AI / ML Learning Check

Preview mode

Published

Copy responder link

Request edit access

JavaScript isn't enabled in your browser, so this file can't be opened. Enable and reload.

AI / ML Learning Check

Full Name

Major

1. How can you remove noise from a biosignal using Python?

Hint: Signal Processing Techniques in Python (SciPy Docs)

2. What time-domain feature measures how spread-out signal values are?

Hint: Feature Extraction from Time Series – sktime Documentation

3. What does the FFT (Fast Fourier Transform) show about a biosignal?

Hint: Harvard: Fourier Transforms and Signal Processing Notes

4. Why use TensorFlow Lite Micro instead of normal TensorFlow on Arduino?

Hint: TensorFlow Lite Micro Documentation

5. When collecting sensor data for supervised ML tasks, why is labeling important?

A) It reduces memory usage

B) It lets the model map inputs to correct outputs

C) It makes signals easier to filter

D) It reduces battery drain

Clear selection

6. Which preprocessing step would you apply to reduce high-frequency noise in biosignal data?

A) Normalization

B) Low-pass filtering

C) Feature extraction

D) Quantization

Clear selection

7. Which of the following is a time-domain feature?

A) Variance

B) Spectrogram

C) Fourier Transform

D) Frequency bin energy

Clear selection

8. Which tool can extract hundreds of statistical features (variance, entropy, zero-crossings, etc.) from time series automatically?

A) Edge Impulse

B) scikit-learn

C) TSFresh

D) TensorFlow Lite

Clear selection

9. Edge Impulse is especially useful because:

A) It allows training and deploying ML models on embedded devices with minimal setup

B) It is only for cloud-based training

C) It can only run on desktop GPUs

D) It doesn’t require labeled data

Clear selection

10. Which Python library is best for quickly testing baseline models like Random Forest or Logistic Regression?

A) TensorFlow Lite

B) scikit-learn

C) Arduino IDE

D) TSFresh

Clear selection

11. What’s the difference between supervised and unsupervised learning?

12. What is the purpose of quantization when deploying ML models to microcontrollers?

A) To increase model accuracy

B) To shrink model size and reduce memory/latency

C) To add more neurons to the model

D) To change the training dataset

Clear selection

13. Which toolkit would you use to prune, quantize, or otherwise optimize a TensorFlow model for deployment on microcontrollers?

A) Edge Impulse CLI

B) TensorFlow Model Optimization Toolkit

C) scikit-learn pipelines

D) Arduino IDE

Clear selection

14. What’s one trade-off of quantization (e.g., converting from float32 to int8 weights)?

15. According to TinyML examples, how small can a deep learning model get and still run effectively on a microcontroller?

A) 1 GB

B) 14 KB

C) 100 MB

D) 1 MB

Clear selection

Submit

Clear form

Never submit passwords through Google Forms.

This content is neither created nor endorsed by Google. - Terms of Service - Privacy Policy

Does this form look suspicious? Report

Forms