Overview
Yakareba Yenguva Yenguva Yekurangarira (LSTM) maseru imhando yakakosha yeinodzokororwa neural network unit yakavakirwa kurangarira ruzivo mukati mekutevedzana kwakareba. Vakagadzirisa dambudziko rekunyangarika-gradient iro rakaremara maRNN apfuura, vachisimbisa makore gumi ekubudirira mumutauro, kutaura, uye kududzira.
Yakareba Yenguva Ipfupi Memory Masero anogara mumusimboti AI toolkit. Paunonzwisisa, mamwe maAI misoro inova nyore kuongorora uye kuenzanisa.
Deep Dive
Yakaunzwa naSepp Hochreiter naJurgen Schmidhuber muna 1997, sero reLSTM rinochengeta 'cell state' inoita sebhandi rekutakura rendangariro richimhanya nemukutevedzana. Magedhi matatu akadzidza anoitonga: gedhi rekukanganwa rinosarudza chekudzima, gedhi rekupinda rinosarudza ruzivo rutsva rwekuchengeta, uye gedhi rekubuda rinosarudza chekufumura sekubuda kwesero. Gedhi rega rega rinoshandisa sigmoid (inoburitsa 0 kusvika 1) kuita seyakapfava switch. Nekuti iyo sero state inogadziridzwa kazhinji nekuwedzera kwete kudzokororwa kuwanda, ma gradients anogona kuyerera achidzokera kumashure pane akawanda nguva nhanho pasina kuderera kusvika zero, zvichiita kuti LSTMs idzidze kutsamira mazana ematanho akaparadzana. Pamberi peTransformers, maLSTM anotsigirwa Google Shandura, kuziva kutaura, uye kugadzira zvinyorwa.
Technical Insight
Iyo inonyangarika-gradient inogadzirisa inobva kune cell state's near-linear update: c_t = f_t * c_{t-1} + i_t * g_t. Iyo yekukanganwa gedhi f_t (a sigmoid) inogona kugara padyo ne1, ichigadzira 'nguva dzose kukanganisa carousel' kuitira kuti masaini ekukanganisa ararame mukudzokera-kuburikidza-nguva mukati menguva refu. Magedhi ndiwo pachawo madiki neural layer (sigmoid yekugezera, tanh yemakoshero evamiriri), ese akadzidziswa akabatana ne gradient descent. Iri gedhi rinoita kuti network idzidze zvekuchengeta uye zvekurasa.
Kubata Memory Memory Cells Yenguva Yakareba Yakareba
Yakareba Yenguva Yenguva Yekurangarira (LSTM) maseru imhando yakakosha yeinodzokororwa neural network unit yakavakirwa kurangarira ruzivo mukati mekutevedzana kwakareba. Vakagadzirisa dambudziko rekunyangarika-gradient iro rakaremara maRNN apfuura, vachisimbisa makore gumi ekubudirira mumutauro, kutaura, uye kududzira. Yakareba Yenguva Ipfupi Memory Masero anogara mumusimboti AI toolkit. Paunonzwisisa, mamwe maAI misoro inova nyore kuongorora uye kuenzanisa. Kuti uvake kunzwisisa kwakadzama, tora Masero Akareba-Nguva Yekurangarira semuenzaniso wekushandisa, kwete chinhu chimwe chete: tsanangura zvinodikanwa, kujekesa fungidziro, uye patsanura zvinogona kuitwa nehurongwa hwakavimbika kubva kune zvichiri kuda kutonga kwenyanzvi.
Mukuita, zvikwata zvakasimba zvinoshandisa Yakareba-Yenguva Yenguva Memory Masero anovaka akasimba epfungwa mamodheru kutanga, obva anyora iwo modhi kune zvipingaidzo chaizvo zvekugadzira. Ivo vanonyora zvakajeka maitiro ebudiriro, bvunzo vachipokana ne data rechokwadi uye mafambiro ebasa, uye iterate zvichibva pane zvakacherechedzwa maitiro ekutadza kwete kuhwina-nguva imwe chete yebhenji. Apa ndipo apo kunzwisisa kwe theoretical kunoshanduka kuve kugona kwakasimba pane chigadzirwa, mutemo, uye mashandiro.
Inokubatsira kuparadzanisa zvakajeka zvichemo zvehunyanzvi kubva mumutauro wekushambadzira. Panguva imwecheteyo, Zvikwata zvakasiyana zvinogona kushandisa izwi rimwechete zvakasiyana, saka tsanangura nzvimbo nekukasira. Nzira yakatsiga ndeyekubatanidza kukurumidza kuyedza nekutonga: mhanyisa vatyairi vendege, tora humbowo, buritsa matanda esarudzo, uye urambe uchivandudza chengetedzo semaitiro emuenzaniso, zvinotarisirwa nemushandisi, uye zvinodikanwa zvekutonga.
Strategic Impact
Inokubatsira kuparadzanisa zvakajeka zvichemo zvehunyanzvi kubva mumutauro wekushambadzira.
Inokubatsira kuparadzanisa zvakajeka zvichemo zvehunyanzvi kubva mumutauro wekushambadzira. Mukutumirwa kwemhando yepamusoro, izvi zvinoshandurirwa kuita mitemo inoyerwa yekushanda, miganhu yevaridzi, uye tsika dzekudzokorora dzinodzokororwa kuitira kuti zvikwata zvikwire kuvimba pane kukwidza kusajeka.
Iwe unogona kubvunza zvirinani kuita mibvunzo usati washandisa mari kana nguva.
Iwe unogona kubvunza zvirinani kuita mibvunzo usati washandisa mari kana nguva. Mukutumirwa kwemhando yepamusoro, izvi zvinoshandurirwa kuita mitemo inoyerwa yekushanda, miganhu yevaridzi, uye tsika dzekudzokorora dzinodzokororwa kuitira kuti zvikwata zvikwire kuvimba pane kukwidza kusajeka.
Zvikwata zvine nzwisiso yakagovaniswa inoita zvirinani chigadzirwa, mutemo, uye sarudzo dzekudzidza.
Zvikwata zvine nzwisiso yakagovaniswa inoita zvirinani chigadzirwa, mutemo, uye sarudzo dzekudzidza. Mukutumirwa kwemhando yepamusoro, izvi zvinoshandurirwa kuita mitemo inoyerwa yekushanda, miganhu yevaridzi, uye tsika dzekudzokorora dzinodzokororwa kuitira kuti zvikwata zvikwire kuvimba pane kukwidza kusajeka.
Real-World Implementation
Kushandura nemuchina wemagetsi kutanga Google Translate's neural system Transformers isati yatanga kutonga.
Kuzivikanwa kwekutaura-kune-mavara muvabatsiri vezwi uye software yekuraira.
Kufanotaura kukosha kweramangwana munguva dzakatevedzana sekuda kwesimba, kuverenga sensor, kana mitengo yemasheya.
Kugadzira zvinyorwa kana mimhanzi tokeni imwe panguva uye kugadzirisa otomatiki kutevedzana.
Maitiro Ekuita
Yakareba Yenguva Ipfupi Memory Masero mukuita
Kushandura nemuchina wemagetsi kutanga Google Translate's neural system Transformers isati yatanga kutonga.
Kushandura muchina wemagetsi kutanga Google Shandura's neural system isati yatora Transformers Matimu anowanzowana mibairo iri nani paanotsanangura mabhindauko emhando kumberi, chengetedza nzira yekukwira kwevanhu yekesi dzemupendero, uye kuronda zvose zvinobudirira kubudirira uye kukanganisa kukanganisa nekufamba kwenguva.
Yakareba Yenguva Ipfupi Memory Masero mukuita
Kuzivikanwa kwekutaura-kune-mavara muvabatsiri vezwi uye software yekuraira.
Kuzivikanwa kwekutaura-kune-mavara muvabatsiri vezwi uye yekuraira software Matimu anowanzo kuwana mhedzisiro iri nani kana achinge atsanangura emhando yepamusoro kumberi, chengetedza nzira yekukwira kwevanhu yemakesi ekumucheto, uye kuteedzera zvese zvakawanikwa zvechigadzirwa uye mutengo wekukanganisa nekufamba kwenguva.
Yakareba Yenguva Ipfupi Memory Masero mukuita
Kufanotaura kukosha kweramangwana munguva dzakatevedzana sekuda kwesimba, kuverenga sensor, kana mitengo yemasheya.
Kufanotaura hunhu hweramangwana munguva dzakateerana senge simba rekuda, kuverenga sensor, kana mitengo yemasheya Matimu anowanzo kuwana mhedzisiro iri nani kana achinge atsanangura mhando yepamusoro kumberi, chengetedza nzira yekukwira kwevanhu yemakesi ekumucheto, uye kuteedzera zvese zvakawanikwa zvechigadzirwa uye mutengo wekukanganisa nekufamba kwenguva.
Yakareba Yenguva Ipfupi Memory Masero mukuita
Kugadzira zvinyorwa kana mimhanzi tokeni imwe panguva uye kugadzirisa otomatiki kutevedzana.
Kugadzira mameseji kana mimhanzi tokeni imwe panguva uye kuzadzisa otomatiki kutevedzana Matimu anowanzo kuwana mhedzisiro iri nani kana achinge atsanangura emhando yepamusoro kumberi, chengetedza nzira yekukwira kwevanhu yemakesi ekumucheto, uye kuteedzera zvese zvakawanikwa zvechigadzirwa nemitengo yekukanganisa nekufamba kwenguva.
Njodzi & Guardrails
Zvikwata zvakasiyana zvinogona kushandisa izwi rimwechete zvakasiyana, saka tsanangura nzvimbo nekukurumidza.
Benchmarks inogona kutaridzika yakasimba nepo chaiyo-yenyika kuita isina kuenzana.
Kuregeredza mhando yedata uye zvirongwa zvekuongorora zvinowanzogadzira mhedzisiro isina kusimba.
Implementation Roadmap
Tanga netsanangudzo yemutauro wakajeka yemhedzisiro yaunoda.
Tanga netsanangudzo yemutauro wakajeka yemhedzisiro yaunoda. Bata nhanho yega yega segedhi rehumbowo: kana maitiro asina kusangana, imbomira kuburitsa, vhara gaka, uye wobva wawedzera kushandiswa.
Sarudza metric imwe yekubudirira uye imwe yekutadza mamiriro usati waedzwa.
Sarudza metric imwe yekubudirira uye imwe yekutadza mamiriro usati waedzwa. Bata nhanho yega yega segedhi rehumbowo: kana maitiro asina kusangana, imbomira kuburitsa, vhara gaka, uye wobva wawedzera kushandiswa.
Mhanya mutyairi mudiki ane data remumiriri, kwete demo rakakwenenzverwa.
Mhanya mutyairi mudiki ane data remumiriri, kwete demo rakakwenenzverwa. Bata nhanho yega yega segedhi rehumbowo: kana maitiro asina kusangana, imbomira kuburitsa, vhara gaka, uye wobva wawedzera kushandiswa.
Nyora apo Masero Marefu-Nguva Yekurangarira anobatsira uye uko nzira dzakareruka dziri nani.
Nyora apo Masero Marefu-Nguva Yekurangarira anobatsira uye uko nzira dzakareruka dziri nani. Bata nhanho yega yega segedhi rehumbowo: kana maitiro asina kusangana, imbomira kuburitsa, vhara gaka, uye wobva wawedzera kushandiswa.