Which of the following statements accurately describe the capabilities and functions of sequence transduction models? Select all that apply.

Sequence transduction models take an input sequence and produce an output sequence, so they can translate sentences, transcribe speech, or predict future tokens. They learn a mapping that can change the length of the output, allowing the model to insert or delete symbols as needed. During training the model learns the alignment between input and output positions, often using attention or encoder‑decoder architectures. For example, a transformer can read the phrase “I love you” in English and output “Je t’aime” in French, learning to match words even though the lengths differ. These models are end‑to‑end trainable and can generate new sequences that follow the learned statistical patterns.