1 00:00:00,000 --> 00:00:02,928 >> [Music kucheza] 2 00:00:02,928 --> 00:00:11,230 3 00:00:11,230 --> 00:00:12,790 >> DAVID Malan: Hii ni CS50. 4 00:00:12,790 --> 00:00:14,072 Hii ni mwisho wa wiki 10. 5 00:00:14,072 --> 00:00:16,030 Na kijana, kufanya tuna nzuri darasani kwako leo. 6 00:00:16,030 --> 00:00:20,040 Sisi ni hivyo msisimko kukaribisha wawili wa marafiki zetu kutoka Yale juu yetu leo 7 00:00:20,040 --> 00:00:23,920 na kuangalia katika makutano ya bandia akili, robotteknik, 8 00:00:23,920 --> 00:00:25,710 usindikaji lugha ya asili, na zaidi. 9 00:00:25,710 --> 00:00:27,090 >> Na hakika, juu ya wiki chache zilizopita, tumekuwa 10 00:00:27,090 --> 00:00:29,714 hakika alitumia muda mwingi, hasa katika psets awali, 11 00:00:29,714 --> 00:00:31,560 kulenga pretty maelezo ngazi ya chini. 12 00:00:31,560 --> 00:00:34,520 Na ni rahisi sana kupoteza mbele ya msitu kwa ajili ya miti 13 00:00:34,520 --> 00:00:38,170 na kupata Hung juu ya mizunguko na hali na kuyatumia, hakika, na kadhalika. 14 00:00:38,170 --> 00:00:41,770 Lakini ukweli ni nyie sasa wana viungo na ambayo unaweza kweli 15 00:00:41,770 --> 00:00:45,300 kutatua baadhi ya matatizo ya kuvutia, miongoni mwa nao wale ambao marafiki zetu katika Yale 16 00:00:45,300 --> 00:00:48,790 kazi ya aibu tu ya Cambridge. 17 00:00:48,790 --> 00:00:52,410 >> Hivyo naomba kwanza kuanzisha kichwa wetu kufundisha msaidizi kutoka Yale, Andy. 18 00:00:52,410 --> 00:00:55,182 >> [Makofi] 19 00:00:55,182 --> 00:00:57,030 20 00:00:57,030 --> 00:01:00,710 >> ANDY: Awali ya yote, kumshukuru tu wewe kwa kuruhusu Yalies wanandoa 21 00:01:00,710 --> 00:01:02,700 pop juu ya chini ya Cambridge leo. 22 00:01:02,700 --> 00:01:05,299 Kweli sisi kufahamu hilo. 23 00:01:05,299 --> 00:01:07,090 Pili, kwa marafiki zetu nyuma home-- Jason, 24 00:01:07,090 --> 00:01:09,670 shukrani kwa ajili ya kukaa na kukimbia hotuba. 25 00:01:09,670 --> 00:01:12,970 Matumaini ni njema katika New Haven. 26 00:01:12,970 --> 00:01:15,720 >> Hivyo yeah, mimi nina super msisimko kuanzisha Scaz leo. 27 00:01:15,720 --> 00:01:17,020 Scaz anaendesha robotteknik maabara. 28 00:01:17,020 --> 00:01:19,690 Yeye ni profesa wa, kama, tano idara mbalimbali katika Yale. 29 00:01:19,690 --> 00:01:23,159 Katika maabara yake, ana wengi, wengi robots kwamba yeye anapenda kucheza na. 30 00:01:23,159 --> 00:01:24,950 Ana, kama, coolest kazi katika dunia. 31 00:01:24,950 --> 00:01:27,116 Na yeye anapata aina ya fujo karibu na kwamba siku zote 32 00:01:27,116 --> 00:01:29,090 muda mrefu na kufanya baadhi ya kazi, kama vile. 33 00:01:29,090 --> 00:01:33,070 >> Na hivyo sisi kweli kuletwa moja wao huko pamoja nasi leo. 34 00:01:33,070 --> 00:01:36,900 Hivyo bila wasiwasi zaidi, Scaz ni kwenda mbele na kuanzisha sisi 35 00:01:36,900 --> 00:01:38,660 kwa robot rafiki yake. 36 00:01:38,660 --> 00:01:41,546 >> [Makofi] 37 00:01:41,546 --> 00:01:42,935 38 00:01:42,935 --> 00:01:44,310 BRIAN SCASSELLATI: Shukrani, Daudi. 39 00:01:44,310 --> 00:01:45,380 Shukrani, Andy. 40 00:01:45,380 --> 00:01:50,050 Ni ajabu sana kuwa hapa na kila mtu leo. 41 00:01:50,050 --> 00:01:56,490 Nataka kwanza kuwa wazi kabisa kuwa CS50 wafanyakazi hapa katika Cambridge 42 00:01:56,490 --> 00:01:59,190 imekuwa incredibly ukarimu kwetu. 43 00:01:59,190 --> 00:02:02,130 Sisi ni hivyo kushukuru kwa kila kitu wamefanya kutuunga mkono. 44 00:02:02,130 --> 00:02:05,690 Na hivyo tunatarajia kuwa na uwezo kurudi wema. 45 00:02:05,690 --> 00:02:09,370 >> Hivyo leo, sisi kupata kutangaza kwamba sisi ni kwenda na mpya, 46 00:02:09,370 --> 00:02:15,240 moja ya-ya aina ya tukio CS50 kinachotokea katika New Haven wiki ijayo. 47 00:02:15,240 --> 00:02:18,040 Na hii ni CS50 Utafiti Expo. 48 00:02:18,040 --> 00:02:21,300 Hivyo sisi ni kwenda kuwa kuwakaribisha everyone-- CS50 wanafunzi, 49 00:02:21,300 --> 00:02:26,510 wafanyakazi kutoka pande zote mbili Harvard na Yale-- kwa ashuke na kutembelea na sisi siku ya Ijumaa. 50 00:02:26,510 --> 00:02:30,400 Kutakuwa na aina mbalimbali ya juu ya 30 watu mbalimbali kuwasilisha 51 00:02:30,400 --> 00:02:34,830 na upperclassmen exhibiting-- kuonyesha mbali baadhi ya bidhaa zao za utafiti. 52 00:02:34,830 --> 00:02:38,480 Kutakuwa na baadhi startups, hata, kuangalia kwa kidogo ya mwezi tech vipaji, 53 00:02:38,480 --> 00:02:40,460 startups kutoka kwa Harvard na Yale. 54 00:02:40,460 --> 00:02:44,550 Na kutakuwa na baadhi ya vikundi mwanafunzi kuangalia kwa baadhi ya uanachama mapya. 55 00:02:44,550 --> 00:02:46,357 >> Ni kwenda kuwa wakati kusisimua sana. 56 00:02:46,357 --> 00:02:49,190 Hopefully wale ambao ni wakishuka kwa mchezo Harvard-Yale 57 00:02:49,190 --> 00:02:51,360 utakuwa na uwezo wa kuacha na kidogo mapema, 58 00:02:51,360 --> 00:02:54,060 haki ya katikati ya chuo hicho, Sterling Memorial Library. 59 00:02:54,060 --> 00:02:58,040 Tunakwenda kuwa na seti ya maonyesho mbalimbali kutoka uhuru 60 00:02:58,040 --> 00:03:04,460 sailboats kwa njia za kutumia programu kuhifadhi miswada medieval. 61 00:03:04,460 --> 00:03:07,860 >> Tunakwenda kuwa na tangazo hoc mitandao na watu 62 00:03:07,860 --> 00:03:11,230 kufundisha programu coding katika Cape Town. 63 00:03:11,230 --> 00:03:13,730 Tutaweza kuwa na kompyuta muziki maandamano. 64 00:03:13,730 --> 00:03:16,020 Na tutaweza bila shaka na robots zaidi. 65 00:03:16,020 --> 00:03:18,900 Hivyo hatuna matumaini utasikia kujiunga na sisi kwa ajili ya tukio hili. 66 00:03:18,900 --> 00:03:21,350 Ni lazima kuwa mengi ya furaha, kidogo kidogo ya chakula, 67 00:03:21,350 --> 00:03:24,430 na mengi ya kuvutia mambo ya kuzungumzia. 68 00:03:24,430 --> 00:03:28,230 >> Hivyo leo, sisi ni kwenda kuzungumza kuhusu usindikaji lugha ya asili. 69 00:03:28,230 --> 00:03:32,560 Na hii ni jaribio kwa ajili yetu kujenga njia mpya ya interfacing 70 00:03:32,560 --> 00:03:35,150 na vifaa yetu kwa sababu kwa muda wa wiki chache zilizopita, 71 00:03:35,150 --> 00:03:40,800 tumekuwa kulenga jinsi gani ni kwamba unaweza kuandika kanuni, kuandika programu 72 00:03:40,800 --> 00:03:47,110 kuwa ni njia ya kuwa na uwezo wa kusema kwa mashine, hii ni nini nataka ufanye. 73 00:03:47,110 --> 00:03:50,210 >> Lakini hatupaswi haja ya kutarajia kwamba kila kitu 74 00:03:50,210 --> 00:03:53,760 hiyo ni huko nje kwamba kutumika na kila mtu katika dunia 75 00:03:53,760 --> 00:03:57,480 ni kwenda kuwa na magari katika aina hii ya mafundisho. 76 00:03:57,480 --> 00:04:02,540 Hivyo sisi kutofautisha kati ya kompyuta Lugha na languages-- asili 77 00:04:02,540 --> 00:04:06,720 yaani, mambo ambayo binadamu matumizi kuwasiliana na wanadamu wengine. 78 00:04:06,720 --> 00:04:12,270 Na sisi kujaribu kujenga interfaces kwamba matumizi hizi taratibu za mawasiliano ya asili. 79 00:04:12,270 --> 00:04:16,029 >> Sasa, tu kama kila mada mengine kwamba tumekuwa ilianza na katika CS50, 80 00:04:16,029 --> 00:04:19,589 tunakwenda kuanza na rahisi kidogo ya usindikaji lugha za asili 81 00:04:19,589 --> 00:04:21,269 tuweze kufikiria. 82 00:04:21,269 --> 00:04:24,940 Tunakwenda kuanza na sehemu ya kihistoria ya lugha za asili. 83 00:04:24,940 --> 00:04:28,740 Na kisha tutaweza kujenga hadi zaidi na zaidi mifumo ya hivi karibuni 84 00:04:28,740 --> 00:04:31,450 na kuwa na baadhi ya kujifurahisha demos njiani. 85 00:04:31,450 --> 00:04:35,900 >> Hivyo sisi ni kwenda kuanza na nini ilikuwa pengine kwanza lugha ya asili 86 00:04:35,900 --> 00:04:38,210 mfumo usindikaji. 87 00:04:38,210 --> 00:04:45,080 Hii ilikuwa ni mfumo wa programu iliyoandikwa katika 1966 na Joseph Weizenbaum aitwaye ELIZA. 88 00:04:45,080 --> 00:04:49,640 Na ELIZA ilitengenezwa kuiga aina ya mwingiliano 89 00:04:49,640 --> 00:04:53,850 ingekuwa na Rogerian tibamaungo. 90 00:04:53,850 --> 00:04:57,210 Sasa, Rogerians, walikuwa na wazo kwamba tiba za kisaikolojia 91 00:04:57,210 --> 00:05:02,800 wanaohusika kuwa na uwezo wa kioo nyuma kwa mgonjwa na kuzungumza nao, 92 00:05:02,800 --> 00:05:08,100 kimsingi, na tu kuwapa vidogo kidogo ya mtaalamu. 93 00:05:08,100 --> 00:05:09,920 Yaani, kila kitu kuwa mtaalamu alisema 94 00:05:09,920 --> 00:05:16,500 ilitakiwa kuwa tu tafakari ya nini mgonjwa alikuwa akiwaambia kwao. 95 00:05:16,500 --> 00:05:18,990 >> Basi hebu jaribu na demo hii. 96 00:05:18,990 --> 00:05:22,820 Je, tuna kujitolea ambao d kuwa tayari kushiriki ndani kabisa wao, 97 00:05:22,820 --> 00:05:26,650 giza, na / au siri bandia na ELIZA? 98 00:05:26,650 --> 00:05:29,680 Hakika, kwa nini wewe kuja juu juu. 99 00:05:29,680 --> 00:05:30,847 Ajabu. 100 00:05:30,847 --> 00:05:31,430 Na jina lako? 101 00:05:31,430 --> 00:05:32,000 >> MILES: Maili. 102 00:05:32,000 --> 00:05:33,041 >> BRIAN SCASSELLATI: Maili. 103 00:05:33,041 --> 00:05:34,270 Vyema kukutana na wewe, Maili. 104 00:05:34,270 --> 00:05:36,270 Ukitaka kushikilia kwenye hilo. 105 00:05:36,270 --> 00:05:40,840 Na tunakwenda kutumia toleo la ELIZA 106 00:05:40,840 --> 00:05:44,640 kwamba tumekuwa kuigwa ndani ya CS50 IDE. 107 00:05:44,640 --> 00:05:47,140 Na tutaangalia chanzo kificho kwa hii katika muda mfupi. 108 00:05:47,140 --> 00:05:49,910 Lakini kwa sasa hivi, tuko tu kwenda waomba hilo kwa kuandika eliza. 109 00:05:49,910 --> 00:05:55,480 >> Na ELIZA ni kwenda kukupa majibu kwamba ni kwenda kuwa katika miji mikuu yote. 110 00:05:55,480 --> 00:05:57,950 Na kwenda kuuliza baadhi ya maswali. 111 00:05:57,950 --> 00:06:02,130 Na tutaweza kujibu na aina katika majibu ya ELIZA. 112 00:06:02,130 --> 00:06:04,310 Hivyo ELIZA amesema, "Hi, mimi nina ELIZA. 113 00:06:04,310 --> 00:06:05,462 Nini tatizo lako? " 114 00:06:05,462 --> 00:06:08,092 Nini ungependa kuwaambia ELIZA kuhusu leo? 115 00:06:08,092 --> 00:06:10,750 >> MILES: Naweza tu usingizi kama nina Bunny slippers juu. 116 00:06:10,750 --> 00:06:12,440 >> BRIAN SCASSELLATI: zote haki. 117 00:06:12,440 --> 00:06:13,360 Sawa. 118 00:06:13,360 --> 00:06:15,300 >> [Kicheko] 119 00:06:15,300 --> 00:06:22,410 >> "Siwezi tu usingizi na Bunny slippers juu. " 120 00:06:22,410 --> 00:06:29,000 Na ELIZA anasema, "Sema, je, kuwa na matatizo yoyote ya kisaikolojia? 121 00:06:29,000 --> 00:06:30,450 Labda hii ni kiashiria nzuri. " 122 00:06:30,450 --> 00:06:30,780 >> SAWA. 123 00:06:30,780 --> 00:06:32,245 Hivyo nini ungependa kujibu? 124 00:06:32,245 --> 00:06:36,110 125 00:06:36,110 --> 00:06:36,970 >> MILES: Ndiyo. 126 00:06:36,970 --> 00:06:38,660 >> BRIAN SCASSELLATI: Ndiyo. 127 00:06:38,660 --> 00:06:39,570 Mtu wa maneno mengi. 128 00:06:39,570 --> 00:06:41,870 Mimi kufurahia jambo hilo. 129 00:06:41,870 --> 00:06:50,200 Eliza anasema, "Wewe wanaonekana kabisa chanya." 130 00:06:50,200 --> 00:06:53,200 >> MILES: Hapana, mimi ni, kwa kweli, elektroni. 131 00:06:53,200 --> 00:06:55,150 >> BRIAN SCASSELLATI: "Hapana 132 00:06:55,150 --> 00:07:02,240 Kwa kweli, mimi ni elektroni. " "Je, wewe unakuja kwangu 133 00:07:02,240 --> 00:07:06,954 kwa sababu wewe ni Electron? " 134 00:07:06,954 --> 00:07:08,870 Sasa, tunaweza kuona baadhi things-- mimi nina kwenda 135 00:07:08,870 --> 00:07:10,200 kutulia wewe kwa haki ya pili, Maili. 136 00:07:10,200 --> 00:07:11,010 Je, si kwenda popote. 137 00:07:11,010 --> 00:07:12,650 Sisi siyo kosa kwa wewe bado. 138 00:07:12,650 --> 00:07:16,490 ELIZA ni, katika baadhi ya njia, kukabiliana na mambo fulani 139 00:07:16,490 --> 00:07:21,140 kwamba yeye kusema lakini bila uelewa halisi. 140 00:07:21,140 --> 00:07:26,090 Na ni pia parroting nyuma baadhi sehemu ya adhabu yake. 141 00:07:26,090 --> 00:07:27,310 >> Hivyo sisi alisema, "Hapana 142 00:07:27,310 --> 00:07:29,090 Kwa kweli, mimi ni elektroni. " 143 00:07:29,090 --> 00:07:34,190 Eliza alijibu nyuma, "Je, wewe kuja kwangu kwa sababu wewe ni Electron? " 144 00:07:34,190 --> 00:07:34,690 Sawa. 145 00:07:34,690 --> 00:07:36,050 Hebu kuwapa moja zaidi majibu hapa. 146 00:07:36,050 --> 00:07:37,140 Kuwapa moja nzuri, Maili. 147 00:07:37,140 --> 00:07:37,806 >> MILES: Good moja? 148 00:07:37,806 --> 00:07:48,400 149 00:07:48,400 --> 00:07:49,570 >> Ndiyo, tafadhali nisaidie. 150 00:07:49,570 --> 00:07:50,570 BRIAN SCASSELLATI: "Ndiyo. 151 00:07:50,570 --> 00:07:58,010 Tafadhali nisaidie. "" Una uhakika? " 152 00:07:58,010 --> 00:07:58,510 MILES: Hapana 153 00:07:58,510 --> 00:08:00,860 BRIAN SCASSELLATI: Hapana 154 00:08:00,860 --> 00:08:03,562 "Je, hiyo inaashiria nini?" 155 00:08:03,562 --> 00:08:06,760 >> MILES: Mimi inaweza kuwa Electron. 156 00:08:06,760 --> 00:08:11,170 >> BRIAN SCASSELLATI: "Mimi inaweza kuwa Electron. " 157 00:08:11,170 --> 00:08:13,790 "Je, akisema hakuna tu kuwa mbaya?" 158 00:08:13,790 --> 00:08:14,390 >> Sawa. 159 00:08:14,390 --> 00:08:16,030 Hivyo Maili, asante sana. 160 00:08:16,030 --> 00:08:16,860 >> MILES: Asante. 161 00:08:16,860 --> 00:08:18,360 >> BRIAN SCASSELLATI: Asante sana. 162 00:08:18,360 --> 00:08:20,230 [Makofi] 163 00:08:20,230 --> 00:08:21,230 164 00:08:21,230 --> 00:08:26,700 >> Hivyo mpango huu ELIZA, kwa njia nyingi, ni kukabiliana tu kwa vipande 165 00:08:26,700 --> 00:08:32,730 kwamba sisi ni kutoa bila kina uelewa wa nini kinatokea hapa. 166 00:08:32,730 --> 00:08:36,640 Hii ni aina ya mfumo aitwaye muundo vinavyolingana, ambapo 167 00:08:36,640 --> 00:08:40,490 sisi ni kuangalia kwa baadhi ya vipande vya maandishi kwamba tuko kisha 168 00:08:40,490 --> 00:08:44,860 kwenda kuchukua nje ya kile yalitolewa kama pembejeo, 169 00:08:44,860 --> 00:08:52,580 kubadilisha, uwezekano, kwa namna fulani, na kisha kuwapa nyuma kwa mtumiaji. 170 00:08:52,580 --> 00:08:55,150 >> Je katika unadhani kwamba ELIZA ni kweli 171 00:08:55,150 --> 00:08:58,230 kufanya halali kisaikolojia hapa? 172 00:08:58,230 --> 00:08:59,250 Mtu mmoja, labda. 173 00:08:59,250 --> 00:09:00,166 >> Watazamaji: [inaudible]. 174 00:09:00,166 --> 00:09:03,315 175 00:09:03,315 --> 00:09:05,440 BRIAN SCASSELLATI: Na jinsi je, hiyo kufanya kujisikia? 176 00:09:05,440 --> 00:09:06,530 Ndiyo, kwa kweli, ni gani. 177 00:09:06,530 --> 00:09:10,890 Na tunakwenda kuona, kwa kweli, chanzo kanuni kwa hiyo katika muda tu. 178 00:09:10,890 --> 00:09:13,580 Na hivyo wewe ni kwenda kuwa uwezo wa kufanya hasa hili. 179 00:09:13,580 --> 00:09:17,420 >> Sasa, ELIZA ni aina moja ya yale tunataka kuwaita leo mazungumzo bot. 180 00:09:17,420 --> 00:09:19,950 Ni tu inakwenda kupitia Nakala kwamba wewe ni kutoa, 181 00:09:19,950 --> 00:09:24,030 hutoa kiasi cha chini wazi ya uelewa au usindikaji, 182 00:09:24,030 --> 00:09:26,790 na kisha parrots nyuma na wewe. 183 00:09:26,790 --> 00:09:31,830 Hivyo hebu tuangalie, conceptually, na kuzungumza kuhusu nini 184 00:09:31,830 --> 00:09:34,690 ni kwamba ELIZA ni kweli kufanya. 185 00:09:34,690 --> 00:09:42,000 >> ELIZA ni kuchukua sentence-- hebu kusema, "Nataka kumvutia bosi wangu." 186 00:09:42,000 --> 00:09:45,130 Na ELIZA ni kutafuta kupitia hukumu hiyo 187 00:09:45,130 --> 00:09:48,730 na kujaribu kutafuta na mechi ya mifumo fulani. 188 00:09:48,730 --> 00:09:52,850 Hivyo, kwa mfano, moja ya mifumo kwamba ELIZA ni kutafuta ni maneno 189 00:09:52,850 --> 00:09:55,110 "Nataka." 190 00:09:55,110 --> 00:09:59,330 Na wakati wowote anaona kitu ambayo ina "Nataka" ndani yake, 191 00:09:59,330 --> 00:10:01,770 ni kujiwekea majibu. 192 00:10:01,770 --> 00:10:05,040 Na majibu kwamba ni kamba fasta. 193 00:10:05,040 --> 00:10:07,915 Katika kesi hiyo, ni "kwa nini unataka?" 194 00:10:07,915 --> 00:10:11,330 Na mimi kuweka nyota kidogo katika mwisho kwa sababu hiyo ni 195 00:10:11,330 --> 00:10:13,310 mwanzo wa mwitikio wetu. 196 00:10:13,310 --> 00:10:16,310 Na nyota inaonyesha kwamba tunakwenda kuchukua mapumziko 197 00:10:16,310 --> 00:10:19,850 ya mtumiaji utterance-- "kumvutia bosi wangu" - 198 00:10:19,850 --> 00:10:24,500 na tunakwenda append kwamba kwenye mwisho wa kamba hii. 199 00:10:24,500 --> 00:10:28,990 >> Hivyo sasa, badala ya kusema, "kwa nini Je, unataka kumvutia bosi wangu, " 200 00:10:28,990 --> 00:10:31,800 kuna kidogo ya ziada usindikaji kwamba tutaweza kufanya. 201 00:10:31,800 --> 00:10:34,440 Yaani, sisi itabidi kubadilisha baadhi ya Viwakilishi 202 00:10:34,440 --> 00:10:38,670 hapa kutoka "bosi wangu" na "bosi wako." 203 00:10:38,670 --> 00:10:41,300 Na huenda kuna wengine wachache mabadiliko ambayo tunahitaji kufanya. 204 00:10:41,300 --> 00:10:44,990 Hivyo badala ya kujikita ni moja kwa moja kwenye mwisho, nini tutaweza kufanya 205 00:10:44,990 --> 00:10:49,160 ni tutaweza kuchukua mapumziko ya mtumiaji utterance-- katika nyeupe here-- 206 00:10:49,160 --> 00:10:54,090 na tutaweza kuchukua ni kitu kimoja katika muda na kubadilisha kila kamba 207 00:10:54,090 --> 00:10:58,180 ishara, kila neno, katika hukumu. 208 00:10:58,180 --> 00:10:59,580 >> Hivyo tutaweza kuchukua neno "kwa". 209 00:10:59,580 --> 00:11:01,650 Hakuna kubadilika kuwa tunahitaji kufanya hivyo. 210 00:11:01,650 --> 00:11:02,340 "Kumvutia." 211 00:11:02,340 --> 00:11:04,140 Hakuna kubadilika tunahitaji kufanya huko. 212 00:11:04,140 --> 00:11:06,670 "Yangu" itakuwa kubadilisha na "wako." 213 00:11:06,670 --> 00:11:10,070 Na "bosi" tutaweza kuondoka tu kama "bosi." 214 00:11:10,070 --> 00:11:12,740 Na kisha hatimaye, kitu chochote kwamba mwisho kwa kipindi, 215 00:11:12,740 --> 00:11:16,640 tutaweza kubadili kwenye swali. 216 00:11:16,640 --> 00:11:22,600 >> Hii rahisi sana muundo vinavyolingana ni kweli na mafanikio. 217 00:11:22,600 --> 00:11:27,260 Na wakati huu ulianzishwa katika 1966-- Joseph Weizenbaum 218 00:11:27,260 --> 00:11:28,986 iliyowekwa hii kwenye kompyuta. 219 00:11:28,986 --> 00:11:31,110 Sasa, kompyuta wakati huo hawakuwa mifano desktop. 220 00:11:31,110 --> 00:11:33,950 Walikuwa pamoja rasilimali. 221 00:11:33,950 --> 00:11:39,090 Na wanafunzi wake ungekuwa kwenda na kuzungumza na ELIZA. 222 00:11:39,090 --> 00:11:41,570 Hatimaye, yeye alikuwa na kudhibiti upatikanaji wa ni 223 00:11:41,570 --> 00:11:43,890 kwa sababu wanafunzi wake walikuwa si kupata kazi yoyote kufanyika. 224 00:11:43,890 --> 00:11:46,190 Walikuwa kuzungumza tu kwa ELIZA. 225 00:11:46,190 --> 00:11:48,850 Na, kwa kweli, yeye alikuwa na moto msaidizi wake, ambaye 226 00:11:48,850 --> 00:11:55,840 alitumia wote wa muda wake kuzungumza na ELIZA kuhusu matatizo yake ya kina na inatia mashaka. 227 00:11:55,840 --> 00:12:00,350 >> Kila mtu ambaye alitumia mifumo hii kuanza anthropomorphize yao. 228 00:12:00,350 --> 00:12:04,490 Walianza kufikiri wao kama kuwa hai na watu halisi. 229 00:12:04,490 --> 00:12:07,969 Walianza kutambua baadhi ya mambo ambayo wao walikuwa wakisema 230 00:12:07,969 --> 00:12:09,010 wanakuja nyuma yao. 231 00:12:09,010 --> 00:12:12,120 Nao walikuwa kutafuta mambo kuhusu wao wenyewe. 232 00:12:12,120 --> 00:12:17,290 Na, kwa kweli, hata wataalam, hata psychotherapists, 233 00:12:17,290 --> 00:12:22,930 kuanza kuwa na wasiwasi kwamba, kwa kweli, labda ELIZA itakuwa kuondoa yao. 234 00:12:22,930 --> 00:12:25,640 Na hata kompyuta wanasayansi wasiwasi kwamba tulikuwa 235 00:12:25,640 --> 00:12:30,040 karibu na kutatua lugha ya asili. 236 00:12:30,040 --> 00:12:33,520 >> Sasa, hiyo haikuwa mahali popote karibu na kweli. 237 00:12:33,520 --> 00:12:37,280 Lakini hiyo ni jinsi ya kuvutia mifumo hii inaweza kuonekana. 238 00:12:37,280 --> 00:12:40,080 Basi hebu kuanza kuangalia chini na kujaribu 239 00:12:40,080 --> 00:12:46,190 kupata kidogo ya swali wapi kificho hii kwa kweli hutokea. 240 00:12:46,190 --> 00:12:48,170 Hivyo tutaweza kufanya kanuni hii inapatikana baadaye. 241 00:12:48,170 --> 00:12:50,880 Na hii ni sana bandari rahisi na ya moja kwa moja 242 00:12:50,880 --> 00:12:53,240 ya awali ELIZA utekelezaji. 243 00:12:53,240 --> 00:12:56,350 >> Hivyo baadhi ya hizi Stylistic mambo ambayo utaona hapa 244 00:12:56,350 --> 00:12:59,360 si stylistically nini tunataka wanataka kufanya 245 00:12:59,360 --> 00:13:01,480 au nini tumekuwa kufundisha wewe kufanya. 246 00:13:01,480 --> 00:13:04,770 Lakini tumekuwa alijaribu kuwaweka namna tofauti katika bandari wengi 247 00:13:04,770 --> 00:13:08,087 kwamba hii imekuwa na hivyo kuwa ni ina ladha ya awali. 248 00:13:08,087 --> 00:13:09,920 Hivyo sisi ni kwenda kwa pamoja rundo la vitu, 249 00:13:09,920 --> 00:13:12,920 na kisha tutaweza kuwa seti ya maneno, mambo 250 00:13:12,920 --> 00:13:16,460 kwamba ELIZA kutambua na kujibu moja kwa moja. 251 00:13:16,460 --> 00:13:20,780 Hivyo kama una maneno kama "unaweza" au "I do not" au "hapana" 252 00:13:20,780 --> 00:13:24,680 au "ndiyo" au "ndoto" au "hello," kisha ELIZA 253 00:13:24,680 --> 00:13:27,920 kujibu selectively kwa wale. 254 00:13:27,920 --> 00:13:30,010 Tutaweza pia kuwa baadhi idadi ya mambo 255 00:13:30,010 --> 00:13:34,940 kwamba sisi wabadilishane, kama kuwabadili "yangu" na "wako." 256 00:13:34,940 --> 00:13:39,920 >> Na kisha tutaweza kuwa na seti ya majibu kwamba kwa kila moja ya maneno haya, 257 00:13:39,920 --> 00:13:42,580 tutaweza mzunguko kupitia hizi majibu tofauti. 258 00:13:42,580 --> 00:13:45,350 Hivyo kama mimi kusema "ndiyo" mara tatu mfululizo, mimi 259 00:13:45,350 --> 00:13:50,429 anaweza kupata tatu tofauti majibu kutoka ELIZA. 260 00:13:50,429 --> 00:13:52,345 Kanuni zetu, basi, ni kweli inashangaza rahisi. 261 00:13:52,345 --> 00:13:59,490 Kama mimi kitabu chini uliopita yote haya majibu kwamba tuna iliyowekwa katika 262 00:13:59,490 --> 00:14:02,920 na sisi kupata chini ya kuu yetu, tunakwenda initialize 263 00:14:02,920 --> 00:14:06,540 michache ya vigezo mbalimbali na kufanya kidogo ya utunzaji wa nyumba 264 00:14:06,540 --> 00:14:08,480 katika mwanzo. 265 00:14:08,480 --> 00:14:11,760 Lakini basi kuna kabisa kuweka wa kanuni uweze kuelewa. 266 00:14:11,760 --> 00:14:15,820 Moja kubwa kitanzi wakati kwamba anasema mimi nina kwenda kurudia hiki tena na tena. 267 00:14:15,820 --> 00:14:20,420 Mimi itabidi kusoma katika mstari, na mimi itabidi kuhifadhi kwamba katika kamba pembejeo. 268 00:14:20,420 --> 00:14:23,880 Mimi itabidi kuangalia na kuona kama ni keyword maalum "bye," ambayo 269 00:14:23,880 --> 00:14:26,199 ina maana kujinasua mpango. 270 00:14:26,199 --> 00:14:29,240 Na kisha mimi itabidi kuangalia na kuona kama mtu ni kurudia tu wenyewe 271 00:14:29,240 --> 00:14:29,800 tena na tena. 272 00:14:29,800 --> 00:14:31,174 Na mimi itabidi yell saa yao kama hawana. 273 00:14:31,174 --> 00:14:34,820 Mimi itabidi kusema "wala kurudia mwenyewe." 274 00:14:34,820 --> 00:14:40,500 >> Muda mrefu kama hakuna wa wale kutokea, tutaweza kisha Scan kupitia na kitanzi kwa njia, 275 00:14:40,500 --> 00:14:45,330 juu ya mistari 308-313 hapa, na angalia uone 276 00:14:45,330 --> 00:14:49,090 ni yoyote ya wale keyword misemo zilizomo katika pembejeo 277 00:14:49,090 --> 00:14:50,620 kwamba nilikuwa tu kupewa? 278 00:14:50,620 --> 00:14:54,845 Kama kuna mechi kwa ajili yao, vizuri basi, mimi itabidi kukumbuka eneo hilo. 279 00:14:54,845 --> 00:14:57,050 Mimi itabidi kukumbuka kwamba keyword. 280 00:14:57,050 --> 00:14:58,620 Na mimi itabidi kuwa na uwezo wa kujenga majibu. 281 00:14:58,620 --> 00:15:03,150 >> Kama mimi si kupata moja, vizuri basi, Jambo la mwisho katika safu yangu keyword 282 00:15:03,150 --> 00:15:08,070 itakuwa majibu yangu ya msingi, wakati kitu kingine mechi. 283 00:15:08,070 --> 00:15:14,160 Mimi itabidi kuuliza maswali kama "Kwa nini wewe kuja hapa? "au" Ninawezaje kukusaidia? " 284 00:15:14,160 --> 00:15:19,710 kwamba ni tu sehemu sahihi bila kujali pembejeo ni. 285 00:15:19,710 --> 00:15:22,580 >> Tutaweza basi kujenga majibu ELIZA ya. 286 00:15:22,580 --> 00:15:26,040 Tutaweza kuwa na uwezo wa kuchukua kwamba majibu ya msingi, 287 00:15:26,040 --> 00:15:28,370 tu kama tulivyofanya kwa kuwa "bosi wangu" mfano. 288 00:15:28,370 --> 00:15:30,970 289 00:15:30,970 --> 00:15:33,990 Kama hayo ni yote kwamba kuna is-- kama ni tu moja 290 00:15:33,990 --> 00:15:36,860 kamba kwamba natakiwa respond-- Naweza tu kutuma nyuma nje. 291 00:15:36,860 --> 00:15:40,610 Kama ana kinyota kwenye mwisho wake, basi mimi itabidi 292 00:15:40,610 --> 00:15:45,710 mchakato kila ishara ya mtu binafsi katika mapumziko ya majibu ya mtumiaji 293 00:15:45,710 --> 00:15:51,590 na kuongeza wale walio katika, swapping nje neno kwa neno kama nahitaji. 294 00:15:51,590 --> 00:15:56,100 >> Yote hii ni kabisa kitu ambacho ungeweza kujenga. 295 00:15:56,100 --> 00:15:59,230 Na kwa kweli, njia ambazo sisi na kusindika hoja mstari amri, 296 00:15:59,230 --> 00:16:03,570 njia ambao una kusindika kwa njia ya maombi HTTP 297 00:16:03,570 --> 00:16:05,510 kufuata aina hiyo ya sheria. 298 00:16:05,510 --> 00:16:08,220 Wao ni muundo vinavyolingana. 299 00:16:08,220 --> 00:16:15,170 >> Hivyo ELIZA alikuwa na kiasi muhimu athari kwa lugha za asili 300 00:16:15,170 --> 00:16:21,620 kwa sababu alifanya hivyo kuonekana kama ilivyokuwa lengo attainable sana, kama namna fulani tunatarajia 301 00:16:21,620 --> 00:16:25,550 kuwa na uwezo wa kutatua tatizo hili moja kwa moja. 302 00:16:25,550 --> 00:16:30,670 Sasa, kwamba si kusema kwamba ELIZA gani kila kitu kwamba tunataka wanataka kufanya. 303 00:16:30,670 --> 00:16:33,710 Hakika si. 304 00:16:33,710 --> 00:16:35,660 Lakini tunapaswa kuwa na uwezo kufanya kitu zaidi. 305 00:16:35,660 --> 00:16:38,280 306 00:16:38,280 --> 00:16:41,490 >> Hatua yetu ya kwanza kwenda zaidi ya ELIZA ni kwenda 307 00:16:41,490 --> 00:16:44,840 kuwa na uwezo wa kuangalia si maandishi kuwa aliingia 308 00:16:44,840 --> 00:16:53,750 ndani ya keyboard lakini hotuba, halisi hotuba kumbukumbu katika kipaza sauti. 309 00:16:53,750 --> 00:16:56,880 Hivyo kama sisi kuangalia hizi vipande mbalimbali, tuko 310 00:16:56,880 --> 00:17:00,304 kwenda na kujenga seti ya mifano. 311 00:17:00,304 --> 00:17:02,970 Tunakwenda kuwa na uwezo kwenda kutoka ngazi ya chini akustisk 312 00:17:02,970 --> 00:17:07,180 information-- lami, amplitude, frequency-- 313 00:17:07,180 --> 00:17:09,530 na kubadilisha kwamba katika baadhi ya vitengo kuwa tuko 314 00:17:09,530 --> 00:17:14,619 uwezo kwa urahisi zaidi kuendesha na, hatimaye, kuendesha yao 315 00:17:14,619 --> 00:17:18,609 ndani ya maneno na sentensi. 316 00:17:18,609 --> 00:17:22,880 >> Kwa hiyo wengi kutambua hotuba mifumo ya kwamba ni huko nje leo 317 00:17:22,880 --> 00:17:26,069 kufuata takwimu mfano ambayo sisi kujenga 318 00:17:26,069 --> 00:17:35,090 uwakilishi matatu tofauti ya yale kuwa ishara ya kusikiliza kweli ina. 319 00:17:35,090 --> 00:17:38,640 Tunaweza kuanza na mtindo wa kifonetiki kwamba mazungumzo juu ya msingi tu 320 00:17:38,640 --> 00:17:41,250 sauti kwamba mimi nina kuzalisha. 321 00:17:41,250 --> 00:17:46,900 Mimi ni kuzalisha kitu ambacho ni B kama katika mvulana au D kama katika mbwa? 322 00:17:46,900 --> 00:17:53,220 Je, mimi kutambua hayo mawili tofauti Simu za kama tofauti na tofauti? 323 00:17:53,220 --> 00:17:56,600 >> Juu ya kwamba, tutaweza kisha kujenga neno matamshi mfano, 324 00:17:56,600 --> 00:18:01,350 kitu ambacho viungo pamoja Simu za wale mtu binafsi 325 00:18:01,350 --> 00:18:04,724 na unachanganya yao katika neno. 326 00:18:04,724 --> 00:18:07,890 Na baada ya hapo, tutaweza kuchukua maneno na tutaweza kuwakusanya na lugha 327 00:18:07,890 --> 00:18:13,010 mfano katika hukumu kamili. 328 00:18:13,010 --> 00:18:17,230 >> Sasa, sisi ni kwenda kuzungumza kuhusu kila ya hizi kujitegemea na tofauti. 329 00:18:17,230 --> 00:18:21,580 Lakini mifano ya haya tatu ni wote tu kwenda kuwa takwimu. 330 00:18:21,580 --> 00:18:23,502 Na hiyo ina maana wakati sisi kazi nao, tutaweza 331 00:18:23,502 --> 00:18:25,376 kuwa na uwezo wa kufanya kazi na wote wakati huo huo. 332 00:18:25,376 --> 00:18:28,100 333 00:18:28,100 --> 00:18:28,600 Sawa. 334 00:18:28,600 --> 00:18:30,890 Hebu kuanza na mtindo wetu fonetiki. 335 00:18:30,890 --> 00:18:34,470 Hivyo mifano ya fonetiki kutegemea mbinu Computational 336 00:18:34,470 --> 00:18:37,320 aitwaye siri mifano Markov. 337 00:18:37,320 --> 00:18:43,050 Hizi ni mifano graphical ambayo mimi kuwa na kutambua hali ya dunia 338 00:18:43,050 --> 00:18:46,500 kama kuwa na sifa na seti ya makala. 339 00:18:46,500 --> 00:18:51,960 Na hali ambayo inaeleza sehemu moja wa utekelezaji kwamba mimi nina kushiriki katika. 340 00:18:51,960 --> 00:19:00,190 >> Hivyo kama nadhani kuhusu kufanya sauti "ma" kama mama, 341 00:19:00,190 --> 00:19:03,970 kuna tofauti vipengele na ile sauti. 342 00:19:03,970 --> 00:19:07,230 Kuna sehemu ambapo mimi kuteka katika pumzi. 343 00:19:07,230 --> 00:19:09,560 Na kisha mimi mfuko wa fedha midomo yangu. 344 00:19:09,560 --> 00:19:13,710 Na mimi unaendelea midomo yangu nyuma kidogo kidogo kufanya kwamba "ma" sauti. 345 00:19:13,710 --> 00:19:15,340 Na kisha kuna habari. 346 00:19:15,340 --> 00:19:17,020 Midomo yangu kuja mbali. 347 00:19:17,020 --> 00:19:19,030 Hewa ni kufukuzwa. 348 00:19:19,030 --> 00:19:22,650 "Ma." 349 00:19:22,650 --> 00:19:29,250 >> Wale sehemu tatu tofauti itakuwa inawakilishwa na mataifa katika graph-- hii 350 00:19:29,250 --> 00:19:33,420 mwanzo, kati, na mwisho. 351 00:19:33,420 --> 00:19:38,060 Na mimi ingekuwa mabadiliko kwamba kuruhusiwa mimi kusafiri kutoka hali moja 352 00:19:38,060 --> 00:19:42,260 hadi kingine na uwezekano fulani. 353 00:19:42,260 --> 00:19:47,250 Hivyo, kwa mfano, kwamba M sauti inaweza kuwa sana, 354 00:19:47,250 --> 00:19:51,850 mfupi sana ulaji katika beginning-- "mm" - na kisha kwa muda mrefu, 355 00:19:51,850 --> 00:19:55,640 awamu vibratory ambapo mimi nina kufanya yangu midomo pamoja na karibu humming-- 356 00:19:55,640 --> 00:20:05,090 "mmmm" - na kisha mfupi sana plosive ambapo mimi kufukuza breath-- "ma." 357 00:20:05,090 --> 00:20:09,370 >> Siri Markov mfano ni iliyoundwa na kukamata ukweli 358 00:20:09,370 --> 00:20:13,340 kwamba kwa njia hiyo mimi kufanya kwamba sauti "ma" ni kwenda 359 00:20:13,340 --> 00:20:17,350 kuwa tofauti kidogo katika majira yake, ni mzunguko, 360 00:20:17,350 --> 00:20:21,030 na sifa yake ya njia hiyo, unaweza kufanya hivyo 361 00:20:21,030 --> 00:20:23,300 au njia ili nipate kufanya hivyo wakati mimi kuzungumza 362 00:20:23,300 --> 00:20:26,030 kuhusu matumizi mbalimbali ya barua. 363 00:20:26,030 --> 00:20:33,240 "Mama" na "naomba" mapenzi sauti tofauti kidogo. 364 00:20:33,240 --> 00:20:36,800 >> Hivyo kutambua sauti fulani, tunataka 365 00:20:36,800 --> 00:20:42,020 kujenga Markov mifano, haya Markov siri mifano, ya kila simu inawezekana kwamba mimi 366 00:20:42,020 --> 00:20:45,840 kutaka kutambua, kila sauti inawezekana, 367 00:20:45,840 --> 00:20:49,750 na kisha kuangalia data akustisk kwamba nina 368 00:20:49,750 --> 00:20:54,430 na kuamua takwimu ambayo moja ni moja na uwezo mkubwa 369 00:20:54,430 --> 00:20:58,110 kwa zinazozalishwa sauti hii. 370 00:20:58,110 --> 00:20:58,610 SAWA. 371 00:20:58,610 --> 00:21:01,540 372 00:21:01,540 --> 00:21:06,750 Kwa mfano kwamba, sisi kisha kuanza kujenga juu yake. 373 00:21:06,750 --> 00:21:09,330 Tunachukua matamshi mfano wa kuigwa. 374 00:21:09,330 --> 00:21:11,790 Sasa, wakati mwingine matamshi mifano ni rahisi na rahisi 375 00:21:11,790 --> 00:21:14,440 kwa sababu kuna moja tu njia ya kutamka kitu. 376 00:21:14,440 --> 00:21:17,990 Wakati mwingine, wao uko kidogo ngumu zaidi. 377 00:21:17,990 --> 00:21:21,340 Hapa ni matamshi kuongoza kwa kuwa jambo nyekundu kwamba ni 378 00:21:21,340 --> 00:21:25,210 matunda kwamba kufanya ketchup nje ya. 379 00:21:25,210 --> 00:21:27,360 Watu sidhani ni matunda. 380 00:21:27,360 --> 00:21:27,860 Sawa? 381 00:21:27,860 --> 00:21:30,880 382 00:21:30,880 --> 00:21:35,300 >> Sasa, kuna njia nyingi tofauti kwamba watu kutamka neno hili. 383 00:21:35,300 --> 00:21:37,780 Baadhi watasema "toe-may-toe." 384 00:21:37,780 --> 00:21:40,880 Baadhi watasema "toe-mah-toe." 385 00:21:40,880 --> 00:21:44,800 Na tunaweza kukamata kwamba kwa moja ya mifano ya haya graphical 386 00:21:44,800 --> 00:21:48,305 ambapo, tena, sisi kuwakilisha mabadiliko kama kuwa na uwezekano fulani 387 00:21:48,305 --> 00:21:51,360 na kuhusishwa na uwezekano pamoja nao. 388 00:21:51,360 --> 00:21:58,290 >> Hivyo katika kesi hii, kama ningekuwa kufuata Njia juu kupitia graph hii nzima, 389 00:21:58,290 --> 00:22:03,330 Mimi itakuwa kuanzia saa barua juu ya mbali kushoto, "ta" sauti. 390 00:22:03,330 --> 00:22:07,570 Napenda kuchukua nusu juu, "loo," na kisha "ma," 391 00:22:07,570 --> 00:22:14,530 na kisha "," na kisha "ta," na "loo." "Toe-may-toe." 392 00:22:14,530 --> 00:22:19,610 Kama mimi alichukua njia chini kupitia hivyo, mimi kupata "ta-mah-toe." 393 00:22:19,610 --> 00:22:26,810 Na kama nilikwenda chini na kisha up, napenda kupata "ta-may-toe." 394 00:22:26,810 --> 00:22:29,950 >> Mifano ya haya kukamata hawa tofauti kwa sababu wakati wowote 395 00:22:29,950 --> 00:22:32,410 sisi kupeleka mmoja wa haya mifumo ya kutambua, 396 00:22:32,410 --> 00:22:35,340 itakuja na kufanya kazi na kura ya aina tofauti ya watu, 397 00:22:35,340 --> 00:22:39,295 kura ya accents mbalimbali, na hata matumizi mbalimbali ya maneno yaleyale. 398 00:22:39,295 --> 00:22:42,204 399 00:22:42,204 --> 00:22:44,120 Hatimaye, juu ya kwamba, tutaweza kujenga kitu 400 00:22:44,120 --> 00:22:48,780 kwamba inaonekana kweli ngumu, aitwaye mfano lugha, 401 00:22:48,780 --> 00:22:52,950 lakini kwa kweli ni rahisi ya tatu kwa sababu hizi kazi 402 00:22:52,950 --> 00:22:56,041 juu ya kile kinachojulikana mifano n-gram. 403 00:22:56,041 --> 00:23:02,270 Na katika kesi hii, mimi nina kuonyesha wewe sehemu mbili mfano n-gram, bigram. 404 00:23:02,270 --> 00:23:08,910 Tunakwenda kufanya kimwili wazo kwamba wakati mwingine, maneno fulani ni 405 00:23:08,910 --> 00:23:14,680 zaidi uwezekano wa kufuata kutokana na neno kuliko wengine. 406 00:23:14,680 --> 00:23:25,210 Kama mimi tu alisema "hali ya hewa," neno ijayo inaweza uwezekano kuwa "leo" 407 00:23:25,210 --> 00:23:31,510 au inaweza kuwa "hali ya hewa utabiri kesho. " 408 00:23:31,510 --> 00:23:38,870 Lakini ni uwezekano wa kuwa na " hali ya hewa utabiri artichoke. " 409 00:23:38,870 --> 00:23:42,980 >> Nini mtindo wa lugha gani ni captures wale takwimu 410 00:23:42,980 --> 00:23:47,450 kwa kuhesabu, kutoka kwa baadhi kubwa sana corpus, wote wa matukio 411 00:23:47,450 --> 00:23:50,890 ambao neno moja ifuatavyo mwingine. 412 00:23:50,890 --> 00:23:54,300 Hivyo kama mimi kuchukua corpus-- kubwa kama kila Wall Street Journal 413 00:23:54,300 --> 00:24:00,750 ambayo imekuwa zinazozalishwa tangu 1930, ambayo ni moja ya corpuses-- kiwango 414 00:24:00,750 --> 00:24:03,910 na mimi kuangalia njia zote Nakala kwamba, na mimi kuhesabu 415 00:24:03,910 --> 00:24:09,770 up mara ngapi baada ya "Utabiri" kufanya naona "leo" 416 00:24:09,770 --> 00:24:17,454 na ni mara ngapi kufanya naona "Utabiri" ikifuatiwa na "artichoke," 417 00:24:17,454 --> 00:24:19,370 moja ya kwanza ni kwenda kuwa mkubwa zaidi. 418 00:24:19,370 --> 00:24:21,540 Ni kwenda itaonekana mbali zaidi mara kwa mara. 419 00:24:21,540 --> 00:24:24,610 Na hivyo itabidi juu uwezekano yanayohusiana na hayo. 420 00:24:24,610 --> 00:24:27,340 >> Kama mimi nataka kufikiri uwezekano wa kutamka nzima, 421 00:24:27,340 --> 00:24:29,940 basi, mimi tu kuvunja up. 422 00:24:29,940 --> 00:24:35,990 Hivyo uwezekano wa kusikia hukumu "panya wakala jibini" 423 00:24:35,990 --> 00:24:39,110 ni uwezekano wa neno "" kuanzia hukumu, 424 00:24:39,110 --> 00:24:42,540 na kisha uwezekano kwamba neno "panya" ifuatavyo neno "," 425 00:24:42,540 --> 00:24:44,910 na uwezekano kwamba neno "wakala" ifuatavyo "panya," 426 00:24:44,910 --> 00:24:51,120 na uwezekano kwamba "cheese" ifuatavyo "wakala". 427 00:24:51,120 --> 00:24:55,160 >> Hii inaonekana kama mengi ya takwimu, mengi ya probabilities. 428 00:24:55,160 --> 00:24:57,510 Na kwamba wote kwamba ni. 429 00:24:57,510 --> 00:25:02,920 Lakini jambo la kushangaza ni kama wewe kufanya hivyo na kubwa sampuli kutosha wa takwimu, 430 00:25:02,920 --> 00:25:03,670 inafanya kazi. 431 00:25:03,670 --> 00:25:05,250 Na ni kazi kwa kiasi kikubwa vizuri. 432 00:25:05,250 --> 00:25:07,810 433 00:25:07,810 --> 00:25:11,420 Sote tunajua teknolojia hizi. 434 00:25:11,420 --> 00:25:16,500 Mifumo ya uendeshaji kuja na kutambua sauti katika hatua hii. 435 00:25:16,500 --> 00:25:20,940 Sisi kutumia Siri na Cortana na Echo. 436 00:25:20,940 --> 00:25:25,070 Na mambo haya ni ya msingi juu aina hii ya tatu-safu model-- 437 00:25:25,070 --> 00:25:30,620 mfano fonetiki chini, a matamshi mfano katikati, 438 00:25:30,620 --> 00:25:33,690 na mfano wa kuigwa lugha juu yao. 439 00:25:33,690 --> 00:25:37,630 >> Sasa, wao kufanya kidogo zaidi kuliko ile ili kujibu maswali. 440 00:25:37,630 --> 00:25:43,000 Lakini utambuzi wa nini wewe akisema inategemea hasa juu ya jambo hilo. 441 00:25:43,000 --> 00:25:45,700 Basi hebu chukua mfano hapa. 442 00:25:45,700 --> 00:25:52,020 Hivyo nina simu yangu ameketi juu hapa chini ya hati kamera. 443 00:25:52,020 --> 00:25:56,110 Na tunakwenda anauliza Siri maswali kadhaa. 444 00:25:56,110 --> 00:25:57,150 Sawa? 445 00:25:57,150 --> 00:25:59,940 >> Basi hebu kuamka simu yangu hapa. 446 00:25:59,940 --> 00:26:02,710 447 00:26:02,710 --> 00:26:05,000 Siri, ni nini hali ya hewa kama katika New Haven leo? 448 00:26:05,000 --> 00:26:07,670 449 00:26:07,670 --> 00:26:10,780 >> Siri: Hapa ni ya hali ya hewa kwa New Haven, Connecticut leo. 450 00:26:10,780 --> 00:26:11,890 >> BRIAN SCASSELLATI: Sawa. 451 00:26:11,890 --> 00:26:16,720 Hivyo kwanza ninyi kuona kwamba Siri kutambuliwa kila moja ya maneno ya mtu binafsi 452 00:26:16,720 --> 00:26:19,050 na kisha zinazozalishwa majibu. 453 00:26:19,050 --> 00:26:22,277 Tutaweza majadiliano juu ya jinsi jibu kwamba huja juu katika kidogo. 454 00:26:22,277 --> 00:26:24,110 Lakini sasa kwa kuwa tunajua kwamba hii ni msingi tu 455 00:26:24,110 --> 00:26:28,880 juu ya takwimu ghafi na hii muundo vinavyolingana aina ya mbinu, 456 00:26:28,880 --> 00:26:31,120 tunaweza kucheza baadhi ya michezo na Siri. 457 00:26:31,120 --> 00:26:34,560 >> Hivyo siwezi kujaribu tena. 458 00:26:34,560 --> 00:26:38,864 Siri, ni nini hali ya hewa kiboko New Haven, leo? 459 00:26:38,864 --> 00:26:39,810 >> Siri: Sawa. 460 00:26:39,810 --> 00:26:44,245 Hapa ni ya hali ya hewa kwa ajili ya New Haven, Connecticut kwa leo. 461 00:26:44,245 --> 00:26:46,120 BRIAN SCASSELLATI: Siri ya si daunted na kwamba 462 00:26:46,120 --> 00:26:50,980 kwa sababu ni kupatikana pattern-- "hali ya hewa," "leo," "New Haven." 463 00:26:50,980 --> 00:26:54,420 Hiyo ni nini ni kukabiliana kwa, tu kama ELIZA. 464 00:26:54,420 --> 00:26:54,920 Sawa. 465 00:26:54,920 --> 00:26:59,390 Hebu kuwapa moja zaidi hata zaidi ujinga mfano. 466 00:26:59,390 --> 00:27:03,075 Siri, hali ya hewa artichoke kakakuona kiboko New Haven? 467 00:27:03,075 --> 00:27:06,806 468 00:27:06,806 --> 00:27:08,400 >> Siri: Hebu kuangalia juu ya jambo hilo. 469 00:27:08,400 --> 00:27:11,280 Hapa ni nini nimeona kwenye mtandao kwa nini ni aubergine kakakuona 470 00:27:11,280 --> 00:27:13,780 kiboko New Haven. 471 00:27:13,780 --> 00:27:14,760 >> BRIAN SCASSELLATI: Sawa. 472 00:27:14,760 --> 00:27:20,400 Hivyo kama mimi kwenda mbali kutosha mbali na mtindo huu, 473 00:27:20,400 --> 00:27:24,365 Nina uwezo wa kuchanganya hili kwa sababu hakuna tena mechi mfano kwamba ina. 474 00:27:24,365 --> 00:27:27,370 475 00:27:27,370 --> 00:27:29,390 Na kwamba takwimu injini hiyo akisema, 476 00:27:29,390 --> 00:27:32,850 nini uwezekano kwamba nimepata Maneno kiboko na artichoke 477 00:27:32,850 --> 00:27:34,440 pamoja, na Kakakuona? 478 00:27:34,440 --> 00:27:36,050 Hiyo got kuwa kitu kipya. 479 00:27:36,050 --> 00:27:38,840 480 00:27:38,840 --> 00:27:40,610 >> Hivyo teknolojia hizi tunatumia kila siku. 481 00:27:40,610 --> 00:27:43,670 482 00:27:43,670 --> 00:27:47,800 Kama tunataka kuwapeleka hatua moja zaidi, ingawa, kama sisi kweli 483 00:27:47,800 --> 00:27:53,930 wanataka kuwa na uwezo wa kuzungumza kuhusu nini ni kwamba mifumo hii ni kukabiliana na, 484 00:27:53,930 --> 00:28:00,630 tuna kuzungumza, tena, kuhusu zaidi ya msingi seti ya maswali. 485 00:28:00,630 --> 00:28:05,370 Na hiyo ndiyo mada katika mawasiliano kwamba tunatoa wito swali kujibu. 486 00:28:05,370 --> 00:28:07,028 Yaani, tunataka kuwa na uwezo to-- yeah? 487 00:28:07,028 --> 00:28:07,944 Watazamaji: [inaudible]. 488 00:28:07,944 --> 00:28:10,789 489 00:28:10,789 --> 00:28:13,330 BRIAN SCASSELLATI: Je, sisi kupata ndani ya latent semantic usindikaji? 490 00:28:13,330 --> 00:28:14,070 Hivyo ndiyo. 491 00:28:14,070 --> 00:28:17,820 Kuna mambo mengi ambayo ni kinachotokea chini ya uso na Siri 492 00:28:17,820 --> 00:28:20,210 na katika baadhi ya mifano Mimi nina kwenda kuonyesha ijayo 493 00:28:20,210 --> 00:28:22,610 ambapo kuna kidogo kabisa katika suala la muundo 494 00:28:22,610 --> 00:28:25,260 ya unachosema hiyo ni muhimu. 495 00:28:25,260 --> 00:28:31,890 Na, kwa kweli, kwamba ni kubwa mtangulizi kwa slide ijayo kwa ajili yangu. 496 00:28:31,890 --> 00:28:35,110 >> Hivyo katika njia sawa kwamba yetu kutambua hotuba ilijengwa juu 497 00:28:35,110 --> 00:28:39,620 ya tabaka mbalimbali, kama tunataka kuelewa ni nini hiyo ni kweli 498 00:28:39,620 --> 00:28:44,620 kuwa alisema, tunakwenda tena kutegemea uchambuzi tabaka mbalimbali 499 00:28:44,620 --> 00:28:47,020 Nakala ya hiyo kuwa kutambuliwa. 500 00:28:47,020 --> 00:28:52,560 Hivyo wakati Siri ni kweli na uwezo wa kusema, tazama nimeona maneno haya. 501 00:28:52,560 --> 00:28:55,230 Sasa nini mimi pamoja nao? 502 00:28:55,230 --> 00:28:59,110 Sehemu ya kwanza ni mara nyingi kwa kwenda kwa njia na kujaribu kuchambua 503 00:28:59,110 --> 00:29:03,010 muundo wa hukumu. 504 00:29:03,010 --> 00:29:05,410 Na katika kile tumeona katika shule ya daraja, mara nyingi, 505 00:29:05,410 --> 00:29:08,920 kama aina ya diagramming hukumu, tunakwenda 506 00:29:08,920 --> 00:29:12,774 kutambua kuwa fulani Maneno wana majukumu fulani. 507 00:29:12,774 --> 00:29:13,440 Hizi ni majina. 508 00:29:13,440 --> 00:29:14,231 Hizi ni viwakilishi. 509 00:29:14,231 --> 00:29:16,200 Hizi ni verbs. 510 00:29:16,200 --> 00:29:19,460 Na tunakwenda kutambua kwamba kwa sarufi maalum, 511 00:29:19,460 --> 00:29:24,700 katika hii kesi Kiingereza sarufi, kuna halali njia ambazo siwezi kuchanganya yao 512 00:29:24,700 --> 00:29:26,280 na njia nyingine ambazo si halali. 513 00:29:26,280 --> 00:29:29,920 514 00:29:29,920 --> 00:29:33,870 >> Kutambua kuwa, muundo kwamba, inaweza kuwa kutosha kusaidia kutuongoza 515 00:29:33,870 --> 00:29:36,720 kidogo. 516 00:29:36,720 --> 00:29:39,820 Lakini siyo kabisa kutosha kwa sisi kuwa na uwezo wa kutoa 517 00:29:39,820 --> 00:29:43,290 maana yoyote kwa nini kuwa alisema hapa. 518 00:29:43,290 --> 00:29:46,615 Ili kufanya hivyo, tutaweza kuwategemea baadhi ya kiasi cha semantic usindikaji. 519 00:29:46,615 --> 00:29:49,590 520 00:29:49,590 --> 00:29:55,080 Yaani, tunakwenda na kuangalia katika chini ya kile kila moja ya maneno haya 521 00:29:55,080 --> 00:29:57,400 kweli hubeba kama maana. 522 00:29:57,400 --> 00:30:01,150 Na katika njia rahisi ya kufanya hivyo, tunakwenda kujiunga na kila neno 523 00:30:01,150 --> 00:30:06,930 kwamba tunajua kazi fulani, mabadiliko fulani kwamba 524 00:30:06,930 --> 00:30:09,300 inaruhusu kutokea. 525 00:30:09,300 --> 00:30:14,470 >> Katika kesi hiyo, tupate studio neno "John" kama kuwa jina sahihi, 526 00:30:14,470 --> 00:30:18,160 kuwa ni inaleta utambulisho. 527 00:30:18,160 --> 00:30:21,530 Na tupate studio "Mariamu" kama njia hiyo hiyo. 528 00:30:21,530 --> 00:30:27,900 Wakati kitendo kama "anapenda," kwamba hufanya uhusiano fulani 529 00:30:27,900 --> 00:30:31,582 kwamba sisi ni uwezo wa kuwakilisha. 530 00:30:31,582 --> 00:30:33,290 Sasa, hiyo haina maana kwa kuwa tunafahamu 531 00:30:33,290 --> 00:30:37,680 upendo ni nini lakini tu kwa kuwa tunafahamu hivyo kwa njia ya mfumo wa mfano. 532 00:30:37,680 --> 00:30:40,480 Hiyo ni, tunaweza studio hilo na kuendesha hiyo. 533 00:30:40,480 --> 00:30:44,230 534 00:30:44,230 --> 00:30:49,120 >> Kwa kila moja ya aina hii ya mbinu, aina yoyote ya semantic usindikaji 535 00:30:49,120 --> 00:30:57,060 hapa ni kwenda zinahitaji kidogo kidogo ya elimu na kazi nyingi za 536 00:30:57,060 --> 00:30:59,020 kwa upande wetu. 537 00:30:59,020 --> 00:31:03,590 Tuko tena katika ulimwengu wa ambapo takwimu wazi tu 538 00:31:03,590 --> 00:31:07,320 ni kwenda kuwa kutosha kwa ajili yetu. 539 00:31:07,320 --> 00:31:11,330 Sasa, ili kwenda kutoka hatua hii ya kuwa 540 00:31:11,330 --> 00:31:15,520 uwezo wa kuzungumzia ndani ya nini hasa kinachotokea hapa, 541 00:31:15,520 --> 00:31:19,640 kwa kuwa na uwezo wa kuendesha hii muundo na kuelewa swali 542 00:31:19,640 --> 00:31:23,160 na kisha kuwa na uwezo kwenda nje na kutafuta, 543 00:31:23,160 --> 00:31:27,290 ambayo inahitaji zaidi tata utambuzi mfano wa kuigwa. 544 00:31:27,290 --> 00:31:34,880 >> Njia ambayo mifumo hii ni kujengwa ni kwa sehemu kubwa sana, kazi sana 545 00:31:34,880 --> 00:31:36,350 kubwa. 546 00:31:36,350 --> 00:31:39,490 Wao kuhusisha binadamu matumizi ya mpango mkubwa 547 00:31:39,490 --> 00:31:44,100 muda jinsia njia katika ambayo aina hii ya hukumu 548 00:31:44,100 --> 00:31:47,270 inaweza kuwakilishwa katika baadhi ya mantiki. 549 00:31:47,270 --> 00:31:51,639 550 00:31:51,639 --> 00:31:53,430 Anapata hata kidogo ngumu zaidi, ingawa. 551 00:31:53,430 --> 00:31:56,400 552 00:31:56,400 --> 00:31:59,660 >> Hata mara moja tumekuwa kushughulikiwa na semantiki, tutaweza 553 00:31:59,660 --> 00:32:03,860 bado una kuangalia pragmatics ya nini kuwa alisema. 554 00:32:03,860 --> 00:32:08,620 Yaani, jinsi gani mimi yanahusiana maneno kwamba mimi kuwa na kitu kimwili nje 555 00:32:08,620 --> 00:32:12,054 kuna katika dunia au katika angalau baadhi chanzo maelezo 556 00:32:12,054 --> 00:32:12,970 niweze kuendesha? 557 00:32:12,970 --> 00:32:15,780 558 00:32:15,780 --> 00:32:20,790 >> Wakati mwingine, hizi kusababisha bits ya ajabu ya utata. 559 00:32:20,790 --> 00:32:24,470 "Red-moto nyota wa kiungwana falaki." 560 00:32:24,470 --> 00:32:25,630 SAWA. 561 00:32:25,630 --> 00:32:28,540 Sasa, tunasoma kwamba kama funny aina ya kichwa cha habari 562 00:32:28,540 --> 00:32:34,690 kwamba tunataka kuona juu ya marehemu usiku TV kwa sababu hatuna kutafsiri "nyota" 563 00:32:34,690 --> 00:32:38,630 kuwa na mwili wake maana ya mbinguni. 564 00:32:38,630 --> 00:32:43,390 Tunajua kwamba maana yake ni zaidi kawaida muigizaji au mwigizaji 565 00:32:43,390 --> 00:32:45,240 kwa kiasi kikubwa cha kujulikana. 566 00:32:45,240 --> 00:32:47,770 567 00:32:47,770 --> 00:32:51,950 >> "Squad husaidia mbwa kuumwa waathirika." 568 00:32:51,950 --> 00:32:55,550 Je, ni kwamba kikosi ni kweli huko nje kusaidia mbwa 569 00:32:55,550 --> 00:32:59,620 katika kwenda kuzunguka na kuuma waathirika? 570 00:32:59,620 --> 00:33:02,380 Au ni kwamba kulikuwa na mtu ambaye alikuwa 571 00:33:02,380 --> 00:33:04,625 kuumwa na mbwa waliohitaji baadhi ya msaada? 572 00:33:04,625 --> 00:33:07,650 573 00:33:07,650 --> 00:33:11,480 Tu kutoka kuangalia syntax na semantiki ya hukumu, 574 00:33:11,480 --> 00:33:14,660 hatuwezi kuamua kwamba. 575 00:33:14,660 --> 00:33:22,000 >> "Helikopta kinatumia inzi binadamu." 576 00:33:22,000 --> 00:33:27,330 Je, kuna, kwa kweli, halisi mambo madogo kuruka karibu kwamba 577 00:33:27,330 --> 00:33:34,510 ni watu wenye mbawa driva helikopta kwa manufaa ya wanadamu? 578 00:33:34,510 --> 00:33:38,960 Au kuna kifaa kimoja mitambo kwamba hupata nguvu yake katika mtu? 579 00:33:38,960 --> 00:33:41,600 580 00:33:41,600 --> 00:33:46,500 >> Tunapoangalia swali mifumo ya kujibu, 581 00:33:46,500 --> 00:33:51,110 hizi ni tabaka kwamba tunahitaji kushughulikia. 582 00:33:51,110 --> 00:33:54,890 Sasa, bora na aina zaidi ya imefikia ya mifumo hii 583 00:33:54,890 --> 00:34:00,640 leo ni uwezo wa kufanya mambo katika ngazi hiyo ni 584 00:34:00,640 --> 00:34:05,050 mbali zaidi ya kile tulikuwa miaka 10 iliyopita. 585 00:34:05,050 --> 00:34:08,210 Hapa ni moja ya mifumo bora hiyo ni huko nje leo. 586 00:34:08,210 --> 00:34:10,519 Na hii ni mfumo kutoka IBM aitwaye Watson. 587 00:34:10,519 --> 00:34:13,040 588 00:34:13,040 --> 00:34:15,500 Hii ni kutokana na marafiki zetu katika IBM. 589 00:34:15,500 --> 00:34:19,810 Hii ni Katherine benki, Jia Chen, na Prisila Moraes anatupa 590 00:34:19,810 --> 00:34:26,323 maandamano ya Watson jana katika IBM Yorktown. 591 00:34:26,323 --> 00:34:26,989 [VIDEO avspelning] 592 00:34:26,989 --> 00:34:27,690 -Jinsi Wewe ni? 593 00:34:27,690 --> 00:34:30,600 594 00:34:30,600 --> 00:34:31,980 >> -I Am vizuri. 595 00:34:31,980 --> 00:34:35,429 Mimi ni furaha daima wakati mimi niko kufikiri, hasa wakati mimi 596 00:34:35,429 --> 00:34:38,920 wanaweza kukaa baridi katika vyumba haya server. 597 00:34:38,920 --> 00:34:40,170 -Nani Ni mwigizaji your favorite? 598 00:34:40,170 --> 00:34:43,870 599 00:34:43,870 --> 00:34:47,555 >> Favorite mwigizaji -my Angelina Jolie kama Lara Croft 600 00:34:47,555 --> 00:34:54,800 katika Tomb Raider kwa sababu awali yake fomu ilikuwa kama tabia virtual kama mimi. 601 00:34:54,800 --> 00:34:56,760 >> -Jinsi Watson inaweza kusaidia wagonjwa? 602 00:34:56,760 --> 00:34:59,700 603 00:34:59,700 --> 00:35:03,755 >> -Je Mpango kwa ajili ya wagonjwa kwa wanapata Watson moja kwa moja? 604 00:35:03,755 --> 00:35:08,050 Upatikanaji wa Watson ushauri maombi itakuwa kutofautiana na matumizi ya kesi. 605 00:35:08,050 --> 00:35:12,630 Lakini sasa, Watson tu kutoa utambuzi iwezekanavyo na tiba 606 00:35:12,630 --> 00:35:14,725 chaguzi mtaalamu wa matibabu. 607 00:35:14,725 --> 00:35:17,680 Pale inapobidi, wagonjwa inaweza kutoa taarifa 608 00:35:17,680 --> 00:35:21,210 kwa Watson, labda kupitia portal mtandao interface, kibao, 609 00:35:21,210 --> 00:35:23,500 au hata simu za mkononi. 610 00:35:23,500 --> 00:35:27,180 Kwa mfano, mgonjwa inaweza kuwa na aliuliza kuelezea dalili zao 611 00:35:27,180 --> 00:35:30,460 katika lugha ya asili moja kwa moja katika ufumbuzi Watson, 612 00:35:30,460 --> 00:35:34,050 kuruhusu daktari kwa kuzingatia mkusanyiko utambuzi na si data. 613 00:35:34,050 --> 00:35:38,379 614 00:35:38,379 --> 00:35:42,716 >> -Nani Aliuawa Abraham Lincoln? 615 00:35:42,716 --> 00:35:46,153 >> Yohana Wilkes Booth. 616 00:35:46,153 --> 00:35:52,373 >> -Katika Mwaka gani Arizona Diamondbacks kushinda World Series? 617 00:35:52,373 --> 00:35:55,295 >> -2001. 618 00:35:55,295 --> 00:35:58,710 >> [Mwisho avspelning] 619 00:35:58,710 --> 00:36:01,610 >> BRIAN SCASSELLATI: Kwa hiyo aina hii ya mifumo ya 620 00:36:01,610 --> 00:36:07,430 na kwa kutegemea juu ya kwanza ya yote kutambua hotuba; pili, 621 00:36:07,430 --> 00:36:12,200 kuwabadili ndani ya ndani ya maana uwakilishi; na kisha, tatu, 622 00:36:12,200 --> 00:36:17,090 kuwa na uwezo wa kwenda nje na kupata Maelezo chanzo hicho 623 00:36:17,090 --> 00:36:21,140 inaruhusu yao ya kujibu swali hilo. 624 00:36:21,140 --> 00:36:27,320 Kiwango hiki cha utata inahusisha aina hiyo ya mambo ya programu 625 00:36:27,320 --> 00:36:31,790 kwamba umekuwa kufanya katika seti tatizo. 626 00:36:31,790 --> 00:36:38,000 >> Sisi ni uwezo wa kuchanganua maombi HTTP katika aina moja ya ngazi ya chini mfano 627 00:36:38,000 --> 00:36:40,810 vinavyolingana kwamba ELIZA wanaweza kufanya. 628 00:36:40,810 --> 00:36:45,070 Sisi ni uwezo wa kubadilisha wale ndani ya uwakilishi wa ndani, 629 00:36:45,070 --> 00:36:50,360 na kisha kuzitumia na swala baadhi Mbegu za nje, pengine kwa kutumia SQL. 630 00:36:50,360 --> 00:36:53,530 631 00:36:53,530 --> 00:36:56,260 Wote wa mifumo ya kwamba zinajengwa leo 632 00:36:56,260 --> 00:37:00,520 kufanya aina hii ya asili mawasiliano lugha 633 00:37:00,520 --> 00:37:04,100 zinajengwa juu ya sawa kanuni hizi. 634 00:37:04,100 --> 00:37:09,530 >> Sasa, hata kama mfumo Watson si tata kutosha 635 00:37:09,530 --> 00:37:14,820 kuwa na uwezo wa kujibu holela maswali kuhusu mada yoyote. 636 00:37:14,820 --> 00:37:20,060 Na kwa kweli, wao kuwa muundo ndani ya uwanja huo. 637 00:37:20,060 --> 00:37:24,440 Hivyo unaweza kwenda online na unaweza kupata matoleo ya Watson kwamba kazi vizuri 638 00:37:24,440 --> 00:37:27,700 ndani ya Informatics matibabu. 639 00:37:27,700 --> 00:37:31,490 Au kuna moja online kwamba tu inahusika na jinsi 640 00:37:31,490 --> 00:37:34,540 kutoa mapendekezo mazuri kuhusu nini bia utakwenda pamoja ambayo chakula. 641 00:37:34,540 --> 00:37:37,060 642 00:37:37,060 --> 00:37:41,870 Na ndani ya nyanja hizo, inaweza kujibu maswali, 643 00:37:41,870 --> 00:37:46,130 kupata taarifa kwamba inahitaji. 644 00:37:46,130 --> 00:37:48,270 >> Lakini huwezi kuchanganya na mechi yao. 645 00:37:48,270 --> 00:37:53,150 Mfumo kwamba imekuwa mafunzo na database ya chakula na bia 646 00:37:53,150 --> 00:37:56,830 haifanyi kazi vizuri wakati wewe ghafla kuiweka katika na Informatics matibabu 647 00:37:56,830 --> 00:37:59,770 database. 648 00:37:59,770 --> 00:38:05,680 Hivyo hata mifumo yetu bora leo kutegemea juu ya kiwango cha usindikaji 649 00:38:05,680 --> 00:38:11,570 ambayo sisi ni mkono coding na kujenga katika miundombinu ili 650 00:38:11,570 --> 00:38:13,275 kufanya mfumo huu kukimbia. 651 00:38:13,275 --> 00:38:16,360 652 00:38:16,360 --> 00:38:20,710 >> Sasa, mada mwisho nataka kuwa na uwezo wa kupata leo 653 00:38:20,710 --> 00:38:23,960 ni kuhusu mawasiliano kwa ishara. 654 00:38:23,960 --> 00:38:29,290 Umati mkubwa wa habari kwamba sisi kuwasiliana na wengine 655 00:38:29,290 --> 00:38:35,490 haina kuja juu kupitia maneno ya mtu binafsi kwamba sisi ni kuomba. 656 00:38:35,490 --> 00:38:40,290 Ni ina nini na mambo kama ukaribu, macho, sauti yako, 657 00:38:40,290 --> 00:38:42,270 inflection yako. 658 00:38:42,270 --> 00:38:46,620 Na kwamba mawasiliano pia ni jambo ambalo wengi interfaces tofauti 659 00:38:46,620 --> 00:38:49,960 huduma mengi juu. 660 00:38:49,960 --> 00:38:51,500 Ni nini Siri anayejali kuhusu. 661 00:38:51,500 --> 00:38:56,250 Siwezi kuuliza Siri kitu kwa sauti moja au kwa sauti mbalimbali za sauti, 662 00:38:56,250 --> 00:38:59,840 na Siri kinaendelea nipe jibu moja. 663 00:38:59,840 --> 00:39:05,260 Lakini si kwamba kile sisi kujenga kwa wengine wengi aina ya interfaces. 664 00:39:05,260 --> 00:39:09,120 >> Mimi nataka kuanzisha wewe sasa na mmoja wa robots. 665 00:39:09,120 --> 00:39:12,720 Hii ilijengwa na wa muda mrefu yangu rafiki na mwenzake Cynthia 666 00:39:12,720 --> 00:39:16,010 Breazeal na kampuni yake Jibo. 667 00:39:16,010 --> 00:39:20,090 Na robot-- hii tunakwenda kuwa na kujitolea wanandoa 668 00:39:20,090 --> 00:39:22,520 kuja kuingiliana na hii. 669 00:39:22,520 --> 00:39:26,200 Hivyo naweza kuwa na watu wawili tayari kucheza na robot kwa ajili yangu? 670 00:39:26,200 --> 00:39:29,936 Mbona wewe kuja juu juu, na kwa nini wewe kuja juu juu. 671 00:39:29,936 --> 00:39:31,310 Kama utaweza kujiunga na mimi hapa, tafadhali. 672 00:39:31,310 --> 00:39:36,520 673 00:39:36,520 --> 00:39:39,670 >> Na kama mimi naweza kuwa na wewe kuja haki zaidi ya hapa. 674 00:39:39,670 --> 00:39:40,170 Shukrani. 675 00:39:40,170 --> 00:39:40,480 Hi. 676 00:39:40,480 --> 00:39:41,400 >> ALFREDO: Nice kukutana na wewe. 677 00:39:41,400 --> 00:39:42,010 Alfredo. 678 00:39:42,010 --> 00:39:42,520 >> BRIAN SCASSELLATI: Alfredo. 679 00:39:42,520 --> 00:39:43,146 >> RACHEL: Rachel. 680 00:39:43,146 --> 00:39:44,228 BRIAN SCASSELLATI: Rachel. 681 00:39:44,228 --> 00:39:45,154 Vyema kukutana na wewe wote. 682 00:39:45,154 --> 00:39:46,820 Alfredo, mimi nina kwenda na kwenda kwanza. 683 00:39:46,820 --> 00:39:47,990 Kuja haki juu hapa. 684 00:39:47,990 --> 00:39:51,870 Mimi nina kwenda kuanzisha you-- kama naweza kupata hii mbali 685 00:39:51,870 --> 00:39:58,450 bila kugonga microphone-- kwa robot mdogo aitwaye Jibo. 686 00:39:58,450 --> 00:40:00,140 SAWA? 687 00:40:00,140 --> 00:40:04,260 >> Sasa, Jibo imeundwa kuwa shirikishi. 688 00:40:04,260 --> 00:40:09,339 Na ingawa inaweza kutoa hotuba wewe, mengi ya mwingiliano na robot 689 00:40:09,339 --> 00:40:09,880 ni mawasiliano kwa ishara. 690 00:40:09,880 --> 00:40:12,450 691 00:40:12,450 --> 00:40:17,070 Alfredo, mimi nina kwenda kuuliza wewe kusema kitu kizuri na complimentary 692 00:40:17,070 --> 00:40:19,554 kwa robot, tafadhali. 693 00:40:19,554 --> 00:40:20,845 ALFREDO: Nadhani ukiangalia cute. 694 00:40:20,845 --> 00:40:24,114 695 00:40:24,114 --> 00:40:25,611 >> [Whirring SOUND] 696 00:40:25,611 --> 00:40:26,192 697 00:40:26,192 --> 00:40:27,108 BRIAN SCASSELLATI: Sawa. 698 00:40:27,108 --> 00:40:30,110 699 00:40:30,110 --> 00:40:33,180 Majibu yake ni si matusi. 700 00:40:33,180 --> 00:40:35,180 Na hata hivyo, aliwapa wote kukiri wazi 701 00:40:35,180 --> 00:40:39,680 kuwa ni wamesikia nini alisema na pia kwa namna fulani kuelewa kwamba. 702 00:40:39,680 --> 00:40:40,530 SAWA? 703 00:40:40,530 --> 00:40:42,070 Hatua haki ya nyuma hapa kwa ajili ya pili moja. 704 00:40:42,070 --> 00:40:43,130 Asante. 705 00:40:43,130 --> 00:40:44,090 >> Rachel, kama wewe ingekuwa. 706 00:40:44,090 --> 00:40:46,070 Sasa, mimi nina kwenda kutoa wewe kazi ngumu sana. 707 00:40:46,070 --> 00:40:48,361 Kama utaweza kusimama hapa hapa, kuunga kidogo tu hivyo 708 00:40:48,361 --> 00:40:50,280 tunaweza kupata wewe kwenye kamera na kuangalia njia hii. 709 00:40:50,280 --> 00:40:56,840 Mimi nina kwenda kuuliza wewe kusema kitu kweli maana na nasty kwa robot. 710 00:40:56,840 --> 00:41:02,900 >> RACHEL: Nini tu walionekana kufanya ni ujinga kabisa. 711 00:41:02,900 --> 00:41:03,840 >> [Humming SOUND] 712 00:41:03,840 --> 00:41:07,610 713 00:41:07,610 --> 00:41:09,030 >> Hiyo ilikuwa ni hata zaidi ya ajabu. 714 00:41:09,030 --> 00:41:10,120 Nini kinaendelea na wewe? 715 00:41:10,120 --> 00:41:13,487 716 00:41:13,487 --> 00:41:16,207 Aw, si kujisikia vibaya. 717 00:41:16,207 --> 00:41:17,040 Mimi nitakupa kumkumbatia. 718 00:41:17,040 --> 00:41:19,882 719 00:41:19,882 --> 00:41:21,090 BRIAN SCASSELLATI: zote haki. 720 00:41:21,090 --> 00:41:22,280 Shukrani, Rachel. 721 00:41:22,280 --> 00:41:24,565 Alfredo, Rachel, shukrani guys sana. 722 00:41:24,565 --> 00:41:26,840 >> [Makofi] 723 00:41:26,840 --> 00:41:28,660 724 00:41:28,660 --> 00:41:34,470 >> Hivyo aina hii ya mwingiliano ina katika njia nyingi baadhi ya sheria hiyo 725 00:41:34,470 --> 00:41:36,950 na baadhi ya sawa muundo kama kile sisi 726 00:41:36,950 --> 00:41:39,950 wanaweza kuwa katika mahusiano lugha. 727 00:41:39,950 --> 00:41:44,530 Hivyo ni wa mawasiliano na mtumishi lengo muhimu. 728 00:41:44,530 --> 00:41:48,590 Na mwingiliano kwamba, katika njia nyingi, imeundwa 729 00:41:48,590 --> 00:41:52,890 kuwa na athari hasa juu ya mtu mazungumzo na au kusikiliza 730 00:41:52,890 --> 00:41:54,410 kwa robot. 731 00:41:54,410 --> 00:41:56,450 >> Sasa, mimi nina bahati kuwa na Jibo hapa leo. 732 00:41:56,450 --> 00:42:00,550 Sam Spaulding ni hapa kusaidia sisi nje na robot. 733 00:42:00,550 --> 00:42:07,470 Na mimi nina kwenda kuuliza Sam kutoa sisi moja nzuri demo ya Jibo kucheza 734 00:42:07,470 --> 00:42:09,720 tuweze kuangalia mwishoni hapa. 735 00:42:09,720 --> 00:42:10,590 Hivyo kwenda mbele, Jibo. 736 00:42:10,590 --> 00:42:11,550 >> SAM: Sawa, Jibo. 737 00:42:11,550 --> 00:42:14,430 Tuonyeshe ngoma hatua yako. 738 00:42:14,430 --> 00:42:17,310 >> [Music kucheza] 739 00:42:17,310 --> 00:42:43,114 740 00:42:43,114 --> 00:42:44,780 BRIAN SCASSELLATI: zote haki, kila mtu. 741 00:42:44,780 --> 00:42:46,865 Shukrani kwa marafiki zetu katika Jibo. 742 00:42:46,865 --> 00:42:49,426 >> [Makofi] 743 00:42:49,426 --> 00:42:50,140 744 00:42:50,140 --> 00:42:54,990 >> Na shukrani kwa marafiki zetu katika IBM kwa ajili ya kusaidia nje leo. 745 00:42:54,990 --> 00:42:57,300 Mawasiliano ni kitu kwamba wewe ni kwenda 746 00:42:57,300 --> 00:43:02,280 kuona kuja juu zaidi na zaidi kama tunajenga interfaces ngumu zaidi. 747 00:43:02,280 --> 00:43:05,760 Wiki ijayo, tutaweza kuwa na kuzungumza kuhusu jinsi ya interface 748 00:43:05,760 --> 00:43:08,890 na wapinzani kompyuta katika michezo. 749 00:43:08,890 --> 00:43:12,950 Lakini kama una maswali kuhusu hili, Mimi itakuwa karibu katika masaa ya ofisi usiku wa leo. 750 00:43:12,950 --> 00:43:17,610 Nina furaha ya kuzungumza na wewe kuhusu ai mada au kupata katika undani zaidi. 751 00:43:17,610 --> 00:43:18,927 Na mwishoni mwa wiki kubwa. 752 00:43:18,927 --> 00:43:21,409 >> [Makofi] 753 00:43:21,409 --> 00:43:21,909 754 00:43:21,909 --> 00:43:26,141 [Music kucheza] 755 00:43:26,141 --> 00:46:42,879