Hey Habr! Ke fana ka tlhokomelo ea hau phetolelo ea sehlooho ka kstencarter7519.
Leha ho bonahala eka re saletsoe ke matsatsi a seng makae feela ho tloha 2020, matsatsi ana le 'ona a bohlokoa tšimong ea nts'etsopele ea software. Mona sengolong sena, re tla bona kamoo selemo se tlang sa 2020 se tla fetola bophelo ba baetsi ba software.
Bokamoso ba nts'etsopele ea software bo fihlile!
Nts'etsopele ea software ea setso ke nts'etsopele ea software ka ho ngola khoutu ho latela melao e itseng e tsitsitseng. Empa nts'etsopele ea software ea sejoale-joale e bone phetoho ea paradigm ka tsoelo-pele ea bohlale ba maiketsetso, ho ithuta ka mochini le thuto e tebileng. Ka ho kopanya mekhoa ena e meraro ea theknoloji, bahlahisi ba tla khona ho etsa litharollo tsa software tse ithutang ho latela litaelo le ho eketsa likarolo le mekhoa e meng ho data e hlokahalang ho hlahisa sephetho se lakatsehang.
Ha re leke ka khoutu e itseng
Ha nako e ntse e ea, mekhoa ea ntlafatso ea software ea neural network e fetohile e rarahaneng haholoanyane mabapi le ho kopanya hammoho le maemo a ts'ebetso le li-interfaces. Ka mohlala, bahlahisi ba ka haha marang-rang a bonolo haholo a neural ka Python 3.6. Mona ke mohlala oa lenaneo le etsang binary classification le 1s kapa 0s.
Ehlile, re ka qala ka ho theha sehlopha sa neural network:
kenya NumPy joalo ka NP
X=np.array([[0,1,1,0],[0,1,1,1],[1,0,0,1]])
y=np.array([[0],[1],[1]])Tšebeliso ea mosebetsi oa sigmoid:
def sigmoid ():
return 1/(1 + np.exp(-x))
def derivatives_sigmoid ():
return x * (1-x)Koetlisa mohlala o nang le boima ba pele le leeme:
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))Bakeng sa ba qalang, haeba u hloka thuso mabapi le marang-rang a marang-rang, u ka batla marang-rang bakeng sa marang-rang a lik'hamphani tse holimo tsa nts'etsopele ea software kapa u ka hira baetsi ba AI/ML ho sebetsa morerong oa hau.
Phetoho ea khoutu ho sebelisoa neuron ea lesela la tlhahiso
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)
Phoso ea lipalo bakeng sa lera la khoutu e patiloeng
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) *lrTloha
print (output)
[[0.03391414]
[0.97065091]
[0.9895072 ]]Kamehla ho bohlokoa ho lula u tseba ka lipuo tsa morao-rao tsa mananeo le mekhoa ea ho khouta, 'me baetsi ba mananeo le bona ba lokela ho tseba lisebelisoa tse ngata tse ncha tse thusang ho etsa hore lits'ebetso tsa bona li sebetse ho basebelisi ba bacha.
Ka 2020, baetsi ba software ba lokela ho nahana ka ho kenyelletsa lisebelisoa tsena tse 5 tsa nts'etsopele ea software lihlahisoa tsa bona, ho sa tsotelehe hore na ba sebelisa puo efe ea lenaneo:
1. Tlhaho ea Puo ea Tlhaho (NLP)
Ka chatbot e tsamaisang litšebeletso tsa bareki, NLP e hapa tlhokomelo ea baetsi ba mananeo ba sebetsang ho nts'etsopele ea software ea sejoale-joale. Ba sebelisa lisebelisoa tsa NLTK tse kang Python NLTK ho kenyelletsa NLP ka potlako ho li-chatbots, bathusi ba dijithale le lihlahisoa tsa dijithale. Bohareng ba 2020 kapa haufinyane, u tla bona NLP e ba bohlokoa ho feta nthong e 'ngoe le e' ngoe ho tloha likhoebong tsa mabenkele ho isa likoloing tse ikemetseng le lisebelisoa tsa lapeng le ofisi.
Ha u hatela pele ka lisebelisoa tse ntle tsa nts'etsopele ea software le mahlale a morao-rao, u ka lebella hore baetsi ba software ba sebelise NLP ka mekhoa e fapaneng, ho tloha ho li-interfaces tse thehiloeng ho lentsoe ho isa ho menu ea ho tsamaea habonolo, tlhahlobo ea maikutlo, ho tsebahatsa maemo, maikutlo, le phihlello ea data. Sena sohle se tla fumaneha ho basebelisi ba bangata, 'me lik'hamphani li tla khona ho fihlela kholo ea tlhahiso ho fihlela ho $430 limilione tse likete ka 2020 (ho ea ka IDC, e boletsoeng ke Deloitte).
2. GraphQL e nkela sebaka sa REST Apis
Ho latela baetsi ba feme ea ka, e leng k'hamphani ea nts'etsopele ea software e kantle ho naha, REST API e lahleheloa ke taolo holim'a bokahohle ba ts'ebeliso ka lebaka la ho jarolla butle ha data e lokelang ho etsoa ho tsoa ho li-URL tse ngata ka bonngoe.
GraphQL ke mokhoa o mocha le o mong o betere ho feta meaho e thehiloeng ho REST e fumanang lintlha tsohle tse amehang ho tsoa libakeng tse ngata ho sebelisa potso e le 'ngoe. Sena se ntlafatsa tšebelisano ea bareki le seva mme se fokotsa latency, se etsa hore sesebelisoa se arabele haholoanyane bakeng sa mosebelisi.
U ka ntlafatsa tsebo ea hau ea nts'etsopele ea software ha u sebelisa GraphQL bakeng sa nts'etsopele ea software. Ho feta moo, e hloka khoutu e nyane ho feta REST Api mme e o lumella ho etsa lipotso tse rarahaneng ka mela e 'maloa e bonolo. E ka boela ea hlomelloa ka likarolo tse ngata tsa Backand as a Service (BaaS) tse etsang hore ho be bonolo ho sebelisoa ke baetsi ba lisebelisoa ka lipuo tse fapaneng tsa lenaneo, ho akarelletsa le Python, Node.js, C++ le Java.
3. Boemo bo tlase ba likhoutu/ho se na khoutu (khoutu e tlase)
Lisebelisoa tsohle tse tlase tsa ntlafatso ea software li fana ka melemo e mengata. E lokela ho sebetsa hantle ka hohle kamoo ho ka khonehang ha o ngola mananeo a mangata ho tloha qalong. Khoutu e tlase e fana ka khoutu e lokiselitsoeng esale pele e ka kengoang ho mananeo a maholo. Sena se lumella le bao e seng baetsi ba mananeo ho etsa lihlahisoa tse rarahaneng ka potlako le habonolo le ho potlakisa tikoloho ea sejoale-joale ea nts'etsopele.
Ho ea ka tlaleho ea TechRepublic, lisebelisoa tsa no-code / low code li se li ntse li sebelisoa ho li-portal tsa marang-rang, lisebelisoa tsa software, lisebelisoa tsa mohala le libaka tse ling. Mmaraka oa lisebelisoa tsa khoutu e tlase o tla hola ho $15 bilione ka 2020. Lisebelisoa tsena li sebetsana le ntho e 'ngoe le e' ngoe, ho kenyelletsa le ho laola logic ea mosebetsi, ho sefa data, ho kenya le ho romela thepa. Mona ke li-platform tsa khoutu tse tlase tse tlase ka 2020:
- Microsoft PowerApps
- Mendix
- Mekhoa ea kantle
- 'Mopi oa Zoho
- Salesforce App Cloud
- Base e potlakileng
- Boot ea selemo
4. Leqhubu la 5G
Khokahano ea 5G e tla ama haholo app ea mobile le nts'etsopele ea software hammoho le nts'etsopele ea webo. Ntle le moo, ka mahlale a joalo ka IoT, tsohle li hokahane. Ka hona, sesebelisoa sa sesebelisoa se tla sebelisa bokhoni ba marang-rang a marang-rang a marang-rang ka 5G.
Puisanong ea morao-rao le Digital Trends, Dan Dery, motlatsi oa mopresidente oa sehlahisoa sa Motorola, o itse "lilemong tse tlang, 5G e tla fana ka data e potlakileng, bandwidth e phahameng, le ho potlakisa software ea mohala ka makhetlo a 10 ho feta mahlale a teng a se nang mohala."
Ka leseli lena, lik'hamphani tsa software li tla sebetsa ho tlisa 5G lits'ebetsong tsa sejoale-joale. Hajoale, basebelisi ba fetang 20 ba phatlalalitse ntlafatso ea marang-rang a bona. Kahoo, bahlahisi ba tla qala ho sebetsa ho sebelisa li-API tse nepahetseng ho nka monyetla oa 5G. Theknoloji e tla ntlafatsa haholo tse latelang:
- Ts'ireletso ea lenaneo la marang-rang, haholo-holo bakeng sa Network Slicing.
- Fana ka mekhoa e mecha ea ho sebetsana le li-ID tsa basebelisi.
- E o lumella ho kenya ts'ebetso e ncha lits'ebetsong tse nang le latency e tlase.
- E tla susumetsa nts'etsopele ea sistimi ea AR / VR.
5. Bopaki bo bonolo
Netefatso e ntse e tsoela pele ho ba mokhoa o sebetsang oa ho sireletsa lintlha tsa bohlokoa. Theknoloji e tsoetseng pele ha e kotsing feela ea li-hacks tsa software, empa hape e tšehetsa bohlale ba maiketsetso esita le quantum computing. Empa 'maraka oa nts'etsopele ea software o se o ntse o bona mefuta e mengata e mecha ea netefatso, joalo ka tlhahlobo ea lentsoe, biometrics le ho lemoha sefahleho.
Mothating ona, linokoane li fumana mekhoa e fapaneng ea ho theha li-ID tsa basebelisi ba inthaneteng le li-password. Kaha basebelisi ba mehala ba se ba tloaetse ho fumana li-smartphone tsa bona ka menoana kapa sefahleho sa sefahleho, kahoo ba sebelisa lisebelisoa tsa ho netefatsa, ba ke ke ba hloka matla a macha a ho netefatsa kaha monyetla oa bosholu ba cyber o tla fokotseha. Mona ke lisebelisoa tse ling tsa netefatso ea lintlha tse ngata tse nang le encryption ea SSL.
- Li-Soft Tokens li fetola li-smartphones tsa hau hore e be li-authenticator tse bonolo tsa lintho tse ngata.
- Li-template tsa Egrid ke mofuta o bonolo oa ho sebelisa le o tsebahalang oa li-authenticators indastering.
- A mang a manane a matle a netefatso bakeng sa likhoebo ke RSA SecurID Access, OAuth, Ping Identity, Authx, le Aerobase.
Ho na le lik'hamphani tsa software India le US tse etsang lipatlisiso tse pharaletseng lefapheng la netefatso le biometrics. Ba boetse ba khothaletsa AI ho theha software e holimo bakeng sa lentsoe, sefahleho-id, boitšoaro le netefatso ea biometric. Hona joale o ka sireletsa liteishene tsa dijithale le ho ntlafatsa bokhoni ba sethala.
fihlela qeto e
Ho bonahala eka bophelo ba baetsi ba mananeo bo tla fokotseha ka 2020 kaha lebelo la nts'etsopele ea software le kanna la potlaka. Lisebelisoa tse teng li tla ba bonolo ho li sebelisa. Qetellong, tsoelopele ena e tla theha lefatše le matla le kenang mehleng e mecha ea dijithale.
Source: www.habr.com