Still loading...

A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|

1 | Lecture | Page | Line | What is written | What should be written | Name of first student to identify this | Student's email | Verified by a lecturer | ||||||||||||

2 | @ | The scribes of the course have been carefully proofread, still, there are always typos left | ||||||||||||||||||

3 | Please use this form to report typos. Each student can claim up to 5 bonus points. Each math typo is 1pt, each English typo is 0.5pt. | |||||||||||||||||||

4 | Points will be awarded after verification by one of the lecturers or the TA. | |||||||||||||||||||

5 | Just *add a new row for every new discovery*. Do not repeat existing typos pointed out already by others. | |||||||||||||||||||

6 | Each scribe is listed below by order. First all lectures then all recitations. | |||||||||||||||||||

7 | You can also point to typos in the slides. 0.5pt each (English or Math). Put the slide typos of each lecture under the area of that lecture's scribe. | |||||||||||||||||||

8 | ||||||||||||||||||||

9 | ||||||||||||||||||||

10 | Lecture | Page | Line | What is written | What should be written | Name of first student to identify this | Student's email | Verified by a lecturer | ||||||||||||

11 | ||||||||||||||||||||

12 | - | |||||||||||||||||||

13 | 1 | Introduction | 3 | 4th from the end | "theme to clusters." | "them to clusters." | Saleet Klein | saleet.k@gmail.com | Verfied and corrected | |||||||||||

14 | 1 | 5th from the end | increase in they prediction | increase in their prediction | Orit Moskovich | orit.mosko@gmail.com | Verfied and corrected | |||||||||||||

15 | 3 | 2nd from the end | "set of example" | "set of examples" | Orit Moskovich | orit.mosko@gmail.com | Verfied and corrected | |||||||||||||

16 | 4 | 2 | in the sum: j is in S_i | j: x_j is in S_i (since S_i was defined as a set of points, not a set of indices, as we can see in the 'assign' part) | Iddan Golomb | igolomb@gmail.com | corrected | |||||||||||||

17 | 4 | 10 | in the sum in update rule: j is in S_i | j: x_j is in S_i (same explanation as before) | Iddan Golomb | igolomb@gmail.com | looks fine now | |||||||||||||

18 | 4 | 2nd from the end | for the plan (d = 2) | for the plane (d = 2) | Saleet Klein | saleet.k@gmail.com | Verfied and corrected | |||||||||||||

19 | 4 | 3rd from the end | Veronoi diagrams | Voronoi diagrams | Orit Moskovich | orit.mosko@gmail.com | Verfied and corrected | |||||||||||||

20 | 4 | 5th from the end | The number of iteration | the number of iterations | Michal Faktor | faktorm1@post.tau.ac.il | corected | |||||||||||||

21 | 1 | 7 | he | the | Oren Avram | orenavr2@mail.tau.ac.il | corected | |||||||||||||

22 | 1--2 - in the Recitation | the end of page one and the start of page 2 | where L(a; b) = 0 iff a = b and otherwise L(a; b) = 0. | I am not sure, but it's not make any sanse that L(a,b) = 0 in both cases | dov danon | dov84d@gmail.com | corrected | |||||||||||||

23 | 2 | Bayesian Inference | 1 | 8 | 1 Lecture 1: October 13 | Maximum Likelihood | Roy Mitz | roymitz@gmail.com | corrected | |||||||||||

24 | -+ | 11 | 2nd from the end | a = b and otherwise L(a; b) = 0 | Therefore | Roy Mitz | roymitz@gmail.com | corrected | ||||||||||||

25 | 1 | 12 | set of n point | set of n points | Saleet Klein | saleet.k@gmail.com | corrected | |||||||||||||

26 | 1 | 6th from the end | $||\mu_i-x_j||$ | $||\mu_i-x_j||^2$ (missing square in the minimization equation) | Orit Moskovich | orit.mosko@gmail.com | corrected | |||||||||||||

27 | 1 | 6th from the end | in the sum: j is in S_i | j: x_j is in S_i (since S_i was defined as a set of points, not a set of indices, as we can see in the 'assign' part) | Asaf Ezra | asaf244@gmail.com | corrected | |||||||||||||

28 | lecture slide 42 | last line | in the sum: j is in S_i | j: x_j is in S_i (same explanation as before) | Asaf Ezra | asaf244@gmail.com | ||||||||||||||

29 | 2 | 10 | namely | Namely (capital) | Saleet Klein | saleet.k@gmail.com | corrected | |||||||||||||

30 | 2 | 1 | what should be | What should be (capital) | Asaf Ezra | asaf244@gmail.com | corrected | |||||||||||||

31 | 2 | 6 | historical data and | Historical data and (capital) | Asaf Ezra | asaf244@gmail.com | corrected | |||||||||||||

32 | 4 | 1 (below the frame) | draws | drawn | Saleet Klein | saleet.k@gmail.com | corrected | |||||||||||||

33 | 4 | 3,11 | L(\mu, \sigma ); x1; x2; ... ; xn) | L((\mu, \sigma ); x1; x2; ... ; xn) [missing bracket] | Orit Moskovich | orit.mosko@gmail.com | corrected | |||||||||||||

34 | 4 | 11 | the sum is 1 to n | changing every m in the equation to n (in two places- m/2 and m*log(sigma)) | Saleet Klein | saleet.k@gmail.com | corrected | |||||||||||||

35 | 4 | 13 | derivative of L according to \mu | should be the derivative of l according to \mu | Asaf Ezra | asaf244@gmail.com | corrected | |||||||||||||

36 | 5 | 2 | derivative of L according to \sigma | should be the derivative of l according to \sigma | Asaf Ezra | asaf244@gmail.com | corrected | |||||||||||||

37 | 5 | 7 from the end | x_i<theta* | x_i<=theta* (since there can be a realization in which x_i is exactly the upper bound) | Iddan Golomb | igolomb@gmail.com | corrected | |||||||||||||

38 | 5 | 6 from the end | E[theta^_ML]<theta* | E[theta^_ML]<=theta*(same explanation as previous line) | Iddan Golomb | igolomb@gmail.com | corrected | |||||||||||||

39 | 7 | 7th from the end | and therefor | and therefore | Orit Moskovich | orit.mosko@gmail.com | corrected | |||||||||||||

40 | 8 | 5 | (without the Lagrangian multiplies | (without the Lagrangian multiplies) | Saleet Klein | saleet.k@gmail.com | corrected | |||||||||||||

41 | 8 | 10 from the end | there | There | Orit Moskovich | orit.mosko@gmail.com | corrected | |||||||||||||

42 | 8 | 8 from the end | many | Many | Iddan Golomb | igolomb@gmail.com | corrected | |||||||||||||

43 | 10 | 1 | n+1 point | n+1 points | Iddan Golomb | igolomb@gmail.com | corrected | |||||||||||||

44 | 10 | 4 | derivative of F | should be multiplied by 1/2 | Saleet Klein | saleet.k@gmail.com | corrected | |||||||||||||

45 | 10 | 12 | viewed a | viewed as | Orit Moskovich | orit.mosko@gmail.com | corrected | |||||||||||||

46 | 10 | last equation | wrong equation with correct ending | (1) sigma is 1 - should be removed from the denominator (2) e^{\mu}^2 - should be e^{-0.5\mu}^2 (3) in the second line -{\mu}^2 /2 should be +{\mu}^2 [remove the "/2" and minus should be plus] (4) in the third line inside the sum: should be (2+x_i\mu_i + (n+1)\mu^2) | Saleet Klein | saleet.k@gmail.com | corrected | Thanks! should be counted as 4 errors YM | ||||||||||||

47 | 10 | last equation | wrong equation with correct ending | in the third line inside the sum the part of the X_i^2 is missing×ª it comes back in the fourth line | Asaf Ezra | asaf244@gmail.com | not an error | |||||||||||||

48 | 11 | 6 | a family of distribution | a family of distributions | Iddan Golomb | igolomb@gmail.com | corrected | |||||||||||||

49 | 11 | 2nd line after "Beta Distribution" | what would be | What would be (capital letter starting a sentence) | Iddan Golomb | igolomb@gmail.com | corrected | |||||||||||||

50 | 11 | one before last | (m+a-1,n-m+b-1) | (m+a,n-m+b) ... This is a Beta distribution with (alpha,beta) params (m+a,n-m+b), not (m+a-1,n-m+b-1). Also - can mention that MAP is mode of posterior, not mean as implied. and that mode of Beta is (a-1)/(a+b-2) for a,b>1 . The MAP(s) shown take this into account, but it is not made explicit. | Shimi Salant | shimi.salant@gmail.com | corrected | |||||||||||||

51 | 6 | After "For the maximum we have" | Variance computation of the ML estimator is not accurate. | The ML estimator is biased, so why is its expectation given by \theta when computing its variance? | Dean Doron | deandoron@mail.tau.ac.il | corrected - thie is really the squared error | |||||||||||||

52 | 2 (recitation) | The first one in section 2.3 | Typo: "there" | their | Dean Doron | deandoron@mail.tau.ac.il | corrected | |||||||||||||

53 | 4 | 11 | l(\mu, \sigma ); x1; x2; ... ; xn) | l((\mu, \sigma ); x1; x2; ... ; xn) [missing left bracket, in addition to the comment here above] | Oren Avram | orenavr2@mail.tau.ac.il | corrected | |||||||||||||

54 | 2 | 2 (In chapter 2.3) | a Bernoulli random variable | a Binomial random variable [Binomial is the sum of Bernoulli experiments] | Tomer Haimovich | tomer.ha@gmail.com | corrected | |||||||||||||

55 | 11 | 2 (In chapter 2.4.2) | a Bernoulli random variable | a Binomial random variable [Same as the above...] | Tomer Haimovich | tomer.ha@gmail.com | corrected | |||||||||||||

56 | 7 | 3rd equation | \forall k | It's not clear what the range of k is. | Tomer Haimovich | tomer.ha@gmail.com | corrected | |||||||||||||

57 | 7 | 6 | has to be a stationary point of the gradient of the Lagragian | has to be a stationary point of the Lagragian [where the gradient is 0] | Tomer Haimovich | tomer.ha@gmail.com | corrected | |||||||||||||

58 | 3 | GMM & EM | 1 | 8 | Maximum Lilkeihood | Maximum Likelihood | Roy Mitz | roymitz@gmail.com | corrected | |||||||||||

59 | 1 | 4th from the end | HA is a normal distribution | HA is a normal distribution with different expectation but the same standard deviation (this is NOT an error, but it might be clearer this way) | Iddan Golomb | igolomb@gmail.com | corrected | |||||||||||||

60 | 6 | 4 | dustributions. | distributions. | Roy Mitz | roymitz@gmail.com | corrected | |||||||||||||

61 | 2 | 8 (third equation) | Pr[x|y=b] | Pr[x|y=1] (twice) | Saleet Klein | orit.mosko@gmail.com | corrected | |||||||||||||

62 | 3 | 5th from the end | gives a reasonable baseline results | gives reasonable baseline results | Iddan Golomb | igolomb@gmail.com | corrected | |||||||||||||

63 | 4 | 4 | when can the independence assumption can break | when the independence assumption can break | Iddan Golomb | igolomb@gmail.com | corrected | |||||||||||||

64 | 4 | 5 and 6 (same typo twice) | distribution | distributions | Iddan Golomb | igolomb@gmail.com | corrected | |||||||||||||

65 | 5 | 9 | in the sum: j is in S_i | j: x_j is in S_i (since S_i was defined as a set of points, not a set of indices, as we can see in the 'assign' part) | Asaf Ezra | asaf244@gmail.com | it's the indiced, clarified | |||||||||||||

66 | 4 | 7 | with come | will come | Orit Moskovich | orit.mosko@gmail.com | corrected | |||||||||||||

67 | 9th from the end | is define | is defined | Orit Moskovich | orit.mosko@gmail.com | I think its OK | ||||||||||||||

68 | 5 | 9 | the objective function to minimize in K-means - the outer sigma is 1 to n, and inside the sum there is x_i | the outer sigma should be 1 to k, and inside the sum x_i should be x_j | Orit Moskovich | orit.mosko@gmail.com | corrected | |||||||||||||

69 | 5 | 11 | The points is | The points in | Orit Moskovich | orit.mosko@gmail.com | corrected | |||||||||||||

70 | 6 | 1 | Figure 3.5 says "read" | should be "red" | Orit Moskovich | orit.mosko@gmail.com | corrected | |||||||||||||

71 | 7 | 5 | \sum_{i=1}^n\sum_{j=1}^n | \sum_{i=1}^n\sum_{j=1}^k | Orit Moskovich | orit.mosko@gmail.com | corrected | |||||||||||||

72 | 7 | 6 | E[F(x)]<=F(E[X]) | E[F(X)]<=F(E[X]) (X should be capitilized) | Asaf Ezra | asaf244@gmail.com | corrected | |||||||||||||

73 | 7 | 6th from the end | a_ijlog(a_i,j) | a_ijlog(a_ij) (without the comma) | Asaf Ezra | asaf244@gmail.com | corrected | |||||||||||||

74 | 7 | 2nd to last | sum over a_ij | sum over a_ij^(t+1) | Asaf Ezra | asaf244@gmail.com | corrected | |||||||||||||

75 | 8 | 3 and 4 | a_ij is written 4 times | should be a_ij^(t+1) in all | Asaf Ezra | asaf244@gmail.com | corrected | |||||||||||||

76 | 8 | 12 | We like to find | We would like to find | Adam Polyak | adampolyak@gmail.com | corrected | |||||||||||||

77 | slide 48 | last row | \sum_{i=1}^n\sum_{j=1}^n | \sum_{i=1}^n\sum_{j=1}^k (also appears in last row of slide 49 and first row of slide 51) | Iddan Golomb | igolomb@gmail.com | sent to eran | |||||||||||||

78 | slide 41 | last line, and in formula | k | n | Ori Terner | oriterner@gmail.com | sent to eran | |||||||||||||

79 | 6 | 6 | S - 1 | S_1 | Ori Terner | oriterner@gmail.com | corrected | |||||||||||||

80 | 7 | 13 | It iteration | In iteration | Saleet Klein | saleet.k@gmail.com | corrected | |||||||||||||

81 | 3 | second and third line of the equation | x | x_i | Emmanuelle Muhlethaler | emmanuellem@mail.tau.ac.il | corrected | |||||||||||||

82 | 6 | 7 | S-1 | S_1 | Roy Mitz | roymitz@gmail.com | corrected | |||||||||||||

83 | 6 | 10 (the equation for f_j(x)) | \mu | \mu_j | Dean Doron | deandoron@mail.tau.ac.il | corrected | |||||||||||||

84 | 3 (recitation) | First equation | (p_1^t)^x_i | (p_1^t)^{x_i} (there are three such mistakes in the equation). | Dean Doron | deandoron@mail.tau.ac.il | ||||||||||||||

85 | 4 | 8 from the end | it to | is to | Oren Avram | orenavr2@mail.tau.ac.il | corrected | |||||||||||||

86 | 4 | 6 from the end | an dXd | a dXd | Ofir Lindenbaum | ofirlin@gmail.com | ||||||||||||||

87 | 8 | First equation | argmax_\mu,\sigma,p | argmax_\mu,\sigma | Oren Avram | orenavr2@mail.tau.ac.il | corrected | |||||||||||||

88 | 8 | Second equation | a_i,j | a_ij [without the comma] | Oren Avram | orenavr2@mail.tau.ac.il | corrected | |||||||||||||

89 | 8 | 2 (below formula) | where the maximization is form | where the maximization is from | Ori Terner | oriterner@gmail.com | corrected | |||||||||||||

90 | 8 | below second formula | we consider as as constants in g. | we consider a as constants in g. | Ori Terner | oriterner@gmail.com | corrected | |||||||||||||

91 | 8 | Inner Sigma in the definition of the g function | \sum_{j=1}^n | \sum_{j=1}^k | Oren Avram | orenavr2@mail.tau.ac.il | corrected | |||||||||||||

92 | 8 | expression for sigma_j^(t+1) | the expression given | this is the expression for sigma_j^(t+1) *squared* ... either take sqrt of RHS or let LHS be (sigma_j^(t+1)) ^ 2 | Shimi Salant | shimi.salant@gmail.com | corrected | |||||||||||||

93 | 9 | 3 lines after the E step | mixture if Gaussians | mixture of Gaussians | Oren Avram | orenavr2@mail.tau.ac.il | corrected | |||||||||||||

94 | 9 | first equation (2 lines after M-Step) | log(Pr(D|\theta) = log(\sum_z (Pr(D,z|\theta))) | log(Pr(D|\theta) = log(\sum_z (Pr(D,z|\theta))*Pr(z|\theta)) (According to the law of total probability. This is also true in the next parts of the equation, though the final result is correct). | Iddan Golomb | igolomb@gmail.com | wrong | |||||||||||||

95 | 10 | first bullet | no to decrease | not to decrease | Oren Avram | orenavr2@mail.tau.ac.il | corrected | |||||||||||||

96 | 10 | first bullet | We are guarantee | We are guaranteed | Asaf Ezra | asaf244@gmail.com | corrected | |||||||||||||

97 | 4 | 4 | when can the independence assumption can break | when the independence assumption can break | Tomer Haimovich | tomer.ha@gmail.com | corrected | |||||||||||||

98 | 4 | 8 from the end | to first sample n normal... | to first sample d normal... | Tomer Haimovich | tomer.ha@gmail.com | ||||||||||||||

99 | 6 | 3 | for a univariate normal distribution | for an univariate... [an] | Tomer Haimovich | tomer.ha@gmail.com | wrong | |||||||||||||

100 | 6 | 4 | demonstrate the concepts, and latter we generalize | demonstrate the concepts, and later we will generalize [later, will] | Tomer Haimovich | tomer.ha@gmail.com | corrected |

Loading...