Federal gosudarstvennoe uchebnoe predpriyatie Chair of the System of Artificial Intelligence



Download 7,58 Mb.
Pdf ko'rish
bet21/37
Sana14.06.2022
Hajmi7,58 Mb.
#667933
1   ...   17   18   19   20   21   22   23   24   ...   37
Bog'liq
zhukov la reshetnikova nv uchebnoe posobie po distsipline pr

kajdom iz imeyushchixsya putey proxojdeniya signalov v seti ot vxoda k
dostatochno malo (2 - 3, no mojet zadavatsya i polzovatelem).
vyxodu, tem
luchshe. 2. Chem menshe chislo neyronov v kajdom sloe seti, tem luchshe. Etot
neodnorodnogo vxoda neyrona. Deystvitelno, neodnorodnÿy vxod neyrona -
ÿÿÿ prosto konstanta, v otlichie ot sinapsa, vÿpolnyayushchego umnojenie
68
Machine Translated by Google


degree of logic prozrachnosti:
chast sinapticheskix vesov prinimaet znacheniya iz mnojestva {-1, 0, 1}.
Sushchestvuyut raznÿe vidy binarizatsii, kotorye predusmatrivayut
vida easily represents in dvoichnom code, chetvertyy je vid (chrome otdelnyx
Iznachalno binarization sozdavalas for uproshcheniya realizatsii
answer).
8. Postroenie mnogestva minimalnyx bezizbytochnyx naborov
iz kriteriev so slaboy stepenyu sootvetstviya. 6.
Binarization neuronnoy set - takoe obuchenie set, chto bolshaya
sets (contrast does not occur).
neuroimitator NeuroPro v. 0.25 realizovanÿ sleduyushchie vidy
deystvitelnyx znacheniy vesov sinapsov k znacheniyam iz mnojestva, kotorye
parameters or reshenie odnoy iz obratnyx zadach (t. e. vÿbor drugogo polya
razlichnÿe konechnÿe mnogestva znacheniy vesov sinapsov. For example, v
7. Neobxodimo privedenie znacheniy nastraivaemyx parametrov seti
neyronnoy seti on tsifrovyx EVM, chto podrazumevaet privedenie
vxodnyx parameters (for removal of duplication of the 1st generation). Minimal
bezizbytochnÿy set vxodnyx parameters - set, pozvolyayushchiy obuchit neyronnuyu
set for polucheniya answer. Prichem
imeetsya v vidu obuchenie neskolkix neyronnyx setey dlya vÿbrannogo polya
-0.7, -0.6, -0.5, -0.4, -0.3, -0.2, -0.1, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1}.
It should be noted that in NeuroPro binary does not provide priority
extremely rare, almost never. Poetomu imeet smysl razlichat vidy ili
obrabatÿvat binarizovannÿe vesa synapsov pri raschetax. Esli
parameter seti - privedenie vesov sinapsov k znacheniyam -1 i 1.
Pri reshenii prikladnyx zadach logicheski prozrachnÿe seti poluchayut
binarizatsii:
privedenie vesov sinapsov k znacheniyam iz mnojestva {-1, 1};
privedenie vesov sinapsov k znacheniyam iz mnojestva {-1, -0.5, 0.5, 1};
privedenie vesov sinapsov k znacheniyam iz mnojestva {-1, -0.75, -0.5,
-0.25, 0.25, 0.5, 0.75, 1}; privedenie vesov sinapsov k znacheniyam iz mnojestva
{-1, -0.9, -0.8,
udobno predstavit in dvoichnom code. This pozvolyaet EVM bÿstree i tochnee
znacheniy) teoreticheski ne predstavim v dvoichnoy sisteme schisleniya, poetomu
okruglyaetsya do nekotoroy tochnosti. 7. Obuchenie po isxodnoy vyborke na
predskazanie odnogo iz vxodnyx
1) strict - with all the necessary requirements (criteria 1 - 7); 2)
slabo logicheski prozrachnÿe seti - soblyudayutsya tolko nekotorÿe
synapse. Binarization in dannom sluchae not uproshchaet vnutrennyuyu structure
rassmatrivat vidy binarizatsii, realizovannyx v NeuroPro, to pervye tri
vesa sinapsa k nulevomu znacheniyu, chto fakticheski oznachalo by udalenie
k konechnomu naboru vÿdelennyx znacheniy. Gelatelna binary
69
Machine Translated by Google


vnutrenney struktury (udaleniya synapsov, neyronov, neodnorodnyx vxodov), a idet
tolko sokrashchenie chisla vxodnyx parametrov (v NeuroPro etot
opredelyaetsya minimalnÿy bezizbytochnÿy set of input parameters, analogichno
predydushchemu point.
smÿsla termina s blizkimi emu podubiyami (vÿyasnenie - novoe li eto znanie
minimalnizo bezizbÿtochnogo nabora ispolzuetsya operation
dublirovaniya 2-go roda. Krome togo, methods of extraction and removal
smÿslov imeet spetsificheskiy character i dlya kajdoy zadachi vÿpolnenie
yavlyaetsya samoy slojnoy i, in nekotorom smÿsle, intellektualnoy. Search
9. For each input parameter stroim nabor parameters, kotoryy
nujnogo vÿchislennogo polya otveta.
etogo vÿpolnyaetsya obuchenie neyronnyx setey s vÿborom v kachestve polya
sochetaniy («sushchnostey»), voznikayushchix v sxodnyx usloviyax. Imenovanie
oblasti.
Ispolzuya opisannÿy tehnologicheskiy poryadok operatsiy, mojno
in a unique term. Proverka, chto est eta «sushchnost» - obshchee, osobennoe ili
edinichnoe (chastnoe) (proverka - znanie li eto?). Comparison
neyronnaya set posledovatelno uproshchaetsya, no ne v smÿsle uproshcheniya
samogo pervogo vxodnogo parameter (polya) i zakanchivaya poslednim. Then
results.
process avtomaticheski vklyuchaet doobuchenie seti). For postroeniya
Takim obrazom, dublirovanie 1-go roda est chastnyy sluchay
or ono uje sushchestvuet, t. e. izvestno ranee). K sojaleniyu, operation search
this point neskolko otlichaetsya from drugix. Sootvetstvenno, this operation
contrast.
dublirovaniya 2-go roda est razvitie p. 7. 10.
Obuchenie primera s ispolzovaniem obuchennoy neyronnoy seti - vÿyasnenie
minimalnyx izmeneniy v parametrax primera dlya polucheniya
ego odnoznachno opredelyaet (dlya ustraneniya dublirovaniya 2-go roda). For
11. Search terms. Search v rezultatax ustoychivyx, povtoryayushchixsya
smÿslov obÿchno provoditsya sovmestno so spetsialistami v predmetnoy
answer kajdogo from vxodnyx parameters, by ocheredi. Takim obrazom, provodyatsya
ÿksperimentÿ s posledovatelnoy smenoy polya otveta, nachinaya s
otveta (pole otveta - odno i to je dlya vsex neyrosetey). After chego kajdaya
ustoychivyx, povtoryayushchixsya sochetaniy v dannyx («sushchnostey») nekotorÿm
poluchit prakticheskie rezultaty dlya lyubyx dannyx prikladnoy oblasti. Primenyaya
dopolnitelnÿe instrumentsy obrabotki dannyx, naprimer, apparat matematicheskoy
statistiki, proveryayut adekvatnost poluchennyx
70
Machine Translated by Google


neyronnyx setey dlya resheniya prikladnyx zadach na urovne polzovatelya, t. e. bez detalizatsii
ispolzuemyx operatsiy. Operatsii predpolagayutsya uje
0) group operatsiy - predvaritelnÿe, osnovnÿe i vspomogatelnÿe
sravneniyu s osnovnymi. V xode obrabotki dannyx obyazatelnÿm yavlyaetsya
operatsiy budem rassmatrivat tri gruppy operatsiy, skomponovannÿe v
in spetsializirovannyx programmax - predobrabotchikax or others
teknologii neyrosetevoy obrabotki dannyx (numeration nachinaetsya s 0):
In dalneyshem dlya postroeniya vsex privedennyx tehnologicheskix

Download 7,58 Mb.

Do'stlaringiz bilan baham:
1   ...   17   18   19   20   21   22   23   24   ...   37




Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©hozir.org 2025
ma'muriyatiga murojaat qiling

kiriting | ro'yxatdan o'tish
    Bosh sahifa
юртда тантана
Боғда битган
Бугун юртда
Эшитганлар жилманглар
Эшитмадим деманглар
битган бодомлар
Yangiariq tumani
qitish marakazi
Raqamli texnologiyalar
ilishida muhokamadan
tasdiqqa tavsiya
tavsiya etilgan
iqtisodiyot kafedrasi
steiermarkischen landesregierung
asarlaringizni yuboring
o'zingizning asarlaringizni
Iltimos faqat
faqat o'zingizning
steierm rkischen
landesregierung fachabteilung
rkischen landesregierung
hamshira loyihasi
loyihasi mavsum
faolyatining oqibatlari
asosiy adabiyotlar
fakulteti ahborot
ahborot havfsizligi
havfsizligi kafedrasi
fanidan bo’yicha
fakulteti iqtisodiyot
boshqaruv fakulteti
chiqarishda boshqaruv
ishlab chiqarishda
iqtisodiyot fakultet
multiservis tarmoqlari
fanidan asosiy
Uzbek fanidan
mavzulari potok
asosidagi multiservis
'aliyyil a'ziym
billahil 'aliyyil
illaa billahil
quvvata illaa
falah' deganida
Kompyuter savodxonligi
bo’yicha mustaqil
'alal falah'
Hayya 'alal
'alas soloh
Hayya 'alas
mavsum boyicha


yuklab olish