AI in Materials Discovery Quiz

Answers and explanations appear as you take the quiz.

1. How does a graph neural network typically represent a crystal?
   • As atoms (nodes) connected by bonds (edges)
   • As a paragraph of text
   • As a single number
   • As an audio waveform
2. What slow physics calculation do AI surrogates often replace for fast screening?
   • Long division
   • Density functional theory (DFT)
   • Weather forecasting
   • Spell checking
3. What did DeepMind's GNoME contribute in 2023?
   • A new programming language
   • A faster web browser
   • Hundreds of thousands of predicted stable crystal structures
   • A robot vacuum
4. Why is rotational and translational invariance important in materials models?
   • It makes models bigger
   • It speeds up the internet
   • It is required by law
   • It keeps predictions physically consistent regardless of how atoms are positioned or labeled
5. What is a 'self-driving lab' in materials discovery?
   • A facility where AI selects candidates and robots synthesize and test them
   • A car that studies chemistry
   • A simulation with no experiments
   • A database of papers
6. Beyond thermodynamic stability, what challenge is increasingly emphasized?
   • The color of the lab
   • Whether a material can actually be synthesized
   • The brand of the computer
   • The price of electricity only
7. How many GNoME structures did the 2023 study report as independently experimentally verified?
   • 7 structures
   • 73,600 structures
   • 736 structures
   • All 381,000 structures
8. How does MatterGen generate candidate crystal structures?
   • It translates database titles into chemical formulas
   • It copies measured crystals without modification
   • It optimizes only the color of a rendered structure
   • A diffusion process gradually refines atom types, coordinates, and the periodic lattice