Pa Habr! Ndikupereka kwa inu kumasulira kwa nkhaniyi ndi kristencarter7519.
Ngakhale zikuwoneka kuti tangotsala masiku ochepa kuti 2020 achoke, masiku ano ndi ofunikiranso pantchito yopanga mapulogalamu. Pano munkhaniyi, tiwona momwe chaka chomwe chikubwera cha 2020 chidzasinthire moyo wa opanga mapulogalamu.
Tsogolo la chitukuko cha mapulogalamu lili pano!
Kupanga mapulogalamu achikhalidwe ndikukhazikitsa mapulogalamu polemba code kutsatira malamulo okhazikika. Koma chitukuko chamakono cha mapulogalamu chawona kusintha kwa paradigm ndi kupita patsogolo kwa luntha lochita kupanga, kuphunzira makina ndi kuphunzira mozama. Mwa kuphatikiza matekinoloje atatuwa, opanga azitha kupanga mayankho apulogalamu omwe amaphunzira kuchokera ku malangizo ndikuwonjezera zina ndi mawonekedwe kuzinthu zofunikira kuti apange zotsatira zomwe akufuna.
Tiyeni tiyese ndi ma code
M'kupita kwa nthawi, machitidwe opangira mapulogalamu a neural network akhala ovuta kwambiri ponena za kuphatikizika komanso milingo ya magwiridwe antchito ndi mawonekedwe. Madivelopa, mwachitsanzo, amatha kupanga neural network yosavuta kwambiri ndi Python 3.6. Nayi pulogalamu yachitsanzo yomwe imapanga magulu a binary ndi 1s kapena 0s.
Zachidziwikire, titha kuyamba ndikupanga gulu la neural network:
lowetsani NumPy ngati NP
X=np.array([[0,1,1,0],[0,1,1,1],[1,0,0,1]])
y=np.array([[0],[1],[1]])Kugwiritsa ntchito sigmoid ntchito:
def sigmoid ():
return 1/(1 + np.exp(-x))
def derivatives_sigmoid ():
return x * (1-x)Kuphunzitsa chitsanzo chokhala ndi zolemera zoyamba ndi zokondera:
epoch=10000
lr=0.1
inputlayer_neurons = X.shape[1]
hiddenlayer_neurons = 3
output_neurons = 1
wh=np.random.uniform(size=(inputlayer_neurons,hiddenlayer_neurons))
bh=np.random.uniform(size=(1,hiddenlayer_neurons))
wout=np.random.uniform(size=(hiddenlayer_neurons,output_neurons))
bout=np.random.uniform(size=(1,output_neurons))Kwa oyamba kumene, ngati mukufuna thandizo lokhudza ma neural network, mutha kusaka pa intaneti zamakampani opanga mapulogalamu apamwamba kapena mutha kulemba ganyu opanga AI/ML kuti agwire ntchito yanu.
Kusintha ma code pogwiritsa ntchito gawo la neuron
hidden_layer_input1=np.dot(X,wh)
hidden_layer_input=hidden_layer_input1 + bh
hiddenlayer_activations = sigmoid(hidden_layer_input)
output_layer_input1=np.dot(hiddenlayer_activations,wout)
output_layer_input= output_layer_input1+ bout
output = sigmoid(output_layer_input)
Cholakwika chowerengera pama code obisika
E = y-output
slope_output_layer = derivatives_sigmoid(output)
slope_hidden_layer = derivatives_sigmoid(hiddenlayer_activations)
d_output = E * slope_output_layer
Error_at_hidden_layer = d_output.dot(wout.T)
d_hiddenlayer = Error_at_hidden_layer * slope_hidden_layer
wout += hiddenlayer_activations.T.dot(d_output) *lr
bout += np.sum(d_output, axis=0,keepdims=True) *lr
wh += X.T.dot(d_hiddenlayer) *lr
bh += np.sum(d_hiddenlayer, axis=0,keepdims=True) *lrTulukani
print (output)
[[0.03391414]
[0.97065091]
[0.9895072 ]]Ndikoyenera kuti nthawi zonse mukhale ndi zilankhulo zaposachedwa komanso njira zolembera, ndipo opanga mapulogalamu ayeneranso kudziwa zida zambiri zatsopano zomwe zimathandiza kuti mapulogalamu awo akhale ogwirizana ndi ogwiritsa ntchito atsopano.
Mu 2020, opanga mapulogalamu akuyenera kuganizira zophatikiza zida zisanu zopangira mapulogalamuwa muzinthu zawo, posatengera kuti amagwiritsa ntchito chilankhulo chotani:
1. Natural Language Processing (NLP)
Ndi chatbot yomwe imathandizira makasitomala, NLP ikupeza chidwi ndi opanga mapulogalamu omwe akugwira ntchito yopanga mapulogalamu amakono. Amagwiritsa ntchito zida za NLTK monga Python NLTK kuti aphatikizepo NLP muzokambirana, othandizira digito, ndi zinthu zamagetsi. Pofika pakati pa 2020 kapena posachedwa, mudzawona NLP ikukhala yofunika kwambiri pachilichonse kuyambira mabizinesi ogulitsa mpaka magalimoto odziyimira pawokha ndi zida zapanyumba ndi ofesi.
Kupita patsogolo ndi zida zabwinoko zopangira mapulogalamu ndi matekinoloje, mutha kuyembekezera opanga mapulogalamu kuti agwiritse ntchito NLP m'njira zosiyanasiyana, kuchokera pamawu ogwiritsira ntchito mawu kupita kumayendedwe osavuta a menyu, kusanthula kwamaganizidwe, kuzindikiritsa zochitika, kutengeka mtima, ndi kupezeka kwa data. Zonsezi zitha kupezeka kwa ogwiritsa ntchito ambiri, ndipo makampani azitha kukwaniritsa zokolola mpaka $430 biliyoni pofika 2020 (malinga ndi IDC, yotchulidwa ndi Deloitte).
2. GraphQL m'malo mwa REST Apis
Malinga ndi omwe akupanga pakampani yanga, yomwe ndi kampani yopanga mapulogalamu akunyanja, REST API ikutaya mphamvu zake pa chilengedwe chonse chogwiritsa ntchito chifukwa chakutsitsa pang'onopang'ono kwa data yomwe imayenera kuchitidwa kuchokera ku ma URL angapo payekhapayekha.
GraphQL ndi njira yatsopano komanso njira yabwinoko yopangira zomanga za REST zomwe zimapezanso zonse zofunikira kuchokera kumasamba angapo pogwiritsa ntchito funso limodzi. Izi zimathandizira kulumikizana kwa kasitomala ndi seva ndikuchepetsa latency, kupangitsa kuti pulogalamuyo ikhale yomvera kwambiri kwa wogwiritsa ntchito.
Mutha kukulitsa luso lanu lokulitsa mapulogalamu mukamagwiritsa ntchito GraphQL pakupanga mapulogalamu. Kuphatikiza apo, imafunikira kachidindo kakang'ono kuposa REST Api ndipo imakulolani kuti mupange mafunso ovuta m'mizere yosavuta. Itha kukhalanso ndi zida zingapo za Backand as a Service (BaaS) zomwe zimapangitsa kuti zikhale zosavuta kugwiritsa ntchito ndi opanga mapulogalamu m'zilankhulo zosiyanasiyana zamapulogalamu, kuphatikiza Python, Node.js, C++ ndi Java.
3. Mulingo wocheperako / palibe code (kodi yotsika)
Zida zonse zopangira mapulogalamu otsika a code zimapereka zabwino zambiri. Iyenera kukhala yothandiza momwe mungathere polemba mapulogalamu ambiri kuyambira pachiyambi. Khodi yotsika imapereka kachidindo kokonzedweratu komwe kungathe kuikidwa mu mapulogalamu akuluakulu. Izi zimathandiza ngakhale osapanga mapulogalamu kuti azitha kupanga mwachangu komanso mosavuta zinthu zovuta ndikufulumizitsa chitukuko chamakono.
Malinga ndi lipoti la TechRepublic, zida zopanda ma code / low code zikugwiritsidwa ntchito kale pamawebusayiti, mapulogalamu apulogalamu, mafoni am'manja ndi madera ena. Msika wa zida zotsika udzakhala $15 biliyoni pofika 2020. Zida izi zimagwira chilichonse, kuphatikiza kuyang'anira kayendetsedwe ka ntchito, kusefa deta, kulowetsa ndi kutumiza kunja. Nawa nsanja zabwino kwambiri zotsika mu 2020:
- Microsoft PowerApps
- Mendix
- Zakunja
- Zoho Mlengi
- Salesforce App Cloud
- Mwachangu
- Nsapato za Spring
4. 5G chiwombankhanga
Kulumikizana kwa 5G kudzakhudza kwambiri pulogalamu yam'manja ndi chitukuko cha mapulogalamu komanso chitukuko cha intaneti. Kupatula apo, ndi matekinoloje monga IoT, zonse zimalumikizidwa. Chifukwa chake, pulogalamu ya chipangizocho idzapindula kwambiri ndi ma netiweki opanda zingwe othamanga kwambiri ndi 5G.
Poyankhulana ndi Digital Trends, a Dan Dery, wachiwiri kwa purezidenti wa Motorola, adati "m'zaka zikubwerazi, 5G idzapereka deta yofulumira, bandwidth yapamwamba, ndikufulumizitsa mapulogalamu a foni nthawi 10 mofulumira kuposa matekinoloje omwe alipo kale."
Mwa ichi, makampani opanga mapulogalamu agwira ntchito kuti abweretse 5G muzogwiritsa ntchito zamakono. Pakadali pano, opitilira 20 alengeza kukweza maukonde awo. Chifukwa chake, opanga tsopano ayamba kugwira ntchito pogwiritsa ntchito ma API oyenerera kuti apeze mwayi wa 5G. Tekinolojeyi idzawongolera kwambiri zotsatirazi:
- Chitetezo cha pulogalamu yapaintaneti, makamaka pa Network Slicing.
- Perekani njira zatsopano zogwiritsira ntchito ma ID.
- Limakupatsani mwayi wowonjezera magwiridwe antchito omwe ali ndi latency yotsika.
- Zidzakhudza chitukuko cha dongosolo la AR / VR.
5. Easy kutsimikizika
Kutsimikizira kukuchulukirachulukira kukhala njira yabwino yotetezera deta yodziwika bwino. Ukadaulo wotsogola sikuti umangokhala pachiwopsezo cha ma hacks a mapulogalamu, komanso umathandizira luntha lochita kupanga komanso ngakhale quantum computing. Koma msika wopititsa patsogolo mapulogalamu akuwona kale mitundu yambiri yotsimikizika, monga kusanthula mawu, biometrics ndi kuzindikira nkhope.
Pakadali pano, obera amapeza njira zosiyanasiyana zopangira ma ID ndi mapasiwedi ogwiritsa ntchito pa intaneti. Popeza ogwiritsira ntchito mafoni ayamba kale kulumikiza mafoni awo ndi chala kapena jambulani kumaso, motero pogwiritsa ntchito zida zotsimikizira, sangafunikire kutsimikizira kwatsopano chifukwa mwayi wa kuba pa intaneti udzakhala wochepa. Nawa zida zina zotsimikizira zinthu zambiri zokhala ndi encryption ya SSL.
- Ma Soft Tokens amasintha mafoni anu kukhala otsimikizira azinthu zambiri.
- Ma tempulo a Egrid ndi osavuta kugwiritsa ntchito komanso odziwika bwino pamakampani.
- Ena mwa mapulogalamu abwino otsimikizira mabizinesi ndi RSA SecurID Access, OAuth, Ping Identity, Authx, ndi Aerobase.
Pali makampani opanga mapulogalamu ku India ndi ku US omwe akuchita kafukufuku wambiri pankhani ya kutsimikizika ndi ma biometric. Akulimbikitsanso AI kuti apange mapulogalamu apamwamba a mawu, nkhope-id, khalidwe ndi biometric kutsimikizika. Tsopano mutha kuteteza mayendedwe a digito ndikuwongolera luso la nsanja.
Pomaliza
Zikuwoneka kuti moyo wa opanga mapulogalamu udzakhala wovuta kwambiri mu 2020 chifukwa mayendedwe a mapulogalamu akuyenera kukwera. Zida zomwe zilipo zidzakhala zosavuta kugwiritsa ntchito. Pamapeto pake, kupita patsogolo kumeneku kudzapanga dziko lamphamvu kulowa m'badwo watsopano wa digito.
Source: www.habr.com