Question 2 NAT · 1.0 marks
A published solution is not available for this question yet.
Deep Learning for Computer Vision · End Term · Sep 2025 · FN
Deep Learning for Computer Vision · End Term · Sep 2025 FN
← Course papers · Start practice / exam
Questions and published explanations below are available without starting a test. Some questions may not have a published solution yet.
0 / 19 published solutions
Question 3 NAT · 1.0 marks
A published solution is not available for this question yet.
Question 4 MCQ · 1.0 marks
Passing patch embeddings through the Transformer encoder
Linear projection of flattened image patches
Adding a class token
Layer normalization of patch embeddings
A published solution is not available for this question yet.
Question 5 MCQ · 1.0 marks
Flattens the softmax distribution, reducing gradient magnitude
sharpens the softmax distribution, increasing emphasis on hard negatives
has no effect on the loss landscape
makes image and text embeddings orthogonal
A published solution is not available for this question yet.
Question 6 MCQ · 1.0 marks
In a Feature Pyramid Network (FPN), the primary purpose of the top-down pathway with
lateral connections is to:
increase the receptive field at the lowest-resolution level
enrich high-resolution feature maps with strong semantics from deeper layers
perform non-maximum suppression at multiple scales
reduce the number of anchors required per level
A published solution is not available for this question yet.
Question 7 MCQ · 1.0 marks
Consider the following pyTorch code:
tensor([0.09, 0.24, 0.67])
tensor([0.09, 0.24, 0.66])
tensor([0.09, 0.23, 0.68])
tensor([0.10, 0.27, 0.63])
A published solution is not available for this question yet.
Question 8 MCQ · 1.0 marks
Use the test set to select the best hyperparameters
Use k-fold cross-validation on the training data and keep the test set
untouched for final re-porting
Tune on the training loss only; validation is unnecessary
Pick hyperparameters from a prior paper without checking performance
A published solution is not available for this question yet.
Question 9 MCQ · 1.0 marks
How is disentangled representation evaluated?:
Using Mutual Information Gap
Using mean Average Precision
KL divergence
Sparsity in the latent one-hot codes
A published solution is not available for this question yet.
Question 10 MSQ · 1.0 marks
Which statements about generative models are true? (Select all that apply.)
VAEs optimize a lower bound on the data log-likelihood.
GANs perform explicit maximum likelihood estimation.
Diffusion models can trade off speed and quality via the number of sampling
steps.
Autoregressive models factorize the joint distribution into a product of
conditionals.
A published solution is not available for this question yet.
Question 11 MSQ · 1.0 marks
In classifier-free guidance for diffusion models, increasing the guidance scale γ tends to: (Select all
that apply.)
improve fidelity to the conditioning signal
always reduce mode collapse to zero
reduce sample diversity at very high values
be equivalent to classifier guidance
A published solution is not available for this question yet.
Question 12 MSQ · 1.0 marks
Which statements are correct? (Select all that apply.)
A linear autoencoder with MSE and no activation learns a subspace equivalent
to PCA (up to rotation) when the hidden dimension < input dimension.
Global attention assigns alignment weights to all encoder time steps for each
decoder step.
Positional encodings convey order information that self-attention lacks.
it is impossible to generate multiple, semantically similar captions with
different styles for the same image.
A published solution is not available for this question yet.
Question 13 NAT · 1.0 marks
A published solution is not available for this question yet.
Question 14 NAT · 1.0 marks
A published solution is not available for this question yet.
Question 15 NAT · 1.0 marks
A published solution is not available for this question yet.
Question 16 MCQ · 1.0 marks
Mixup data augmentation primarily:
reduces label noise by hardening labels
creates convex combinations of inputs and labels to encourage linear behavior
between classes
prunes redundant training samples
increases input resolution without changing model capacity
A published solution is not available for this question yet.
Question 17 NAT · 1.0 marks
A published solution is not available for this question yet.
Question 18 MCQ · 1.0 marks
During training, Batch Normalization uses:
Running (population) statistics only
batch statistics for normalization and updates running estimates for inference
neither batch nor running statistics
group-wise normalization identical to GroupNorm
A published solution is not available for this question yet.
Question 19 MSQ · 1.0 marks
Which techniques help mitigate exploding/vanishing gradients in RNNs? (Select all that apply.)
Gradient clipping
Orthogonal or identity initialization of recurrent weights
Using LSTM/GRU cells
Randomly reversing input sequences as augmentation
A published solution is not available for this question yet.
Question 20 NAT · 1.0 marks
A published solution is not available for this question yet.