Learn Before
Inputs and Outputs in End-to-End Speech Recognition
Question: In an end-to-end speech recognition system, what is the input and what is the output? How does that output compare in complexity to the result of a simpler prediction task?
Sample answer: The input is a spoken audio recording, and the output is a text transcription. The transcription is much richer than a simple scalar prediction such as a single probability or number.
Key points:
- Input is spoken audio
- Output is text transcription
- The output is more structured than a single number
Rubric: Full credit for naming the audio input and text output, and for noting that the output is more complex than a single numeric prediction.
0
1
Tags
Machine Learning
Deep Learning
Supervised Learning
Dive into Deep Learning @ D2L
Data Science
Machine Learning Strategy
Machine Learning Yearning @ DeepLearning.AI
Related
When an end-to-end speech-to-text model receives an audio recording, what should it output directly?
End-to-end speech recognition has been a successful application of end-to-end learning.
An end-to-end speech recognition model can take a(n) _____ and produce the transcript directly.
Match each item to its role in an end-to-end speech recognition system.
Order the steps in a simple end-to-end speech recognition pipeline.
Which dataset is most important for training an end-to-end speech recognition system?
End-to-end learning is always the most effective choice for every machine learning problem.
End-to-End Systems Can Succeed, Yet Still Be Suboptimal
Match each statement with the best supporting detail about direct input-to-output learning in speech recognition.
Choosing Between End-to-End and a Pipeline for Transcription
Why end-to-end speech systems can predict more than a single value
Choosing an End-to-End Model for Meeting Transcription
Inputs and Outputs in End-to-End Speech Recognition
Traditional vs. End-to-End Speech Recognition Pipelines