AWS Certified AI PractitionerAI/ML and Generative AI FundamentalsHard
A research team is developing a generative AI model to create novel protein structures based on specific functional requirements. The model needs to understand complex relationships between amino acid sequences and their resulting 3D structures, and critically, how changes in one part of the sequence can influence distant parts of the structure. Which neural network architecture is best suited for handling these long-range dependencies and intricate pattern recognition in sequential data for generative tasks?
- AMultilayer Perceptron (MLP)
- BConvolutional Neural Network (CNN)
- CRecurrent Neural Network (RNN)
- DTransformer Architecture
Show answer & explanationAnswer & explanation
Correct answer: D. Transformer Architecture
Transformer architecture, particularly with its self-attention mechanism, excels at capturing long-range dependencies in sequential data like protein sequences. This allows it to understand how distant parts of a sequence influence each other, which is crucial for generating complex, functional protein structures.
Why the other options are wrong
- A. MLPs are basic feedforward networks that treat inputs independently, lacking the ability to process sequences or capture dependencies between elements.
- B. CNNs are primarily designed for spatial data (images) and local patterns, not ideal for long-range dependencies in sequences without significant modifications.
- C. RNNs handle sequential data but struggle with long-range dependencies due to vanishing/exploding gradients and limited memory over long sequences, making them less effective for complex protein structures.
Transformer Architecture
A neural network architecture, primarily relying on self-attention mechanisms, designed to process sequential data and capture long-range dependencies efficiently.
- Uses self-attention to weigh the importance of different parts of the input sequence.
- Processes entire sequences in parallel, unlike RNNs.
- Forms the foundation for many state-of-the-art Generative AI models, especially LLMs.
Memory trick: Networks Learn, Architectures Define How.