Physical Inputs Needed to Gather Training Data
Question: What physical resources are needed to collect the input-output examples required to train a fully end-to-end self-driving system?
Sample answer: You would need a fleet of sensor-equipped cars and human drivers to operate them while data is collected.
Key points:
- Sensor-equipped cars
- Human drivers or operators
Rubric: Award full credit if the student mentions sensor-equipped cars.
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
What labeled pairs are needed to train a pure end-to-end delivery-drone controller?
Even with lots of labeled examples, an end-to-end model is not automatically the best design choice.
To train a camera-guided delivery robot end to end, you need many (Image, _____) pairs.
Match each application to the labeled input-output data it needs for end-to-end training.
Order the chain that explains why an end-to-end warehouse robot is hard to train.
Why would training a pure end-to-end warehouse robot require specially equipped robots?
End-to-end learning usually works best when many labeled input-output pairs are available.
If paired examples are scarce, you should treat end-to-end learning with great _____.
Match each autonomous driving data-collection challenge with its consequence.
Order the reasoning steps for deciding whether a single end-to-end model is practical for a new task.
Training Data Demands for Direct Control Models
When a Direct Model Is a Good Idea
Physical Inputs Needed to Gather Training Data