Questions and Answers of Artificial Intelligence A Modern approach

Figure S13.3 shows pairs of Bayes nets. In each, the original network is shown on the left. The reversed network, shown on the right, has all the arrows reversed. Therefore, the reversed network may
Exercise Exercise 12.FISH introduced the naive Bayes model in Figure S13.6. (F) is true iff today was a good day of fishing. There are three features: whether it rained (R), how many fish were caught
a. For the Bayes net in Figure S13.37, we are given the query P(Z | + y). All variables are Boolean. Assume we run variable elimination to compute the answer to this query, with the following
Assume the Bayes net, and the corresponding distributions over the variables in the Bayes net from Figure S13.46. a. Your task is now to estimate P(+y | + x, +z) using rejection sampling. Below
Recall that for a standard HMM the Elapse Time update and the Observation update are of the respective forms:P(Xt | e1:t−1) = Σxt−1 P(Xt | xt−1)P(xt−1 | e1:t−1)P(Xt | e1:t) ∝ P(Xt |
Consider the following sentence (from The New York Times, July 28, 2008):Banks struggling to recover from multibillion-dollar loans on real estate are curtailing loans to American businesses,
Consider a robot with two simple manipulators, as shown in figure ??. Manipulator A is a square block of side 2 which can slide back and on a rod that runs along the x-axis from x = −10 to x = 10.
Consider the simplified robot shown in Figure ??. Suppose the robot’s Cartesian coordinates are known at all times, as are those of its goal location. However, the locations of the obstacles are
This exercise explores the relationship between workspace and configuration space using the examples shown in Figure ??.a. Consider the robot configurations shown in Figure ??(a) through (c),
Go through Turing’s list of alleged “disabilities” of machines, identifying which have been achieved, which are achievable in principle by a program, and which are still problematic because
Alan Perlis (1982) wrote, “A year spent in artificial intelligence is enough to make one believe in God”. He also wrote, in a letter to Philip Davis, that one of the central dreams of computer
I. J. Good claims that intelligence is the most important quality, and that building ultra intelligent machines will change everything. A sentient cheetah counters that “Actually speed is more
Concerns regarding the effect of technology on the future of work are not new. In a 1964 memo to President Lyndon B. Johnson, the Ad Hoc Committee of the Triple Revolution (1964), one comprised of
Some critics object that AI is impossible, while others object that it is too possible and that ultra intelligent machines pose a threat. Which of these objections do you think is more likely? Would
Consider the Bayes net in Figure S13.36 with the query P(Q | + e). a. Which variables can be ignored when computing the answer to the query? b. Prove a general result concerning the
A smell of sulphur (S) can be caused either by rotten eggs (E) or as a sign of the doom brought by the Mayan Apocalypse (M). The Mayan Apocalypse also causes the oceans to boil (B). The Bayesian
Suppose that a patient can have a symptom (S) that can be caused by two different diseases (A and B). It is known that the variation of gene G plays a big role in the occurrence of disease A. The
Consider a Bayes net over the random variables A, B, C, D, E with the structure shown in Figure S13.2, with full joint distribution P(A, B, C, D, E). a. Consider the marginal distribution P(A,
Scientific inquiry investigates the causal relation of variables; scientists study the effects of interventions. Causal networks provide a framework for such analysis by exactly specifying
PacLabs has just created a new type of mini power pellet that is small enough for Pacman to carry around with him when he’s running around mazes. Unfortunately, these mini-pellets don’t guarantee
Suppose that a patient can have a symptom (S) that can be caused by two different diseases (A and B). Disease A is much rarer, but there is a test T that tests for the presence of A. The Bayes net
Consider the Bayes net and corresponding probability tables shown in Figure S13.50. Fill in the following table with the probabilities of drawing each respective sample given that we are using each
You are an exobiologist, studying the wide range of life in the universe. You are also an avid dancer and have an excellent model of the way species invent dancing. The key variables are:Sound
a. Express the joint probability distribution P(A, B, C) induced by the Bayes net in Figure S13.32 in terms of the associated conditional probability distributions.Compute the values of the following
Consider the joint distribution P(A, B, C, D) defined by the Bayes net in Figure S13.33. Compute the values of the following quantities: a. P(A = true). b. P(A = true, B = false, C = false,
Assume you are given the Bayes net and the corresponding CPTs shown in Figure S13.51. a. Assume we receive evidence that A = + a. If we were to draw samples using rejection sampling, what is the
Consider the sequence of graphs in Figure S13.38 and the application of variable elimination to each in order to compute P(X). For each, regardless of the elimination ordering, the largest factor
In this question, we consider the efficiency of variable elimination as a function of the elimination ordering. For any variable, X, let |X| denote the size of X’s range (the number of values it
A k-zigzag network has k Boolean root variables and k + 1 Boolean leaf variables, where root i is connected to leaves i and i + 1. Figure S13.52 shows an example for k = 3, where each Di represents a
A tree-augmented Naive Bayes model (TANB) is identical to a Naive Bayes model, except the features are no longer assumed conditionally independent given the class Y. Specifically, if (X1, X2, . . . ,
Research the YOLO (You Only Look Once) approach to object detection, and compare it to the RCNN region proposal approach described in the book.
An augmented context-free grammar can represent languages that a regular context-free grammar cannot. Show an augmented context-free grammar for the language an bn cn. The allowable values for
A stereoscopic system is being contemplated for terrain mapping. It will consist of two CCD cameras, each having 512 × 512 pixels on a 10 cm × 10 cm square sensor. The lenses to be used have a
Define the following terms in your own words.a. Active and passive sensingb. Image featurec. Object modeld. Rendering model.
In the shadow of a tree with a dense, leafy canopy, one sees a number of light spots. Surprisingly, they all appear to be circular. Why? After all, the gaps between the leaves through which the sun
Suppose you have a CNN object detection algorithm that runs very well in a data center on a large cluster of computers. Now you want to deploy it onto cell phones (without draining users’
Edges in an image can correspond to a variety of events in a scene. Consider Figure 25.4 (page 887), and assume that it is a picture of a real three-dimensional scene. Identify ten different
Object detectors are commonly evaluated with a metric known as mean average precision (mAP). Research this method and explain how it works.
Name as many perceptual channels as you can that are used by robots today.
Which of the following are true, and which are false? a. Finding corresponding points in stereo images is the easiest phase of the stereo depthfinding process. b. Shape-from-texture can be
Consider a picture of a white sphere floating in front of a black backdrop. The image curve separating white pixels from black pixels is sometimes called the “outline” of the sphere. Show that
The Waltz algorithm takes as input a list of the vertices in a sketch diagram of one or polyhedrons. Each vertex is described by the number of lines that meet and whether the lines are colinear. From
A new object detection approach called CenterNet models an object as a single point, the center of its bounding box. Research the approach and report on your findings.
(Courtesy of Pietro Perona.) Figure ?? shows two cameras at X and Y observing a scene. Draw the image seen at each camera, assuming that all named points are in the same horizontal plane. What can be
Consider an infinitely long cylinder of radius r oriented with its axis along the y-axis. The cylinder has a Lambertian surface and is viewed by a camera along the positive z-axis. What will you
Crowd counting is the task of estimating the number of people in an image (or images). Whenever there is a big event, organizers of the event tend to exaggerate the crowd size, and competitors try to
mage classification relies on labeled image data. Where does that come from? How accurate are the labels? How hard is it to create the labels? Read http://karpathy.github.io/2014/09/02/
Normally we think of more depth of field as a good thing: we want more of the image in focus. But some photographers prize shallow depth of field, because it makes the subject of the photograph stand
Does a small cell phone camera lens with an aperture of f/2 gather as much light from each patch in the scene as an f/2 lens on a large SLR camera that is taking an image of the same scene?
Suppose that you are working with the robot in Exercise 26.ABMA above and you are given the problem of finding a path from the starting configuration of figure ?? to the ending configuration.
(This exercise was first devised by Michael Genesereth and Nils Nilsson. It works for first graders through graduate students.) Humans are so adept at basic household tasks that they often forget how
Attempt to write definitions of the terms “intelligence,” “thinking,” and “consciousness.” Suggest some possible objections to your definitions.
Does a refutation of the Chinese room argument necessarily prove that appropriately programmed computers have mental states? Does an acceptance of the argument necessarily mean that computers cannot
How do the potential threats from AI technology compare with those from other computer science technologies, and to bio-, nano-, and nuclear technologies?
Why might state actors support or oppose autonomous weapons, as discussed in Section ??? What role can technologists play in this debate?
The text describes some of the unintended side effects which might result from artificial intelligence technologies, perhaps from externalities in the objective function of a system. Here consider
Analyze the potential threats from AI technology to society. What threats are most serious, and how might they be combated? How do they compare to the potential benefits?
Compare the social impact of artificial intelligence in the last fifty years with the social impact of the introduction of electric appliances and the internal combustion engine in the fifty years
Recall the discussion of transhumanism from the text. Similar considerations have been made with regard to cyborgs, cybernetic organisms. Either individually or in a discussion group,
Compare the price of various component technologies used in AI, and create charts for how those prices have changed over the last twenty years.
In the network in Figure S13.45, identify the Markov blanket of each variable. Figure S13.45 A H E B 7 F C J G K
Recall that any directed acyclic graph G has an associated family of probability distributions, which consists of all probability distributions that can be represented by a Bayes net with structure
In this problem, we will play economist. Consider four variables, price (P), demand (Q), income (I), and wages (W). More specifically, where Q is the quantity of household demand for a product A, P
In this chapter we present Bayesian networks, originally called belief networks, for the representation of uncertainty but other types of graphical models may also be used. Consider, Markov networks,
Prove that if there is no direct path between two variables, X and Y in a causal network, that there is no causal effect between them.
Consider a simple Bayesian network with root variables Cold, Flu, and Malaria and child variable Fever , with a noisy-OR conditional distribution for Fever as described in Section 13.2.2. By adding
For the Bayes net shown in Figure S13.42, consider the query P(A|h), and the variable elimination ordering B, E, C, F, D. a. In the table below fill in the factor generated at each step—we did
Assume we are running variable elimination, and we currently have the following three factors:The next step in the variable elimination is to eliminate B. a. Which factors will participate in
In the off-switch problem (Section 16.7.2), we have assumed that Harriet acts rationally. Suppose instead that she is Boltzmann-rational, i.e., she follows a randomized policy that chooses action x
Valerie has just found a cookie on the ground. She is concerned that the cookie contains raisins, which she really dislikes but she still wants to eat the cookie. If she eats the cookie and it
We have described three policies for the vacuum robot: (1) A uniform random walk, (2) A bias for wandering southeast, as described in Exercise 14.HMMR, and (3) The policy described in
Consider a decision network with the following structure, where node U is the utility:a. In the graph above, how do we know if a node is guaranteed to have V P I = 0?b. Can any node be guaranteed to
A vehicle is trying to identify the situation of the world around it using a set of sensors located around the vehicle. Each sensor reading is based off of an object’s location (LOC) and an
Suppose that a particular student shows up with red eyes and sleeps in class every day. Given the model described in Exercise 14.SLEP, explain why the probability that the student had enough sleep
Consider the DBN in Figure 14.13(b). In the chapter, the battery level Batteryt and the battery meter reading BMetert are assumed to be integer-valued with a range of 0 to 5. In this exercise, we
(Adapted from David Heckerman.) This exercise concerns the Almanac Game, which is used by decision analysts to calibrate numeric estimation. For each of the questions that follow, give your best
Figure ?? shows a myopic function INFORMATION-GATHERING-AGENT(t) returns hat chooses the next evidence variable to observe according to the ratio VPI(Ej)/C (Ej), where C (Ej) is the cost of observing
You are the latest contestant on Monty Hall’s game show, which has undergone a few changes over the years. In the game, there are n closed doors: behind one door is a car (U(car) = 1000), while the
Consider the following lotteries: • L1 = [1, 1]. • L2 = [0.5, 2; 0.5, 0]. • L3 = [1, 2].Four students have expressed their preferences over these lotteries as follows: •
True or False: Assume Agent 1 has a utility function U1 and Agent 2 has a utility function U2. If U1 = k1U2 + k2 with k1 > 0, k2 > 0 then Agent 1 and Agent 2 have the same preferences.
PacBaby just found a $100 bill—it is the only thing she owns. Ghosts are nice enough not to kill PacBaby, but when they find PacBaby they will steal all her money. The probability of the ghosts
Figure 17.13 shows two MDPs: one, M, represents a two-armed bandit where one has the choice to continue with the first arm or to switch permanently to a second arm with fixed reward λ; the other, a
Your friend claims that he can write an effective Naive Bayes spam detector with only three features: the hour of the day that the email was received (H ∈ {1, 2, . . . , 24}), whether it contains
Let the initial belief state b0 for the 4 × 3 POMDP on page 588 be the uniform distribution over the nonterminal states, i.e., (1/9 , 1/9 , 1/9 , 1/9 , 1/9 , 1/9 , 1/9 , 1/9 , 1/9 , 0, 0). Calculate
Consider training the Naive Bayes model shown on the left with the training data provided in the table on the right.Calculate the maximum likelihood estimate of P(F1 = 1 | Y = 0).
What is the time complexity of d steps of POMDP value iteration for a sensorless environment?
Define the core of a cooperative game G = (N, v), and show by way of example that the core of a cooperative game may be empty.
Either prove or disprove each of the following statements in the context of 2 × 2 games (you may find it helpful to do proofs by providing examples or counter examples). a. If a player i has a
You have classification data with classes Y ∈ {+1, −1} and features Fi ∈ {+1, −1} for i ∈ {1, . . . , K}. Say you duplicate each feature, so now each example has 2K features, with FK+i = Fi
The Na¨ıve Bayes model has been famously used for classifying spam. We will use it in the “bag-of-words” model: • Each email has binary label Y which takes values in {spam,
Stoplights S1 and S2 can each be in one of two states: green (g) or red (r). Additionally, the machinery behind both stoplights (W) can be in one of two states: working (w) or broken (b). We collect
You want to learn a Bayes’ net over the random variables A, B, C. You decide you want to learn not only the Bayes’ net parameters, but also the structure from the data. You are willing to
Abhishek has been getting a lot of spam recently and is not satisfied with his email client’s Naive Bayes spam classifier. Thankfully, he knows about Bayes Nets and has decided to implement his own
You are given a model with two distinct label variables Y1, Y2, and there is a super label Z which conditions all of these labels, thus giving us this hierarchical na¨ıve Bayes model. The
Consider a naive Bayes classifier with two features, shown below. We have prior information that the probability model can be parameterized by λ and p, as shown below:We have a training set that
A softmax layer in a neural network takes an input vector x and produces an output vector y, where a. Obtain the derivatives ∂yj/∂xi in terms of x-values for the cases i = j and i ≠
Answer the following True/False questions. a. In the case of a binary class and all binary features, Naive Bayes a linear classifier. Justify your answer. b. Naive Bayes trained using
Consider the geometric distribution, which has P(X = k) = (1 − θ)k−1 θ. Assume in our training data X took on the values 4, 2, 7, and 9. a. Write an expression for the log-likelihood of
Identical twins are rare, but just how unlikely are they? With the help of the sociology department, you have a representative sample of twins to help you answer the question. The twins data gives
A softmax layer in a neural network takes an input vector x and produces an output vector y, whereShow that the sigmoid function is equivalent to a softmax with d = 2. yj = ej Σ=1 ek

Showing 1 - 100 of 303