AI0-001 AI Concepts and Techniques Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Convolutional Neural Network (CNN)

  Why it's wrong here

  Convolutional neural networks apply local filters with weight sharing over grid-structured inputs such as images, and their fixed receptive fields do not model the temporal dependencies or gating that suppress vanishing gradients in sequential data; recurrent architectures with LSTM or GRU cells, or transformers, address that requirement. CNNs are tempting because they are the default choice for image classification and can be applied to one-dimensional sequences such as audio or text, but they lack the recurrent state needed for long-range order.

• ✗

  Transformer

  Why it's wrong here

  Transformers process sequences in parallel through self-attention, which removes recurrence but does not itself target the vanishing gradient of long backpropagation chains. The option tempts because Transformers now dominate sequential tasks, yet the architecture explicitly designed to mitigate that gradient problem is the LSTM.

• ✗

  Vanilla Recurrent Neural Network (RNN)

  Why it's wrong here

  A vanilla RNN suffers the vanishing gradient problem rather than mitigating it, because gradients shrink as they propagate back through many timesteps. The option tempts because RNNs are indeed built for sequential data, but the gated LSTM architecture was created specifically to address that gradient decay.

• ✓

  Long Short-Term Memory (LSTM) network

  Why this is correct

  LSTM networks use gated cells (input, forget, output) that regulate information flow across time steps, preserving gradients over long sequences. This architecture directly mitigates the vanishing gradient problem while handling sequential data, matching the stem's requirement.