Federal gosudarstvennoe uchebnoe predpriyatie Chair of the System of Artificial Intelligence

Control questions for 2 glave:
Ispolzovanie atributnyx grammatik hotya i uslojnyaet isxodnuyu KS
grammatiku, no pozvolyaet detalno rassmotret osobennosti realizatsii i
dlya razrabotki translyatorov, kak istoricheski teoreticheskaya osnova.
Preimushchestvom atributnogo podkhoda yavlyaetsya funktsionalnaya priroda
3. Training and testing
crossroads and traffic lights. Tem ne menee, yavlyaetsya bolee predpochtitelnÿm
Exit
2. Predobrabotka
attributes. S odnoy storony eto vÿzÿvaet nekotorye trudnosti pri
vse semeystvo grammatik, poluchennyx na osnove pervonachalnoy. Kak
postroenii translyatorov. No pozvolyaet sushchestvenno rasshirit grammar
5. Osnovnÿe tipy zadach, reshaemyx s pomoshchyu neyronnyx setey?
6. Generalnaya sovokupnost, obuchayushchaya vyborka, testovaya
vyborka. 7. Effect pereobucheniya setey
semantiku razlichnyx tehnologicheskix operatsiy i sushchestvenno utochnyaet
atributnogo vychislitelya i neuporyadochennost protsessa vychisleniya
neyrosetevuyu technology. Dannyy podkhod mojet byt rasprostranen i na
4. Contrast: vidy, sposoby
10. Sposoby polucheniya yavnogo znaniya ot neyronnoy seti s uchitelem
postroit translator by analogii s yazÿkami programmirovaniya.
neyrosetevoy teknologii.
8. Vybor polya otveta dlya dvoystvennyx neyronnyx setey
potentsialno dlya lyuboy slojnosti predmetnyx oblastey, v tom chisle i dlya
92
Machine Translated by Google

---

postroen dlya ochen bÿstrogo i effektivnogo resheniya zadachi binarnoy
setey dlya obrabotki dannyx polezny dlya ispolzovaniya na nachalnom etape
imitators MDN, Neurogenesis (A.V. Xomich) and ryade drugix podobnyx
pravilo nedostatochno uchebnoy ili metodicheskoy literatury, ne govorya uje
pomoshchyu programs Monoton i mnogix drugix. Krome togo, in nastoyashchee
i neobxodimyy etap rabotÿ, no on ne yavlyaetsya dostatochnÿm. Ochen vajno
izvestnyx zarubejnyx paketyov bÿl vÿbran pakety SNN. Vybor SNN obosnovÿvaetsya eshche
i tem, chto v bolshinstve imitatorov ili paketov
metodov i algorithms obucheniya neyronnyx setey, no menee udoben dlya
architecture. SNN offers a choice of several different architectures.
neyronnyx setey, stepen realizatsii predlojennoy teknologii v takix
predusmotreny dopolnitelnÿe modules for working with neural networks. For example,
products of the Excel Neural Package series (Winnet, Kohonen map), realized as a set of
add-ons for Microsoft Excel. When choosing a program for dannoy rabotÿ otsenivalsya ryad
objektivnyx
Sredi bolshogo kolichestva gotovyx paketov programm mojno
russkom yazyke ili hotya by nalichie tekstov ili uchebnyx posobiy na russkom
Very good programs MultiNeuron (different versions), EYE, CLAB developed for work
in the operating system MS DOS, sometimes in
kotoryye vklyuchenÿ vozmojnosti ispolzovaniya neyronnyx setey, kak
Described high technologies and methods of using neurons
especially Siberia: Clab (S.E. Gilev), MultiNeuron raznyx versiy (A.A. Rossiev, E. M. Mirkes,
S. E. Gilev et al.), NeuroPro (V. G. Tsaregorodtsev), Eye, FAMaster (D.A. Rossiev), (Zinovev
A.), STATISTICA Neural Networks, MATLAB and many others. Est bolshoe kolichestvo
menee izvestnyx
sovremennymi EVM i operating systems. Chrome togo, CLAB byl
raboty - obdumyvaniya i postanovki zadachi. The task is the first
programm /
Obuchenie neyronnyx setey s uchitelem mojet byt vÿpolneno s
o tom, chto interface bolshinstva iz nix na angliyskom yazyke. Sredi bolee
classification. EYE pozvolyaet reshat raznÿe zadachi dlya izucheniya raznyx
otechestvennogo proizvodstva ispolzuyut neyronnÿe seti odnoy
prakticheskoe ispolzovanie konkretnyx instrumentov - imitatorov
time in some package statistics and in electronic tables
osvoeniya osnov teknologii v sovremennom ponimanii. MultiNeuron bolee
imitators, quality and convenience of work with programs, etc.
and subjective properties of imitators. For example, trebovalsya interface na
_
otmetit nekotorÿe, naibolee izvestnÿe spetsialistam nashey strany i