QuizzesCost/Loss Functions & Convexity — Concept Check
intermediate
math-for-ai

Cost/Loss Functions & Convexity — Concept Check

Questions on Cost/Loss Functions & Convexity. A loss function is the single number a model tries to minimise, and convexity describes whether that landscape is a simple bowl or a mountain range full of traps.

7 minutes 7 questions Pass at 70%
Time Limit
7 minutes - timer starts when you begin
Change Answers Freely
Revisit and change any answer before submitting
Full Review After Submit
Correct answers and explanations shown for every question
Auto-submit on timeout
Quiz submits automatically when time runs out

Questions in this quiz

7 questions on Cost/Loss Functions & Convexity — Concept Check. Work through them here, then start the quiz to answer them and see your score.

  1. 1.A model is trained using mean squared error (MSE) as its loss function. What is the primary objective of this model?

    • AMaximize the probability of correct classification
    • BMinimize the squared differences between predicted and actual values
    • CMinimize the absolute differences between predicted and actual values
    • DMaximize the confidence in incorrect predictions
  2. 2.A deep learning model struggles to converge during training. Which characteristic of its loss function might explain this?

    • AThe loss function is non-convex with multiple local minima
    • BThe loss function is quadratic with a single minimum
    • CThe loss function is convex with a single global minimum
    • DThe loss function is linear with no minima
  3. 3.A logistic regression model is used for binary classification. Which loss function is most appropriate?

    • AAbsolute error
    • BBinary cross-entropy
    • CHinge loss
    • DMean squared error
  4. 4.During training, a model's loss decreases initially but then plateaus. What might be a reason for this behavior?

    • AThe learning rate is too high, causing overshooting
    • BThe loss function is not differentiable
    • CThe model has reached a local minimum in a non-convex loss surface
    • DThe model has reached the global minimum in a convex loss surface
  5. 5.Why might gradient descent fail to find the global minimum in a deep network?

    • AGradient descent only works for linear models
    • BGradient descent is not suitable for any optimization
    • CDeep networks have non-convex loss functions
    • DDeep networks have convex loss functions
  6. 6.A model using cross-entropy loss shows high confident errors. What does this indicate?

    • AThe model's predictions are confidently incorrect
    • BThe model has reached the global minimum
    • CThe model is underfitting and needs more complexity
    • DThe model is overfitting to the training data
  7. 7.What is the main advantage of using a convex loss function in optimization?

    • AIt simplifies the model architecture
    • BIt guarantees finding the global minimum from any starting point
    • CIt allows for non-differentiable optimization
    • DIt ensures the presence of multiple global minima