Question 19: Consider an encoder-decoder architecture trained using m…

Select all that apply.

1. A

   During training, the decoder may receive ground-truth tokens as inputs at subsequent time steps.

2. B

   During inference, the decoder conditions on its own previously generated tokens.

3. C

   Teacher forcing can be used during inference to give more accurate predictions.

4. D

5. E

   Beam search trades computational cost for a broader exploration of possible output sequences.

6. F

   Beam search guarantees recovery of the globally optimal sequence for any fixed beam width.