Ei Habr! Ou te faailoa atu ia te outou le faaliliuga o le tusiga saunia e kristencarter7519.
E ui lava e foliga mai ua na o ni nai aso tatou te o ese mai le 2020, o nei aso e taua foi i le tulaga o le atinaʻeina o polokalama. I lenei tusiga, o le a tatou vaʻavaʻai pe faʻafefea ona suia le olaga o tagata atinaʻe polokalame i le tausaga a sau 2020.
O le lumanaʻi o le atinaʻeina o polokalama o loʻo iinei!
O le atina'eina o polokalame fa'ale-aganu'u o le atina'eina lea o polokalama fa'akomepiuta e ala i le tusiaina o fa'ailoga e mulimuli i nisi o tulafono fa'amautu. Ae o le atinaʻeina o polokalame faʻaonaponei ua molimauina se suiga faʻatusa ma le alualu i luma i le atamai faʻapitoa, masini aʻoaʻoga ma aʻoaʻoga loloto. E ala i le tuʻufaʻatasia o nei tekinolosi e tolu, o le a mafai e le au atinaʻe ona fatuina ni fofo faʻapipiʻi e aʻoaʻo mai faʻatonuga ma faʻaopoopo faʻaopoopoga foliga ma mamanu i faʻamaumauga e manaʻomia e maua ai le manaʻoga.
Tatou taumafai i se code
I le aluga o taimi, neural network software development systems ua sili atu ona lavelave i tulaga o le tuʻufaʻatasia faʻapea foʻi ma tulaga o galuega ma fesoʻotaʻiga. O tagata atiaʻe, mo se faʻataʻitaʻiga, e mafai ona fausia se fesoʻotaʻiga neural faigofie ma le Python 3.6. Ole polokalame fa'ata'ita'iga lea e fa'avasegaina fa'alua ile 1s po'o le 0s.
Ioe, e mafai ona tatou amata i le fatuina o se vasega neural network:
faaulufale mai NumPy pei NP
X=np.array([[0,1,1,0],[0,1,1,1],[1,0,0,1]])
y=np.array([[0],[1],[1]])Faʻaaogaina o le galuega sigmoid:
def sigmoid ():
return 1/(1 + np.exp(-x))
def derivatives_sigmoid ():
return x * (1-x)A'oa'oina se fa'ata'ita'iga ma ulua'i fua ma fa'aituau:
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))Mo tagata amata, afai e te manaʻomia se fesoasoani e uiga i fesoʻotaʻiga neural, e mafai ona e suʻeina le initaneti mo upega tafaʻilagi a kamupani atinaʻe sili ona lelei pe mafai foi ona e faʻafaigaluegaina AI / ML developers e galue i lau poloketi.
Suiga fa'ailoga e fa'aaoga ai se neu fa'apipi'i
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)
Sese fa'atatau mo le fa'ailoga natia
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) *lrTuʻua
print (output)
[[0.03391414]
[0.97065091]
[0.9895072 ]]E aoga i taimi uma le faʻaauau pea i le taimi nei i gagana faʻapolokalame fou ma faiga faʻasologa, ma e tatau foi i le au fai polokalame ona nofouta i le tele o meafaigaluega fou e fesoasoani e faʻaoga ai a latou polokalama i tagata fou.
I le 2020, e tatau i le au atinaʻe polokalame ona mafaufau e tuʻufaʻatasia nei meafaigaluega e 5 mo le atinaʻeina o polokalame i totonu oa latou oloa, e tusa lava po o le a le gagana polokalame latou te faʻaogaina:
1. Gagana Fa'anatura (NLP)
Faatasi ai ma se chatbot e faʻafaigofie ai le auaunaga a tagata faʻatau, o loʻo maua e le NLP le mafaufau o tagata fai polokalame o loʻo galulue i le atinaʻeina o polokalama faʻaonapo nei. Latou te faʻaogaina meafaigaluega NLTK e pei ole Python NLTK e faʻapipiʻi vave ai le NLP i chatbots, fesoasoani numera, ma oloa numera. E oʻo atu i le ogatotonu o le 2020 poʻo le lumanaʻi lata mai, o le a e vaʻai i le NLP e sili atu le taua i mea uma mai pisinisi faʻatau atu i taavale tutoʻatasi ma masini mo le fale ma le ofisa.
I le agai i luma ma sili atu meafaigaluega atinaʻe polokalame ma tekinolosi, e mafai ona e faʻamoemoe e faʻaogaina e le au atinaʻe polokalama le NLP i ni auala eseese, mai le faʻaogaina o tagata faʻaoga leo i le sili atu ona faigofie le faʻaogaina o lisi, suʻesuʻega lagona, faʻamatalaga faʻamatalaga, lagona, ma faʻamatalaga faʻamatalaga. O nei mea uma o le a avanoa mo le toʻatele o tagata faʻaoga, ma o le a mafai e kamupani ona ausia le faʻatupulaia o le gaosiga e oʻo atu i le $ 430 piliona i le 2020 (e tusa ai ma le IDC, na taʻua e Deloitte).
2. GraphQL sui REST Apis
E tusa ai ma le au atinaʻe i laʻu kamupani, o se kamupani atinaʻe polokalame i fafo, o le REST API o loʻo faʻaumatia lona pule i luga o le faʻaogaina o le atulaulau ona o le faʻagesegese o le utaina o faʻamaumauga e manaʻomia ona faia mai le tele o URL taʻitasi.
O le GraphQL ose faiga fou ma ose suiga sili atu ile REST-based architecture e toe maua uma faʻamatalaga talafeagai mai le tele o nofoaga e faʻaaoga ai se fesili e tasi. Ole mea lea e fa'aleleia atili ai feso'ota'iga a le tagata fa'atau ma fa'aitiitia le fa'agaoioiga, ma fa'asili ai ona tali atu le talosaga mo le tagata fa'aoga.
E mafai ona e fa'aleleia atili au tomai tau atina'e pe a e fa'aogaina le GraphQL mo le atina'eina o polokalame. E le gata i lea, e manaʻomia le itiiti ifo o le code nai lo le REST Api ma faʻatagaina oe e fai ni fesili lavelave i nai laina faigofie. E mafai foʻi ona faʻapipiʻiina i le tele o Backand as a Service (BaaS) faʻapitoa e faʻafaigofie ona faʻaogaina e tagata faʻapipiʻi polokalame i gagana eseese polokalame, e aofia ai le Python, Node.js, C++ ma Java.
3. La'asaga maualalo/leai se fa'ailoga (fa'ailoga maualalo)
O mea faigaluega tau atina'e polokalame maualalo e maua ai le tele o fa'amanuiaga. E tatau ona lelei tele pe a tusia le tele o polokalame mai le amataga. Low code e maua ai le code preconfigured e mafai ona fa'apipi'i i polokalame tetele. O lenei mea e mafai ai e oʻo lava i tagata e le o ni polokalame e vave ma faigofie ona faia ni mea lavelave ma faʻavavevaveina le faʻaogaina o meaola faanatura faʻaonaponei.
E tusa ai ma se lipoti a TechRepublic, o meafaigaluega e leai se code / low code ua uma ona faʻaaogaina i luga o upega tafaʻilagi, polokalama faakomepiuta, telefoni feaveaʻi ma isi vaega. Ole maketi o mea faigaluega maualalo ole a tupu ile $15 piliona ile 2020. O nei mea faigaluega e faʻatautaia mea uma, e aofia ai le puleaina o le faʻaogaina o galuega, faʻamama faʻamaumauga, faʻaulufale mai ma faʻatau atu. O pito sili ona maualalo code platforms i le 2020:
- Microsoft PowerApps
- Manatu
- Outsystems
- Zoho Foafoa
- Salesforce App Cloud
- Fa'avae vave
- Tautotogo tautotogo
4. 5G galu
Ole feso'ota'iga 5G ole a a'afia tele ai le telefoni feavea'i ma le atina'eina o polokalame fa'apea fo'i le atina'eina o upegatafa'ilagi. A uma, faʻatasi ai ma tekinolosi e pei o IoT, e fesoʻotaʻi mea uma. O le mea lea, o le masini komepiuta o le a faʻaogaina ai le tele o gafatia o fesoʻotaʻiga uaea televave ma le 5G.
I se faatalanoaga talu ai nei ma Digital Trends, Dan Dery, le sui peresitene o oloa a Motorola, fai mai "i tausaga a sau, o le 5G o le a tuʻuina atu faʻamatalaga vave, maualuga le bandwidth, ma faʻavaveina le telefoni feaveaʻi 10 taimi sili atu nai lo tekinolosi uaea."
I lenei malamalama, o kamupani faakomepiuta o le a galulue e aumai le 5G i faʻaoga faʻaonapo nei. I le taimi nei, e sili atu i le 20 tagata faʻatautaia ua faʻasalalau faʻaleleia a latou fesoʻotaʻiga. O lea la, o le a amata nei e le au atinaʻe ona galulue i le faʻaaogaina o API talafeagai e faʻaoga ai le 5G. O le tekinolosi o le a faʻaleleia atili mea nei:
- Puipuiga o polokalame feso'otaiga, aemaise lava mo le Network Slicing.
- Tuuina atu ni auala fou e taulima ai ID fa'aoga.
- Fa'ataga oe e fa'aopoopo ni galuega fou i tusi talosaga e maualalo le fa'aogaina.
- O le a a'afia ai le atina'eina o le AR/VR system.
5. Fa'amaoni faigofie
O le fa'amaoni ua fa'atuputeleina le avea o se faiga lelei mo le puipuia o fa'amatalaga ma'ale'ale. O tekinolosi fa'apitoa e le gata ina fa'aletonu i masini komepiuta, ae e lagolagoina fo'i le atamai fa'apitoa ma e o'o lava i le fa'aputuga tele. Ae o loʻo vaʻaia e le maketi atinaʻe polokalame le tele o ituaiga fou o faʻamaoniga, e pei o suʻesuʻega leo, biometrics ma le faʻaalia o foliga.
I lenei laʻasaga, e suʻe e tagata taʻavale auala eseese e faʻaogaina ai ID ma upu faʻaoga i luga ole laiga. Talu ai ona ua masani tagata faʻaoga feaveaʻi i le faʻaogaina o latou telefoni feaveaʻi ma se tamatamailima poʻo foliga faʻataʻitaʻiga, ma faʻaaogaina meafaigaluega faʻamaonia, latou te le manaʻomia ni faʻamaoniga fou e mafai ona faʻaitiitia ai le ono mafai ona gaoia le initaneti. O nisi nei o meafaigaluega faʻamaonia faʻapitoa faʻatasi ma faʻailoga SSL.
- Soft Tokens e fa'aliliuina au telefoni i ni fa'amaoniga fa'apitoa e tele.
- EGrid faʻataʻitaʻiga o se faʻaoga faigofie ma taʻutaʻua o faʻamaoniga i totonu o le alamanuia.
- O nisi o polokalame faʻamaonia sili mo pisinisi o le RSA SecurID Access, OAuth, Ping Identity, Authx, ma le Aerobase.
O lo'o iai kamupani fa'akomepiuta i Initia ma Amerika o lo'o faia su'esu'ega tele i le tulaga o le fa'amaoni ma biometrics. O loʻo latou faʻalauiloaina foʻi AI e fatuina polokalama sili atu mo le leo, foliga-id, amio ma le faʻamaoni biometric. Ole taimi nei e mafai ona e puipuia auala numera ma faʻaleleia tulaga gafatia.
iʻuga
E foliga mai o le olaga mo tagata fai polokalame o le a faʻaitiitia le luʻitau i le 2020 ona o le saoasaoa o le atinaʻeina o polokalame e foliga mai e faʻavavevave. O meafaigaluega avanoa o le a faigofie ona faʻaoga. Mulimuli ane, o lenei alualu i luma o le a fausia ai se lalolagi malosi e ulufale atu i se tausaga numera fou.
puna: www.habr.com