Foundations 0/0

Foundations

NLP, vision and recommenders

The same methods applied to text, images, recommendations and graphs: tokenisers and word embeddings, detection and segmentation, matrix factorisation and message-passing networks. Many production systems pair a language model with one of these.

0 / 15 solved

Questions

0/15 solved
Question Module Tags Difficulty Success
1. Applying Merges In Rank Order Natural Language Processing
nlptokenization +2
Medium —
2. Counting Output Vectors Per Pair Natural Language Processing
nlpword2vec +2
Medium —
3. Two Washingtons, One Vector Natural Language Processing
nlpner +2
Medium —
4. Bigram Probability By Hand Natural Language Processing
nlplanguage-models +2
Medium —
5. The Validation Loader That Keeps Moving Computer Vision
computer-visiondata-augmentation +2
Medium —
6. Sobel on a Dark-to-Light Patch Computer Vision
computer-visionconvolution +2
Medium —
7. Two Boxes, One Threshold Computer Vision
computer-visionobject-detection +2
Medium —
8. One Label Per Pixel Computer Vision
computer-visionsegmentation +2
Medium —
9. Weighted By Taste, Not By Genre Recommender Systems
recommender-systemscollaborative-filtering +2
Medium —
10. Thirty-Four Gigabytes Of Nothing Recommender Systems
recommender-systemssparsity +2
Medium —
11. Two Empty Rows Recommender Systems
recommender-systemscold-start +2
Medium —
12. Clicks Don't Lie, They Just Whisper Recommender Systems
recommender-systemsmatrix-factorization +1
Medium —
13. Readout Loses The Molecule Graph ML
graph-mlgnn +2
Medium —
14. Where The Walk Steps Next Graph ML
graph-mlnode2vec +2
Hard —
15. Three Layers Of Averaging Graph ML
graph-mlgnn +2
Hard —