Which is not a deep learning algorithm?

Question

Options

• ASetr-encoder
• BConvolutional neural networks
• CRecurrent neural networks
• DSupport vector machine

• A
  3% (1)
• B
  5% (2)
• C
  10% (4)
• D
  83% (33)

Explanation

Support Vector Machine (SVM) is a classical machine learning algorithm based on finding an optimal hyperplane to separate data classes - it does not use layered neural network architectures or backpropagation, which are the defining characteristics of deep learning.

Options A, B, and C are all deep learning algorithms: Convolutional Neural Networks (CNNs) use stacked convolutional layers to learn spatial features (images, video); Recurrent Neural Networks (RNNs) use sequential, memory-carrying layers for time-series and language data; and SETR (Segmentation Transformer Encoder) is a transformer-based deep architecture for image segmentation tasks. All three rely on multi-layer, differentiable networks trained via gradient descent.

Memory tip: Deep learning = deep neural networks. If an algorithm name includes "neural network," "transformer," or "encoder/decoder," it's almost certainly deep learning. SVM is a kernel-based geometric method - no neurons, no layers, no "deep."

No community discussion yet for this question.