LEAP-2022-Bootcamp

123456789101112131415161718192021222324252627282930313233
Across
  1. 2. technique to create parsimonious representation
  2. 3. influence on the Earth system
  3. 5. a machine learning approach inspired by the brain
  4. 6. greenhouse gas primarily responsible for anthropogenic climate change
  5. 12. an algorithm that learns without a teacher
  6. 13. type of clouds formed by localized upward air motion
  7. 15. method of representing processes that are too small to be resolved in a climate model
  8. 16. absolute temperature scale
  9. 18. computer language used for today's Earth System Models
  10. 19. a variable that influences both the dependent and independent variables, causing a spurious association
  11. 21. transfer of heat by physical contact
  12. 24. clever application of chain rule for neural network gradients
  13. 27. a neural network has many of these
  14. 29. from these deposits on the ocean floor, past climate signals can be infered
  15. 30. neural network that reconstructs the input
  16. 32. ocean circulation by which surface water moves to the deep ocean
  17. 33. unit of energy
Down
  1. 1. matrix decomposition that discovers lower-order structures in a matrix
  2. 4. method to select hyper-parameters
  3. 7. a type of neural network without loops
  4. 8. another name to call independent variables
  5. 9. correlation does not imply ...
  6. 10. a scalar measuring how far the prediction is from the truth
  7. 11. anthropogenic changes on the climate are induced by ...
  8. 14. a generative model that competes with a discriminative model
  9. 17. a type of variables that are not directly observed
  10. 20. when a machine learning system fails to generalize to unseen examples
  11. 22. convection happens when a fluid is heated from ...
  12. 23. a representation of what the machine learning system has learned
  13. 25. probability assigned to models or parameters before any data is taken into account
  14. 26. the component of the Earth system that is referred to as cryosphere
  15. 28. joint probability of the observed data as a function of models and parameters
  16. 31. a complete pass through the dataset in deep learning