How does artificial intelligence learn? - Briana Brownell

637,569 views ・ 2021-03-11

TED-Ed


Videoni ijro etish uchun quyidagi inglizcha subtitrlarga ikki marta bosing.

Translator: Dilnoza Nishanova Reviewer: Nazarbek Nazarov
00:09
Today, artificial intelligence helps doctors diagnose patients,
0
9829
5167
Bugun sun’iy intellekt shifokorlarga tashxis qo’yishda yordam bermoqda,
00:14
pilots fly commercial aircraft, and city planners predict traffic.
1
14996
5042
pilotlarga samolyot boshqarishda, muhandis -larga tirbandliklarni prognoz qilishda.
00:20
But no matter what these AIs are doing, the computer scientists who designed them
2
20038
4250
Lekin SI nima qilishidan qat’iy nazar, uni ishlab chiqqan dasturchilar
00:24
likely don’t know exactly how they’re doing it.
3
24288
2750
katta ehtimol bilan qanday qilayotganini bilmaydilar.
00:27
This is because artificial intelligence is often self-taught,
4
27038
3875
Buning sababi ko’p hollarda SI o’zi o’rgangan,
00:30
working off a simple set of instructions
5
30913
2374
oddiygina qo’llanmalar to’plamidan
00:33
to create a unique array of rules and strategies.
6
33287
3584
o’zgacha qoida va strategiyalar yaratadi.
00:36
So how exactly does a machine learn?
7
36871
2625
Xo’sh aynan qanday qilib mashina o’rganadi?
00:39
There are many different ways to build self-teaching programs.
8
39496
3000
O’zi o’rganadigan programmalar yaratishning ko’p yo’llari bor.
00:42
But they all rely on the three basic types of machine learning:
9
42496
3916
Lekin ularning barchasi uch asosiy mashina o’rganuvchi sistemaga asoslanadi:
00:46
unsupervised learning, supervised learning, and reinforcement learning.
10
46412
5000
nazoratsiz o’rganish, nazoratli o’rganish va mustahkam o’rganish.
00:51
To see these in action,
11
51412
1959
Bularni hayotda ko’rish uchun,
00:53
let’s imagine researchers are trying to pull information
12
53371
3250
tadqiqotchilarni minglab bemorlar tibbiy ma’lumotlari orasidan
00:56
from a set of medical data containing thousands of patient profiles.
13
56621
4208
informatsiya olishini tasavvur qilaylik.
01:01
First up, unsupervised learning.
14
61371
3125
Birinchi o’rinda, nazoratsiz o’rganish.
01:04
This approach would be ideal for analyzing all the profiles
15
64496
3333
Bu usul hamma profillarni umumiy o’xshashlik va
01:07
to find general similarities and useful patterns.
16
67829
3625
foydali yo’llarni topish uchun idealdir.
01:11
Maybe certain patients have similar disease presentations,
17
71454
3417
Balkim ba’zi bemorlarda o’xshash kasallik belgilar bor
01:14
or perhaps a treatment produces specific sets of side effects.
18
74871
4042
yoki ehtimol davolanishning ba’zi nuqsonlari bo’ladi.
01:18
This broad pattern-seeking approach can be used to identify similarities
19
78913
4125
Bu keng yo’l topuvchi usul inson aralashuvisiz
01:23
between patient profiles and find emerging patterns,
20
83038
3375
bemorlarning ma’lumotlaridan o’xshashlikni
01:26
all without human guidance.
21
86413
2291
topishda ishlatish mumkin.
01:28
But let's imagine doctors are looking for something more specific.
22
88704
3209
Lekin tasavvur qiling shifokorlar aniqroq narsani izlashmoqda.
01:32
These physicians want to create an algorithm
23
92371
2208
Bu shifokorlar ma’lum bir kasallikka
01:34
for diagnosing a particular condition.
24
94579
2709
tashxis qo’yish uchun algoritm yaratsishmoqchi.
01:37
They begin by collecting two sets of data—
25
97288
2583
Ular 2 to’plam ma’lumotlarni yig’ishdan boshlashadi--
01:39
medical images and test results from both healthy patients
26
99871
3375
ikkala sog’ bemorlar va kasallikka chalingan bemorlardan
01:43
and those diagnosed with the condition.
27
103246
2417
tibbiy rasmlar va test natijalarini olishadi.
01:45
Then, they input this data into a program
28
105663
2583
So’ng, bu ma’lumotlarni kasal bemorlarning
01:48
designed to identify features shared by the sick patients
29
108246
3375
xususiyatlarini aniqlash uchun dasturga kirg’iziladi
01:51
but not the healthy patients.
30
111621
2125
lekin sog’ bemorlarniki emas.
01:53
Based on how frequently it sees certain features,
31
113746
3083
Ma’lum xususiyatlarni qanchalik tez uchratishiga qarab,
01:56
the program will assign values to those features’ diagnostic significance,
32
116829
4042
dastur har bir xususiyatning muhimligiga qarab diagnostik qymat berib chiqadi,
02:00
generating an algorithm for diagnosing future patients.
33
120871
3708
va bemorlarga tashxis qo’yuvchi algoritm ishlab chiqaradi.
02:04
However, unlike unsupervised learning,
34
124579
3167
Ammo, boshqaruvsiz o’rganishdan farqli ravishda,
02:07
doctors and computer scientists have an active role in what happens next.
35
127746
4625
shifokor va dasturchilar keyingi jarayonda faol rol o’ynaydilar.
02:12
Doctors will make the final diagnosis
36
132371
2083
Shifokorlar oxirgi tashxisni qo’yadilar va
02:14
and check the accuracy of the algorithm’s prediction.
37
134454
2917
algoritm natijasining aniqligini tekshiradilar.
02:17
Then computer scientists can use the updated datasets
38
137871
2917
Dasturchilar yangi ma’lumotlarni dastur
02:20
to adjust the program’s parameters and improve its accuracy.
39
140788
3500
paramaterlarini to’g’irlash va aniqligini yaxshilash uchun ishlatishlari
02:24
This hands-on approach is called supervised learning.
40
144663
3166
Bu amaliy usul boshqaruvsiz o’rganish deyiladi.
02:27
Now, let’s say these doctors want to design another algorithm
41
147829
3167
Endi, ayataylik shifokorlar boshqa davolash rejasi
02:30
to recommend treatment plans.
42
150996
1750
algoritm tuzmoqchilar.
02:32
Since these plans will be implemented in stages,
43
152746
2833
Bu rejalar qadamma-qadam borar ekan va har bir
02:35
and they may change depending on each individual's response to treatments,
44
155579
3959
bemorning holatiga qarab o’zgarar ekan, shifokorlar
02:39
the doctors decide to use reinforcement learning.
45
159538
2833
mustahkamlovchi o’rganishni ishlatishga qaror qilganlar.
02:42
This program uses an iterative approach to gather feedback
46
162746
3167
Bu dastur takroriy usulni qaysi dori va davolash
02:45
about which medications, dosages and treatments are most effective.
47
165913
4583
eng optimali ekanligini aniqlashda feedbek to’playdi.
02:50
Then, it compares that data against each patient’s profile
48
170496
3000
Keyin, har bir bemor bilan optimal bo’lgan
02:53
to create their unique, optimal treatment plan.
49
173496
2833
davolash rejasini tuzish uchun solishtirib chiqadi.
02:56
As the treatments progress and the program receives more feedback,
50
176329
3500
Davolash davom etar va dastur feedback olar ekan,
02:59
it can constantly update the plan for each patient.
51
179829
3292
u doimiy tarzda har bir bemor rejasini yangilab boradi.
03:03
None of these three techniques are inherently smarter than any other.
52
183121
3375
Hech qaysi uch usul bir-biridan kelib chiqishi jihatidan aqlliroq emas.
03:06
While some require more or less human intervention,
53
186496
2667
Ba’zilari ko’proq yo kamroq inson aralashuvini qilsa-da,
03:09
they all have their own strengths and weaknesses
54
189163
2416
barchasini ma’lum ishga mos qiladigan
03:11
which makes them best suited for certain tasks.
55
191579
2250
kuchli va kuchsiz jihatlari bor.
03:14
However, by using them together,
56
194329
2375
Lekin, ularni birga ishlatib, tadqiqotchilar
03:16
researchers can build complex AI systems,
57
196704
2875
mukammal ma’lum dasturlar, bir-birini nazorat qiladigan va
03:19
where individual programs can supervise and teach each other.
58
199579
3292
o’rgatadigan SI sistemalar yarata oladilar.
03:22
For example, when our unsupervised learning program
59
202871
2958
Misol uchun, boshqaruvsiz o’rganadigan dastur
03:25
finds groups of patients that are similar,
60
205829
2334
o’xshash bemorlarni guruhini topsa,
03:28
it could send that data to a connected supervised learning program.
61
208163
3375
buni ulangan boshqaruvli o’rganish dasturiga yuboradi.
03:31
That program could then incorporate this information into its predictions.
62
211829
3500
U dastur keyin bu ma’lumotni bashoratga aylantira oladi
03:35
Or perhaps dozens of reinforcement learning programs
63
215871
2792
Yoki ehtimol ko’plab majburiy o’rganishning ma’lumotlari
03:38
might simulate potential patient outcomes
64
218663
2291
turli davolash rejalari haqidagi mulohazalarni
03:40
to collect feedback about different treatment plans.
65
220954
2750
ehtimoliy bemorning oqibalaridan to’play oladilar.
03:43
There are numerous ways to create these machine-learning systems,
66
223704
3125
O’zi o’rganadigan mashinalar yaratishning bir qancha yo’llari bor,
03:46
and perhaps the most promising models
67
226829
1834
va ehtimol eng optimal model
03:48
are those that mimic the relationship between neurons in the brain.
68
228663
3416
miyadagi neyronlar orasidagi bog’lanishni o’xshatishdir.
03:52
These artificial neural networks can use millions of connections
69
232079
3292
Bu sun’iy neyron bog’lamalari millionlab bog’lanishlarni
03:55
to tackle difficult tasks like image recognition, speech recognition,
70
235371
4417
qiyin muammolarni yechishda ishlata oladilar: surat, ovoz aniqlash
03:59
and even language translation.
71
239788
2041
va hattoki til tarjima qilish.
04:01
However, the more self-directed these models become,
72
241829
3292
Biroq, qanchalik o’zia yo’nalgan model bo’lsa,
04:05
the harder it is for computer scientists
73
245121
2125
dasturchilar uchun shunchalik bu
04:07
to determine how these self-taught algorithms arrive at their solution.
74
247246
3833
o’zi o’rgatilgan algoritmlar ishlashini aniqlash qiyin bo’ladi.
04:11
Researchers are already looking at ways to make machine learning more transparent.
75
251079
4459
Tadqiqotchilar o’zi o’rganadigan mashina- lar yaratishni yo’llarini ko’rmoqdalar.
04:15
But as AI becomes more involved in our everyday lives,
76
255538
2916
Biroq SI ko’proq hayotimizga aralashar ekan,
04:18
these enigmatic decisions have increasingly large impacts
77
258454
2792
bunday o’ylantiradigan qarorlarning ko’proq
04:21
on our work, health, and safety.
78
261246
2875
ish, sog’liq va xavfsizligimizga ta’siri bor.
04:24
So as machines continue learning to investigate, negotiate and communicate,
79
264121
4958
Shunday qilib mashinalar qidirishni, kelishish va gaplashishni o’rganar ekan,
04:29
we must also consider how to teach them to teach each other to operate ethically.
80
269079
5209
biz ularning bir-biriga to’g’ri o’rgatishini ham hisobga olishimiz kerak.
Ushbu veb-sayt haqida

Ushbu sayt sizni ingliz tilini o'rganish uchun foydali bo'lgan YouTube videolari bilan tanishtiradi. Dunyo bo'ylab eng yaxshi o'qituvchilar tomonidan o'qitiladigan ingliz tili darslarini ko'rasiz. Videoni u yerdan o'ynash uchun har bir video sahifasida ko'rsatilgan inglizcha subtitrlarga ikki marta bosing. Subtitrlar video ijrosi bilan sinxronlashtiriladi. Agar sizda biron bir fikr yoki so'rov bo'lsa, iltimos, ushbu aloqa formasi orqali biz bilan bog'laning.

https://forms.gle/WvT1wiN1qDtmnspy7