Refer to Figure 1.7. Which of the following set of points all represent inefficient levels of production, ceteris paribus?

Q7: (8 points)After taking the action suggested in the previous question, suppose the discount factor is γ = 0.9, the state transfers from s₂ to s₄ after taking action aₜ, and the reward r is 0.6. Please update the Q-table and write down the updated Q-table. Note: Only one value in the table needs updating, and you might need the Bellman Equation:

Q5: (6 points)Give a scenario that hard margin SVM doesn’t work well, and we need soft margin SVMs.

Q2: (6 points) Assuming we aim to build a more advanced recommendation system for an online bookstore using matrix factorization-based methods, similar to the one that won the Netflix prize. Suppose the global mean rating of books is 3.6 stars. Bob, a loyal customer, has rated 400 books, and his average rating is 0.3 stars higher than the global average rating. Meanwhile, Pride and Prejudice is a book in the bookstore that has 200,000 ratings, with an average rating that is 0.5 stars lower than the global average. What would be a baseline estimate of Bob’s rating for Pride and Prejudice? (2 points) Illustrate how you arrived at your answer. (2 points)

Q6: (4 points)What is the strategy of AdaBoosting for reweighing the training data points?

Q3: (3 points) Answer True or False only: In Boosting, the performance of the final model heavily depends on the performance of the individual weak learners.

Q4: (8 points) Designing A Machine Learning System.Given user features, item features, and a user-item-rating matrix, if we formulate the problem of recommending personalized items for users as a ranking task, how can we use develop a personalized Learning To Rank (LTR) model for recommendations? Please specify: how you will use the data what is your model structure what is your objective function how to use the learned ranking model to conduct personalized recommendations.

Q6: (15 points)In bipartite ranking, documents are grouped into a “relevant (+)” set and a “non-relevant (–)” set. Since relevant documents should appear earlier than non-relevant documents, the ranking error is given by: Total number of disordered pairs/Total number of item pairs between the relevant set and the non-relevant set. Below are a list of documents and their golden-standard relevance labels, the prediction results of a ranking model, and the prediction results of a binary classification model. Please calculate the bipartite ranking error and the binary classification error.​   Document ID Golden-standard relevance labels The predicted scores of the ranking model The predicted labels of the classification models D1 – 0.91 + D2 – 0.82 + D3 + 0.73 + D4 + 0.64 + D5 + 0.55 + D6 + 0.46 + D7 – 0.37 – D8 – 0.28 –

Q4: (3 points) Answer True or False only: In AdaBoosting, subsequent decision stumps are trained by reweighing misclassified samples of the previous decision stumps. In this way, weak learners are combined parallelly to create a strong learner.

Q2: (4 points)What is the definition of a support vector?