ʻAʻole nahu ʻo Watson i ke kauka, a maikaʻi loa

ʻOiai, e like me ka nui o nā kula ʻē aʻe, ua emi iki ka hoihoi e hoʻololi i nā kauka me AI ma hope o ke ʻano o nā hemahema diagnostic, ke hoʻomau nei ka hana ma ka hoʻomohala ʻana i ka lāʻau lapaʻau AI. No ka mea, eia naʻe, hāʻawi lākou i nā manawa kūpono a me kahi manawa e hoʻomaikaʻi ai i ka pono o nā hana ma nā wahi he nui.

Ua hoʻolaha ʻia ʻo IBM i ka makahiki 2015, a i ka makahiki 2016 ua loaʻa iā ia ke komo i ka ʻikepili mai nā hui ʻikepili maʻi nui ʻehā (1). ʻO ka mea kaulana loa, mahalo i ka nui o nā hōʻike media, a ma ka manawa like ʻo ka papahana makemake nui e hoʻohana ana i ka naʻauao artificial kiʻekiʻe mai IBM e pili ana i ka oncology. Ke hoʻāʻo nei nā kānaka ʻepekema e hoʻohana i nā kumuwaiwai nui o ka ʻikepili e hoʻoponopono iā lākou i mea e hoʻohuli ai iā lākou i mau lāʻau lapaʻau anti-cancer. ʻO ka pahuhopu lōʻihi ka loaʻa ʻana o Watson i luna hoʻoponopono hoʻokolohua hoʻokolohua a me nā hopena e like me ke kauka.

1. ʻO kekahi o nā hiʻohiʻona o ka ʻōnaehana olakino Watson Health

Eia naʻe, ua ʻike ʻia kēlā Watson ʻAʻole hiki ke kuhikuhi kūʻokoʻa i nā palapala lapaʻau, a ʻaʻole hiki ke unuhi i ka ʻike mai nā moʻolelo olakino uila o nā maʻi. Eia naʻe, ʻo ka hoʻopiʻi koʻikoʻi loa iā ia ʻaʻole hiki ke hoʻohālikelike maikaʻi i kahi maʻi hou me nā maʻi maʻi maʻi ʻelemakule ʻē aʻe a ʻike i nā hōʻailona ʻike ʻole ʻia i ka nānā mua ʻana.

ʻOiaʻiʻo, aia kekahi poʻe oncologists i ʻōlelo he hilinaʻi lākou i kāna hoʻoholo ʻana, ʻoiai ʻo ka hapa nui o nā manaʻo o Watson no nā lāʻau lapaʻau maʻamau, a i ʻole he manaʻo lapaʻau hou. He nui ka poʻe i kuhikuhi e lilo kēia ʻōnaehana i hale waihona puke maʻalahi no nā kauka.

Ma muli o nā manaʻo hoʻohiwahiwa ʻole mai IBM nā pilikia me ke kūʻai aku ʻana o ka ʻōnaehana Watson ma nā keʻena olakino US. Ua kūʻai aku nā ʻelele kūʻai IBM i kekahi mau halemai ma India, South Korea, Thailand a me nā ʻāina ʻē aʻe. Ma India, loiloi nā kauka () i nā ʻōlelo aʻoaʻo a Watson no 638 mau hihia o ka maʻi maʻi umauma. ʻO ka helu hoʻokō no nā ʻōlelo aʻoaʻo lapaʻau he 73%. ʻOi aku ka ʻino Watson Ua hoʻokuʻu ʻia ma ka Hale Hōʻikeʻike ʻo Gachon ma South Korea, kahi o kāna mau ʻōlelo aʻoaʻo maikaʻi loa no nā maʻi maʻi maʻi colorectal 656 i kūlike i nā ʻōlelo aʻoaʻo loea wale nō 49 pakeneka o ka manawa. Ua loiloi nā kauka i kēlā ʻAʻole maikaʻi ʻo Watson me nā poʻe maʻi kahikoma ka hāʻawi ʻole ʻana iā lākou i kekahi mau lāʻau lapaʻau maʻamau, a ua hana hewa nui i ka hana ʻana i ka nānā ʻana i ka mālama koʻikoʻi no kekahi mau maʻi me ka maʻi metastatic.

ʻO ka hope loa, ʻoiai ʻo kāna hana ma ke ʻano he diagnostician a me ke kauka i manaʻo ʻia he kūleʻa ʻole, aia kekahi mau wahi i hōʻike ʻia ʻo ia he mea pono loa. Huahana Watson no Genomics, i hoʻomohala ʻia me ka hui pū ʻana me ke Kulanui o North Carolina, Yale University, a me nā ʻoihana ʻē aʻe nā hale hana genetic no ka hoʻomākaukau ʻana i nā hōʻike no ka poʻe oncologists. Hoʻoiho ʻo Watson i ka faila papa inoa hoʻololi genetic i ka mea maʻi a hiki ke hoʻopuka i kahi hōʻike i loko o nā minuke e loaʻa nā manaʻo no nā lāʻau koʻikoʻi āpau a me nā hoʻokolohua lapaʻau. Mālama ʻo Watson i ka ʻike genetic me ka maʻalahino ka mea, ua hōʻike ʻia lākou i loko o nā faila i kūkulu ʻia a ʻaʻole i loaʻa i nā ambiguities - aia he hoʻololi a ʻaʻohe hoʻololi.

Ua paʻi nā hoa IBM ma ke Kulanui o North Carolina i kahi pepa e pili ana i ka pono ma 2017. Ua ʻike ʻo Watson i nā hoʻololi koʻikoʻi i ʻike ʻole ʻia e nā haʻawina kanaka ma 32% o lākou. ua aʻo nā maʻi, e hoʻolilo iā lākou i mau moho maikaʻi no ka lāʻau hou. Eia nō naʻe, ʻaʻohe mea hōʻike e alakaʻi i ka hoʻohana ʻana i nā hopena lapaʻau maikaʻi aʻe.

ʻO ka home o nā protein

ʻO kēia a me nā hiʻohiʻona ʻē aʻe he mea kōkua i ka ulu ʻana o ka manaʻoʻiʻo e hoʻoponopono ʻia nā hemahema āpau o ka mālama ola kino, akā pono mākou e ʻimi i nā wahi e hiki ai ke kōkua maoli kēia, no ka mea ʻaʻole maikaʻi loa ka poʻe ma laila. ʻO ia kahua, no ka laʻana, noiʻi protein. I ka makahiki i hala, ua puka mai ka ʻike e hiki ke wānana pololei i ke ʻano o nā protein e pili ana i kā lākou kaʻina (2). He hana kuʻuna kēia, ma waho aʻe o ka mana o nā kānaka wale nō, akā i nā kamepiula ikaika. Inā ʻike mākou i ka hoʻohālikelike pololei ʻana o ka wili ʻana o nā molekole protein, e loaʻa nā manawa nui no ka lāʻau lapaʻau. Manaʻo nā kānaka ʻepekema me ke kōkua o AlphaFold e aʻo mākou i nā hana o nā tausani, a ʻo ia hoʻi, e ʻae iā mākou e hoʻomaopopo i nā kumu o nā maʻi he nui.

Kiʻi 2. Hoʻohālike ʻia ka wili protein me DeepMind's AlphaFold.

Ano ʻike mākou i ʻelua haneli miliona mau protein, akā hoʻomaopopo pono mākou i ke ʻano a me ka hana o kahi hapa liʻiliʻi o lākou. Nā Paena ʻo ia ke kahua kūkulu kumu o nā mea ola. Na lākou ke kuleana no ka hapa nui o nā kaʻina hana i loko o nā cell. Pehea lākou e hana ai a me kā lākou hana e hoʻoholo ʻia e kā lākou hoʻolālā 50D. Lawe lākou i ke ʻano kūpono me ka ʻole o nā ʻōlelo aʻo, alakaʻi ʻia e nā kānāwai o ka physics. No nā makahiki he nui, ʻo nā ʻano hoʻokolohua ke ala nui no ka hoʻoholo ʻana i ke ʻano o nā protein. I nā makahiki XNUMX, hoʻohana X-ray crystallographic ala. I nā makahiki i hala iho nei, ua lilo ia i mea hana noiʻi o ke koho. microscopy kristal. I nā makahiki 80 a me 90, hoʻomaka ka hana ma ka hoʻohana ʻana i nā kamepiula e hoʻoholo i ke ʻano o nā protein. Akā naʻe, ʻaʻole i ʻoluʻolu nā hopena i ka poʻe ʻepekema. ʻAʻole i hana nā ʻano hana no kekahi mau protein no nā mea ʻē aʻe.

Aia i ka makahiki 2018 ʻAlaFold loaʻa ka ʻike mai ka poʻe akamai ma hoʻohālike pūmua. Eia naʻe, i kēlā manawa ua hoʻohana ʻo ia i nā ala like loa me nā papahana ʻē aʻe. Ua hoʻololi ka poʻe ʻepekema i nā loea a hana i kahi ʻē aʻe, nāna i hoʻohana i ka ʻike e pili ana i nā palena kino a me nā geometric i ka pelu ʻana o nā molekole protein. ʻAlaFold hāʻawi i nā hopena like ʻole. I kekahi manawa ua hana maikaʻi ʻo ia, i kekahi manawa ʻino. Akā, aneane ʻelua hapakolu o kāna mau wānana i hui pū me nā hopena i loaʻa ma nā ʻano hoʻokolohua. I ka hoʻomaka ʻana o ka makahiki 2, ua wehewehe ka algorithm i ke ʻano o nā protein o ka maʻi SARS-CoV-3. Ma hope mai, ua ʻike ʻia ua kūlike nā wānana no ka protein Orf2020a me nā hopena i loaʻa i ka hoʻokolohua.

ʻAʻole wale ia e pili ana i ke aʻo ʻana i nā ala o loko o ka hoʻopili ʻana i nā protein, akā pili pū kekahi i ka hoʻolālā. Ua hoʻohana ʻia nā mea noiʻi mai ka NIH BRAIN initiative aʻo mīkini e hoʻomohala i kahi protein e hiki ke hahai i nā pae serotonin o ka lolo i ka manawa maoli. ʻO Serotonin kahi neurochemical e pāʻani ana i kahi kuleana nui i ka mālama ʻana o ka lolo i ko mākou mau manaʻo a me nā manaʻo. No ka laʻana, nui nā antidepressants i hoʻolālā ʻia e hoʻololi i nā hōʻailona serotonin i lawe ʻia ma waena o nā neurons. Ma kahi ʻatikala ma ka puke pai ʻo Cell, ua wehewehe nā ʻepekema pehea lākou e hoʻohana ai i ka holomua ʻano hana ʻenekinia e hoʻohuli i ka protein bacteria i mea hana noiʻi hou e hiki ke kōkua i ka hoʻouna ʻana i ka serotonin me ka pololei ʻoi aku ka pololei ma mua o nā ʻano hana o kēia manawa. Ua hōʻike ʻia nā hoʻokolohua preclinical, ka hapa nui o nā ʻiole, hiki i ka mea ʻike ke ʻike koke i nā loli liʻiliʻi o ka pae serotonin o ka lolo i ka wā hiamoe, makaʻu a me ka launa pū ʻana, a hoʻāʻo i ka pono o nā lāʻau psychoactive hou.

ʻAʻole i kūleʻa mau ka hakakā ʻana i ka maʻi maʻi

Ma hope o nā mea a pau, ʻo kēia ka maʻi ahulau mua a mākou i kākau ai ma MT. Eia nō naʻe, no ka laʻana, inā mākou e kamaʻilio e pili ana i ke kaʻina hana o ka hoʻomohala ʻana o ka maʻi maʻi, a laila i ka pae mua, ʻike ʻia ʻo AI he mea hemahema. Ua hoʻopiʻi ka poʻe akamai i kēlā Akamai ʻimi hoʻopunipuni ʻAʻole hiki ke wānana pololei i ka nui o ka laha o ka coronavirus ma muli o ka ʻikepili mai nā maʻi maʻi ma mua. “Ke hana maikaʻi nei kēia mau hoʻonā i kekahi mau wahi, e like me ka ʻike ʻana i nā helehelena i loaʻa kekahi helu o nā maka a me nā pepeiao. ʻO ka maʻi ahulau SARS-CoV-2 ʻO kēia nā hanana i ʻike mua ʻole ʻia a me nā ʻano hou he nui, no laila ʻaʻole hana maikaʻi ka ʻike artificial e pili ana i ka ʻikepili mōʻaukala i hoʻohana ʻia e aʻo ai. Ua hōʻike ka maʻi maʻi maʻi e pono mākou e ʻimi i nā ʻenehana a me nā ala ʻē aʻe, "wahi a Maxim Fedorov mai Skoltech i kahi ʻōlelo ʻApelila 2020 i ka pāpāho Lūkini.

Ma ka manawa he mau akā naʻe nā algorithms e ʻike ʻia e hōʻoia i ka pono nui o AI i ka hakakā ʻana me COVID-19. Ua hoʻomohala ka poʻe ʻepekema ma US i kahi ʻōnaehana i ka hāʻule o 2020 e ʻike i nā hiʻohiʻona maʻamau i ka poʻe me COVID-19, ʻoiai inā ʻaʻohe o lākou mau hōʻailona ʻē aʻe.

I ka puka ʻana mai o nā lāʻau lapaʻau, ua hānau ʻia ka manaʻo e kōkua i ka hoʻopaʻa ʻana i ka lehulehu. Hiki iā ia, no ka laʻana kōkua i ka hoʻohālike i ka lawe ʻana a me ka logistic o nā kano. No ka hoʻoholo ʻana i ka heluna kanaka e pono ke hoʻopaʻa mua ʻia i ka maʻi maʻi wikiwiki. E kōkua nō hoʻi ia e wānana i ka noi a hoʻopaʻa i ka manawa a me ka wikiwiki o ka hoʻopaʻa ʻana ma o ka ʻike koke ʻana i nā pilikia a me nā bottlenecks i loko o ka logistics. ʻO ka hui pū ʻana o nā algorithm me ka nānā mau ʻana hiki ke hāʻawi koke i ka ʻike e pili ana i nā hopena ʻaoʻao a me nā hanana olakino.

kēia nā ʻōnaehana hoʻohana AI i ka hoʻomaikaʻi ʻana a me ka hoʻomaikaʻi ʻana i ka mālama olakino ua ʻike mua ʻia. Ua mahalo ʻia ko lākou mau pono kūpono; no ka laʻana, ka ʻōnaehana mālama olakino i hoʻomohala ʻia e Macro-Eyes ma ke Kulanui ʻo Stanford ma US. E like me ka nui o nā keʻena lapaʻau ʻē aʻe, ʻo ka pilikia ka nele o nā maʻi i hōʻike ʻole ʻia no nā koho. Maka Makro ua kūkulu ʻia kahi ʻōnaehana hiki ke wānana pono i nā poʻe maʻi ʻaʻole paha i laila. I kekahi mau kūlana, hiki iā ia ke kuhikuhi i nā manawa a me nā wahi ʻē aʻe no nā keʻena lapaʻau, kahi e hoʻonui ai i ka manawa o kahi maʻi e hōʻike ana. Ma hope mai, ua hoʻohana ʻia ka ʻenehana like ma nā wahi like ʻole mai Arkansas a i Nigeria me ke kākoʻo, ʻo ia hoʻi, ka US Agency for International Development i.

Ma Tanzania, ua hana ʻo Macro-Eyes i kahi papahana i manaʻo ʻia ka hoʻonui ʻana i ka maʻi o nā keiki. Ua nānā ka polokalamu i ka nui o nā ʻano o nā lāʻau lapaʻau e pono e hoʻouna ʻia i kahi kikowaena lāʻau lapaʻau i hāʻawi ʻia. Ua hiki nō iā ia ke loiloi i nā ʻohana e hiki ʻole ke hoʻopaʻa i kā lākou mau keiki, akā hiki iā lākou ke hoʻohuli ʻia me nā manaʻo kūpono a me kahi o kahi kikowaena lāʻau lapaʻau ma kahi kūpono. Me ka hoʻohana ʻana i kēia polokalamu, ua hiki i ke aupuni Tanzanian ke hoʻonui i ka pono o kāna papahana lapaʻau e 96%. a e hoʻemi i ka ʻōpala kano i 2,42 no 100 poʻe.

Ma Sierra Leone, kahi i nalowale ai ka ʻikepili olakino o nā kamaʻāina, ua hoʻāʻo ka hui e hoʻohālikelike i kēia me ka ʻike e pili ana i ka hoʻonaʻauao. Ua ʻike ʻia ua lawa ka helu o nā kumu a me kā lākou mau haumāna wale nō e wānana i ka 70 pakeneka. ʻO ka pololei o ka loaʻa ʻana o ka wai maʻemaʻe i ka hale lapaʻau kūloko, ʻo ia ka wāwae o ka ʻikepili i ke olakino o ka poʻe e noho ana ma laila (3).

3. Hōʻikeʻike Macro-Eyes o nā polokalamu mālama ola i alakaʻi ʻia e AI ma ʻApelika.

ʻAʻole nalowale ka moʻolelo o ke kauka mīkini

ʻOiai nā hemahema Watson ke hoʻomohala ʻia nei nā ala diagnostic hou a manaʻo ʻia he ʻoi aku ka holomua. Hoʻohālikelike i hana ʻia ma Sweden i Kepakemapa 2020. hoʻohana ʻia i ka diagnostics imaging o ka maʻi maʻi umauma hōʻike i ka hana maikaʻi loa o lākou e like me ka radiologist. Ua hoʻāʻo ʻia nā algorithms me ka hoʻohana ʻana i ʻeiwa kaukani kiʻi mammography i loaʻa i ka wā o ka nānā ʻana. ʻEkolu mau ʻōnaehana, i koho ʻia ʻo AI-1, AI-2 a me AI-3, loaʻa ka pololei o 81,9%, 67%. a me 67,4%. No ka hoʻohālikelike, no ka poʻe radiologists nāna e wehewehe i kēia mau kiʻi ma ke ʻano he mua, ʻo kēia helu he 77,4%, a ma ka hihia o nā mea hōʻike lekiōʻo ia ka lua o ka wehewehe ʻana, he 80,1 pakeneka. ʻO ka maikaʻi loa o nā algorithms i hiki ke ʻike i nā hihia i hala ʻole nā ​​​​radiologists i ka wā o ka nānā ʻana, a ua ʻike ʻia nā wahine he maʻi ma lalo o hoʻokahi makahiki.

Wahi a nā mea noiʻi, hōʻoia kēia mau hopena nā algorithms naʻauao hana kōkua i ka hoʻoponopono ʻana i nā maʻi hoʻopunipuni hewa ʻole i hana ʻia e nā radiologists. ʻO ka hoʻohui ʻana i nā mana o AI-1 me kahi radiologist awelika i hoʻonui i ka nui o nā maʻi maʻi umauma i ʻike ʻia e 8%. Ke manaʻo nei ka hui ma ka Royal Institute e alakaʻi nei i kēia noiʻi i ka maikaʻi o nā algorithms AI e hoʻomau i ka ulu. Ua paʻi ʻia kahi wehewehe piha o ka hoʻokolohua ma JAMA Oncology.

W ma ka pālākiō ʻelima. I kēia manawa, ke ʻike nei mākou i kahi wikiwiki ʻenehana koʻikoʻi a hiki i ka pae IV (automate kiʻekiʻe), i ka wā e hoʻoili kūʻokoʻa ai ka ʻōnaehana i ka ʻikepili i loaʻa a hāʻawi i ka loea i ka ʻike i hoʻopaʻa ʻia. Mālama kēia i ka manawa, pale i ka hewa kanaka a hāʻawi i ka mālama maʻi ʻoi aku ka maikaʻi. ʻO ia kāna i hoʻoholo ai i kekahi mau mahina i hala ʻO Stan A.I. ma ke kula lapaau kokoke iaia, prof. ʻO Janusz Braziewicz mai ka Polish Society for Nuclear Medicine ma kahi ʻōlelo i ka Polish Press Agency.

4. Nānā mīkini i nā kiʻi lapaʻau

Algorithms, e like me ka poʻe akamai e like me prof. ʻO Brazievichʻoiai he mea nui i kēia ʻoihana. ʻO ke kumu ka piʻi wikiwiki o ka helu o nā hoʻokolohua diagnostic imaging. No ka manawa 2000-2010 wale nō. ua hoʻonui ʻia ka nui o nā hoʻokolohua MRI a me nā hoʻokolohua. ʻO ka mea pōʻino, ʻaʻole i hoʻonui ʻia ka nui o nā kauka loea i hiki ke hoʻokō wikiwiki a hilinaʻi. Aia nō ka hapa o nā ʻenehana mākaukau. ʻO ka hoʻokō ʻana i nā algorithms i hoʻokumu ʻia AI e mālama i ka manawa a ʻae i ka standardization piha o nā kaʻina hana, a me ka pale ʻana i ka hewa o ke kanaka a ʻoi aku ka maikaʻi, nā mālama pilikino no nā maʻi.

E like me ka mea i ʻike ʻia lāʻau lapaʻau forensic hiki ke pōmaikaʻi mai ka hoʻomohala ʻana o ka naʻauao hana. Hiki i nā poʻe loea ma kēia kahua ke hoʻoholo i ka manawa kūpono o ka make ʻana o ka mea i make ma o ka loiloi kemika o nā mea huna o nā ilo a me nā mea ʻē aʻe e hānai ana i nā ʻiʻo make. Piʻi ka pilikia i ka wā e hoʻohui ʻia ai nā mea huna mai nā ʻano necrophages like ʻole i ka nānā ʻana. ʻO kēia kahi e komo ai ke aʻo mīkini. Ua hoʻomohala nā kānaka ʻepekema ma ke Kulanui o Albany he ʻano hana akamai e hiki ai ke ʻike wikiwiki i nā ʻano ilo ma muli o kā lākou "manamana lima kemika". Ua hoʻomaʻamaʻa ka hui i kā lākou polokalamu kamepiula me ka hoʻohana ʻana i nā hui ʻana o nā hui like ʻole o nā mea huna kemika mai nā ʻano lele ʻeono. Ua wehewehe ʻo ia i nā pūlima kemika o nā larvae insect me ka hoʻohana ʻana i ka spectrometry nuipaʻa, kahi e ʻike ai i nā kemika ma ke ana pololei ʻana i ka ratio o ka nuipaʻa me ka hoʻoili uila o kahi ion.

No laila, e like me kāu e ʻike ai ʻO AI ma ke ʻano he mākaʻi noiʻi ʻaʻole maikaʻi loa, hiki ke hoʻohana nui ʻia i kahi keʻena forensic. Ua manaʻo nui paha mākou iā ia i kēia manawa, e kali ana i nā algorithm e hoʻopau i nā kauka mai ka hana (5). Ke nana kakou Akamai ʻimi hoʻopunipuni ʻoi aku ka ʻoiaʻiʻo, ka nānā ʻana i nā pono kūpono kūikawā ma mua o ka laulā, ʻike hou ʻia kāna ʻoihana i ka lāʻau lapaʻau.

5. Ka hihio o ke kaa o ke kauka