML Pipelines & Validation
Time Series & Forecasting
Deep Learning Fundamentals
TensorFlow & Keras
RNN & Sequences
Transformers & Attention
NLP & Hugging Face
GenAI & LangChain
MLOps & Deployment
Master Data Science & ML for your next interview
Access all questions, flashcards, technical tests, code review exercises and interview simulators.
Start for free