ʻO ka lolo artificial: ka manaʻo hoʻopunipuni i loko o kahi mīkini

ʻAʻole pono ka ʻike artificial i kope o ka naʻauao kanaka, no laila ʻo ka papahana o ka hana ʻana i ka lolo artificial, kahi kope ʻenehana o ke kanaka, he wahi ʻokoʻa iki o ka ʻimi noiʻi. Eia naʻe, hiki i kekahi manawa o ka hoʻomohala ʻana e hui pū kēia papahana me ka hoʻomohala ʻana o AI. He halawai maikai keia.

Ua hoʻokumu ʻia ka European Human Brain Project ma 2013. ʻAʻole ia i wehewehe ʻia ma ke ʻano he "papahana lolo artificial". Akā, hoʻoikaika ia i ka ʻano cognitive, ka makemake e hōʻike maikaʻi i kā mākou kikowaena kauoha. ʻO ka mana hou o ka WBP ʻaʻole ia he mea koʻikoʻi ma ke ʻano he mea hoʻoikaika no ka hoʻomohala ʻana i ka ʻepekema. Eia naʻe, ʻaʻole hiki ke hōʻole ʻia ʻo ka pahuhopu o ka poʻe ʻepekema e hana nei i kēia papahana ʻo ia ka hana ʻana i ka simulation lolo hana, a aia kēia i loko o kahi ʻumi makahiki, ʻo ia hoʻi, mai 2013 a 2023.

Manaʻo ka poʻe ʻepekema ʻo ka palapala ʻāina kikoʻī o ka lolo e hiki ke hana hou i ka lolo kanaka. Hoʻokahi haneli trillion pili i hana ʻia i loko o ia mea i pani ʻia - no laila, ke hoʻomaka nei ka hana koʻikoʻi e hana i kahi palapala ʻāina o kēia paʻakikī hiki ʻole ke ʻike ʻia, i kapa ʻia ʻo ka connectome.

Ua hoʻohana mua ʻia ka huaʻōlelo i nā pepa ʻepekema ma 2005, kūʻokoʻa e nā mea kākau ʻelua: Olaf Spons o ke Kulanui o Indiana a me Patrick Hagmann o ka Halemai Ke Kulanui o Lausanne.

Manaʻo ka poʻe ʻepekema i ka wā e palapala ai lākou i nā mea āpau i ka lolo, a laila hiki ke kūkulu ʻia i ka lolo artificial, e like me ke kanaka, a laila, ʻo wai ka mea ʻike, ʻoi aku ka maikaʻi ... ʻO ka papahana hana connectome, ma ka inoa a me ke ʻano, e pili ana i ka papahana kaulana e wehewehe i ka genome kanaka - ka Human Genome Project. Ma kahi o ka manaʻo genome, hoʻohana ka papahana i hoʻomaka i ka manaʻo connectome e wehewehe i ka huina o nā pilina neural i ka lolo. Manaʻo nā kānaka ʻepekema ʻo ke kūkulu ʻana i kahi palapala ʻāina piha o nā pili neural e hoʻohana ʻia ʻaʻole wale i ka ʻepekema hana, akā i ka mālama ʻana i nā maʻi.

www.humanconnectomeproject.org

ʻO ka mea mua a hiki i kēia manawa ʻo ka mea hoʻohui i ʻike piha ʻia ʻo ia ka pūnaewele o nā pilina neuronal i loko o ka ʻōnaehana nerve o caenorhabditis elegans. Ua hoʻomohala ʻia e ka hana hou ʻana o 1986D o ke ʻano nerve me ka microscopy electron. Ua paʻi ʻia ka hopena o ka hana ma 30. I kēia manawa, ʻo ka papahana noiʻi nui loa i hana ʻia i loko o ke kahua o ka ʻepekema hou i kapa ʻia ʻo connectomics ʻo ia ka Human Connectome Project, i kākoʻo ʻia e ka American National Institutes of Health (ka huina o $ XNUMX miliona).

Algorithm naʻauao

ʻAʻole maʻalahi ka hana ʻana i kahi kope synthetic o ka lolo kanaka. E maʻalahi paha ka ʻike ʻana ʻo ka naʻauao kanaka ka hopena o kahi algorithm maʻalahi i wehewehe ʻia ma ka pukana ʻo Nowemapa 2016 o Frontiers in Systems Neuroscience. Ua loaʻa iā Joe Tsien, he kanaka ʻepekema neuro ma Augusta University of Georgia.

Ua hoʻokumu ʻia kāna noiʻi ʻana ma luna o ka manaʻo i kapa ʻia ʻo connectionism, a i ʻole ke kumumanaʻo o ke aʻo ʻana i ka makahiki kikohoʻe. Hoʻokumu ʻia ma ka manaʻoʻiʻo ʻo ke kumu o ke aʻo ʻana ʻo ke aʻo ʻana i ka noʻonoʻo, ʻoi aku ka mua ma mua o ka loaʻa ʻana o ka ʻike. ʻO nā mea kākau o kēia kumumanaʻo: ʻo George Siemens, nāna i wehewehe i kāna mau manaʻo ma ka pepa Connectivism: A Theory of Learning for the Digital Age, a me Stephen Downes. ʻO ka mākaukau nui ma ʻaneʻi, ʻo ia ka hiki ke hoʻohana pono i nā holomua ʻenehana a loaʻa ka ʻike i loko o nā ʻikepili waho (ka mea i kapa ʻia ʻike-kahi), ʻaʻole mai ka ʻike i aʻo ʻia i ke kaʻina aʻo, a me ka hiki ke hoʻopili a hoʻopili iā lākou me nā ʻike ʻē aʻe.

Ma ka pae neural, wehewehe ke kumumanaʻo i nā pūʻulu o nā neurons i hana i nā hui paʻakikī a pili pū e pili ana i nā manaʻo kumu a me nā ʻike. Ma ke aʻo ʻana i nā holoholona hoʻokolohua me nā electrodes, ua ʻike nā ʻepekema ʻo kēia mau "hui" neural i koho mua ʻia no kekahi mau ʻano hana. Hoʻokumu kēia i kahi ʻano algorithm o ka lolo me kekahi mau pilina pili. Manaʻo ka poʻe ʻepekema ʻaʻole ʻokoʻa ka hana o ka lolo o ke kanaka, me kāna mau pilikia a pau, ma mua o ka lolo o nā rodents.

Ka lolo mai nā memristors

Ke haku mākou i nā algorithms, hiki ke hoʻohana ʻia nā memristors e hoʻohālikelike kino i ka lolo kanaka. Ua hōʻoia hou nā kānaka ʻepekema mai ke Kulanui o Southampton i ka pono ma kēia ʻano.

ʻO nā memristors ʻepekema Pelekania, i hana ʻia mai nā metala oxides, i hana ʻia ma ke ʻano he synapses artificial no ke aʻo ʻana (a me ke aʻo hou ʻana) me ka hoʻohana ʻole ʻana i waho, me ka hoʻohana ʻana i nā datasets i loaʻa i nā ʻike pili ʻole, e like me ke kanaka. No ka hoʻomanaʻo ʻana o nā memristors i kā lākou mau mokuʻāina i ka wā i pio ai, pono lākou e ʻai i ka mana liʻiliʻi ma mua o nā mea kaapuni maʻamau. He mea koʻikoʻi kēia ma ke ʻano o kekahi mau mea liʻiliʻi ʻaʻole hiki ke loaʻa a ʻaʻole pono e loaʻa kahi pākaukau nui.

ʻOiaʻiʻo, ʻo ka hoʻomaka wale nō kēia o ka hoʻomohala ʻana o kēia ʻenehana. Inā hoʻohālike ʻo AI i ka lolo o ke kanaka, pono ia ma kahi o nā haneli haneli o nā synapses. ʻO ka hoʻonohonoho o nā memristors i hoʻohana ʻia e nā mea noiʻi ʻoi aku ka maʻalahi, no laila ua kaupalena ʻia i ka ʻimi ʻana i nā hiʻohiʻona. Eia naʻe, ʻike ka hui ʻo Southampton i ka hihia o nā noi haʻahaʻa, ʻaʻole pono e hoʻohana i ka nui o nā memristors. Mahalo iā lākou, hiki ke kūkulu ʻia, no ka laʻana, nā sensor e hoʻokaʻawale i nā mea a ʻike i nā ʻano me ka ʻole o ke kanaka. ʻO ia mau mea pono e hoʻohana nui ʻia i nā wahi paʻakikī a i ʻole ka pilikia.

Inā mākou e hoʻohui i nā ʻike ākea i hana ʻia e ka Human Brain Project, ka palapala ʻāina o "connectomes", ka ʻike ʻana i nā algorithms naʻauao a me ka ʻenehana o ka uila memristor, malia paha i nā makahiki e hiki mai ana e hiki iā mākou ke kūkulu i kahi lolo artificial, kahi kope pololei. o ke kanaka. ʻO wai ka mea ʻike? Eia kekahi, ʻoi aku ka mākaukau o kā mākou kope synthetic no ka hoʻololi ʻana o ka mīkini ma mua o mākou.