{"description":"Test your understanding of audio preprocessing, spectrograms, CNN architecture, model training, and audio classification.","questions":[{"answer":"A 2D visual representation of frequency vs time","number":1,"options":["The speed of the audio playback","A histogram of sound frequencies","A 2D visual representation of frequency vs time","A list of audio files"],"question":"What does a spectrogram represent in audio processing?"},{"answer":"Short-Time Fourier Transform (STFT)","number":2,"options":["Wavelet transform","Fourier transform","Short-Time Fourier Transform (STFT)","Linear regression"],"question":"What technique is commonly used to convert raw audio into a spectrogram?"},{"answer":"They enable CNNs to extract spatial and frequency features from sound","number":3,"options":["They reduce audio length","They allow visualization of data during testing","They enable CNNs to extract spatial and frequency features from sound","They eliminate the need for activation functions"],"question":"Why are spectrograms useful for CNNs in audio recognition?"},{"answer":"Normalization layer","number":4,"options":["Dense layer","Dropout layer","Normalization layer","ReLU activation"],"question":"Which layer is typically used to standardize input values in CNNs?"},{"answer":"Prevents overfitting by randomly disabling neurons","number":5,"options":["Increases training accuracy","Adds noise to audio inputs","Prevents overfitting by randomly disabling neurons","Improves frequency resolution"],"question":"What is the role of the dropout layer in CNN training?"},{"answer":"Recognizing short spoken commands","number":6,"options":["Multi-language translation","Recognizing short spoken commands","Speech synthesis","Noise cancellation"],"question":"What is a good use case for the Mini Speech Commands dataset?"},{"answer":"Adam","number":7,"options":["RMSprop","SGD","Adam","Nadam"],"question":"Which optimizer is often used in training CNNs for audio recognition?"},{"answer":"Sparse categorical cross-entropy","number":8,"options":["Mean squared error","Binary cross-entropy","Sparse categorical cross-entropy","Hinge loss"],"question":"What loss function is typically used for multi-class classification with integer labels?"},{"answer":"Halts training when validation loss stops improving","number":9,"options":["Increases batch size over time","Stops training if accuracy exceeds 100%","Halts training when validation loss stops improving","Speeds up GPU performance"],"question":"What does EarlyStopping do during training?"},{"answer":"A set of class probabilities","number":10,"options":["A raw audio waveform","A list of sound files","A set of class probabilities","A binary classification score"],"question":"What does the model output after predicting on an audio spectrogram?"}],"title":"Audio Recognition Visual Demo"}