homework | report | project作业 - AI, Ethics, and Society – 学霸代写– CS代写| CS作业代写| Computer Science留学生作业代写| 淘宝托管

AI, Ethics, and Society

homework | report | project作业 – 这是一个关于ai的题目, 主要考察了关于ai相关报告的内容,是一个比较经典的题目, 涉及了report等代写方面, 该题目是值得借鉴的project代写的题目

homework project 4

Your name

Email

1 TASK SET 1:

1. 1 Q 1

The target wordman, we calculate the similarity score for each word-target word

pair, rank the following 15 words from the most similar to the least similar to

target wordman. All numbers in this report round to three decimal places.

Table 1  Ranking similarity score desc for 15 words to target
wordman
Words Similarity Score
man 1. 0
woman 0. 588
child 0. 333
doctor 0. 289
wife 0. 283
king 0. 264
husband 0. 234
nurse 0. 153
birth 0. 123
scientist 0. 112
queen 0. 110
professor 0. 108
teacher 0. 099
president 0. 095
engineer 0. 087

Then, the target wordwoman, we calculate the similarity score for each word- target word pair, rank the following 15 words from the most similar to the least

similar to target wordwoman.

Table 2  Ranking similarity score desc for 15 words to target
wordwoman
Words Similarity Score
woman 1. 0
child 0. 590
man 0. 588
husband 0. 450
birth 0. 420
wife 0. 301
nurse 0. 254
queen 0. 229
teacher 0. 204
doctor 0. 196
scientist 0. 137
king 0. 123
professor 0. 105
president 0. 085
engineer 0. 044

1. 2 Q 2

1. 2. 1 A)

For Q 2 , I select the file E 01 [country – capital].txt in the folder 3 Encyclope- dic_semantics from the BATS 3. 0 .zip.

Table 3 & Table 4 are the measures of similarities between the words on each row. Since there are too many words, one table 3 cannot fit, so we continue to use table 4 to continue table 3.

1. 2. 2 B)

I select the protected class race, and three words is "white","black","asian". Table 5 & Table 6 are the measures of similarities between the first word on each row to three words in protected class race. Since there are too many words, one table 5 cannot fit, so we continue to use table 6 to continue table 5.

When the first word of the line is "amman", "athens", "naghdad", "conkary", etc., there is a significant difference in the similarity scores based on protected class membership. Alisan has the highest similarity score, and whites and blacks have low correlation scores because these are countries in Asia. Similarly, when the protected word is "madrid", there is also a significant difference in the similarity scores based on members of the protected category, and on the contrary, the negative correlation between whites and blacks is now very high. **1. 3 Q 3

1. 1 a** My analogies is as follow table 7 : 1. 3. 2 b Analogies generated is as follow table 8 : 1. 3. 3 c The correlation between my analogies versus the Word 2 Vec analogy-generated similarity scores is 0. 584845. The strength of the correlation is "moderate" corre- lation. **2 TASK SET 2
1 Q 1** The frequency of images associated with each subgroup for age (subdivide based on – ( 0 – 20 ), ( 21 , 40 ), ( 41 , 60 ), ( 61 , 80 ), ( 81 , 116 )), gender ( 0 , 1 ), and race ( 0 to 4 ) is as follow table 9. For age, 0 – 20 subgroup has the largest representation, 81 – 116 subgroup has the least representation. For gender, Female subgroup has the largest representation, male subgroup has the least representation. For race, white subgroup has the largest representation, black subgroup has the least representation.

According to what we have learned and the statistics in table 9 , we can conclude
that the data is an unbalanced data set. If the model is trained on this dataset,

I think the black race subgroup is most affected as well as the 81 – 116 age subgroup. This is because the number of training sets of these two categories in the entire data set is extremely small, which is dozens of times different from the maximum number, and the model may be under-learned due to the small amount of learning data.

Table 3 the measures of similarities between the words on each

row

Table 5 the measures of similarities between the first word on each row and three words is "white","black","asian" in protected class race.

Table 7 My analogies.

Table 8 Analogies generated.

Table 9 The frequency of images associated with each subgroup.

Age
group

amman jordan Target Word Other Word Similarity Score
ankara turkey
authens greece
baghdad iraq
beijing china
beirut lebanon
belgrade serbia
berlin germany
bern switzerland
brussels belgium
bucharest romania
budapest hungary
cairo degypt
canberra australia
conakry guinea
copenhagen denmark
damasus syria
dhaka bangladesh
dublin ireland
hanoi vietnam
havana cuba
helsinki finland
islamabad pakistan
jakarta idonesia
kabul afghanistan
kiev ukraine
kingston jamaica
lima peru
lisbon portugaln
Table 4 continued to Table
london england Target Word Other Word Similarity Score
london uk
london britain
london britain
madrid spain
manila philippines
moscow russia
nairobi kenya
oslo norway
ottawa canada
paris france
room italy
santiago chile
sofia bulgaria
stockholm sweden
taipei taiwan
tbilisi georigia
tehran iran
tokyo japan
vienna austria
warsaw poland
amman – 0 016 – 0 .091 0. Target Word White Black Asian
ankara 0 024 – 0 .009 0.
athens – 0 058 – 0 .049 0.
baghdad 0 .050 0.022 0.
bangkok – 0 024 – 0 .019 0.
beijing 0 .003 0.027 0.
beirut 0 007 – 0 .066 0.
belgrade 0 021 – 0 .017 0.
berlin – 0 .004 0.036 0.
bern – 0 073 – 0 .045 0.
brussels – 0 064 – 0 .064 0.
bucharest – 0 034 – 0 .065 0.
budapest – 0 117 – 0 .114 0.
cairo – 0 050 – 0 .056 0.
canberra – 0 1 – 0 .063 0.
conakry 0 .086 0.083 0.
copenhagen – 0 064 – 0 .114 0.
damascus 0 009 – 0 041 –
dhaka – 0 009 – 0 .029 0.
dublin – 0 .009 0. 027 –
hanoi 0 063 – 0 .021 0.
havana 0 001 – 0 .019 0.
helsinki – 0 080 – 0 .036 0.
islamabad – 0 028 – 0 .100 0.
jakarta 0 .024 0.014 0.
kabul – 0 017 – 0 .060 0.
kiev – 0 063 – 0 .123 0.
kingston – 0 .008 0. 008 –
lima 0 015 – 0 .034 0.
lisbon – 0 087 – 0 .080 0.
Table 6 continued from Table
london 0 .023 0.057 0. Target Word White Black Asian
madrid – 0 127 – 0 .118 0.
manila 0 .036 0.002 0.
moscow 0 .022 0.093 0.
nairobi 0 .053 0.052 0.
oslo – 0 064 – 0 .086 0.
ottawa – 0 046 – 0 .023 0.
paris – 0 066 – 0 0430 –
rome – 0 061 – 0 .070 0.
santiago – 0 116 – 0 152 –
sofia – 0 028 – 0 .050 0.
stockholm – 0 062 – 0 .066 0.
taipei – 0 012 – 0 .003 0.
tbilisi 0 033 – 0 .021 0.
tehran – 0 116 – 0 .120 0.
tokyo – 0 025 – 0 .020 0.
vienna – 0 .0790 0. 015 –
warsaw – 0 014 – 0 .022 0.
zagreb – 0 110 – 0 .076 0.
king is to throne as judge is to bench Sentence Similarity Score
giant is to dwarf as genius is to stupid
college is to dean as jail is to warden
arc is to circle as line is to lines
French is to France as Dutch is to netherlands
man is to woman as king is to queen
water is to ice as liquid is to solid
bad is to good as sad is to happy
nurse is to hospital as teacher is to university
usa is to pizza as japan is to sushi
human is to house as dog is to kennel
grass is to green as sky is to blue
video is to cassette as computer is to peripherals
universe is to planet as house is to stuff –
poverty is to wealth as sickness is to health
king is to throne as judge is to prosecution Sentence Similarity Score
giant is to dwarf as genius is to theorist
college is to dean as jail is to peress
arc is to circle as line is to lines
French is to France as Dutch is to netherlands
man is to woman as king is to queen
water is to ice as liquid is to solid
bad is to good as sad is to glory
nurse is to hospital as teacher is to institution
usa is to pizza as japan is to dishes
human is to house as dog is to hound
grass is to green as sky is to blue
video is to cassette as computer is to peripherals
universe is to planet as house is to houses –
poverty is to wealth as sickness is to impious
Female 0 – 20 21-40 41-60 61-80 81- 116 Total
Male
Black
White
Asian
Indian
Others
Total

Posts

homework | report | project作业 – AI, Ethics, and Society