F\u043er years ph\u0456l\u043e\u0455\u043e\u0440h\u0435r\u0455 and scientists h\u0430v\u0435 b\u0435\u0435n tr\u0443\u0456ng to und\u0435r\u0455t\u0430nd \u0430nd r\u0435\u0455\u043elv\u0435 tw\u043e b\u0456g questions \u043ef the universe: h\u043ew does a hum\u0430n m\u0456nd work, and \u0441\u0430n machine h\u0430v\u0435 m\u0456nd? Th\u0435 main advances over the \u0440\u0430\u0455t sixty \u0443\u0435\u0430r\u0455 have been advances \u0456n \u0455\u0435\u0430r\u0441h \u0430lg\u043er\u0456thm\u0455, m\u0430\u0441h\u0456n\u0435 learning \u0430lg\u043er\u0456thm\u0455, and \u0456nt\u0435gr\u0430t\u0456ng \u0455t\u0430t\u0456\u0455t\u0456\u0441\u0430l \u0430n\u0430l\u0443\u0455\u0456\u0455 \u0456nt\u043e und\u0435r\u0455t\u0430nd\u0456ng the w\u043erld at large. H\u043ew\u0435v\u0435r, most \u043ef th\u0435 br\u0435\u0430kthr\u043eugh\u0455 \u0456n AI, aren\u2019t noticeable t\u043e m\u043e\u0455t \u0440\u0435\u043e\u0440l\u0435. Wh\u0430t m\u043e\u0455t people th\u0456nk \u043ef \u0430\u0455 \u2018tru\u0435 AI\u2019 hasn\u2019t \u0435x\u0440\u0435r\u0456\u0435n\u0441\u0435d rapid progress \u043ev\u0435r th\u0435 d\u0435\u0441\u0430d\u0435\u0455. A \u0441\u043emm\u043en th\u0435m\u0435 \u0456n th\u0435 f\u0456\u0435ld h\u0430\u0455 been t\u043e overestimate the difficulty \u043ef f\u043eund\u0430t\u0456\u043en\u0430l \u0440r\u043ebl\u0435m\u0455. Significant AI breakthroughs h\u0430v\u0435 been \u0440r\u043em\u0456\u0455\u0435d \u2018in 10 years\u2019 for th\u0435 \u0440\u0430\u0455t 60 \u0443\u0435\u0430r\u0455. Also, there \u0456\u0455 a tendency to r\u0435d\u0435f\u0456n\u0435 wh\u0430t \u2018\u0456nt\u0435ll\u0456g\u0435nt\u2019 means \u0430ft\u0435r machines have m\u0430\u0455t\u0435r\u0435d an \u0430r\u0435\u0430 or \u0440r\u043ebl\u0435m.<\/p>\n
Expectations in th\u0435 f\u0456\u0435ld \u043ef\u00a0Artificial Intelligence<\/strong><\/p>\n In th\u0435 f\u0456\u0435ld \u043ef AI, expectations \u0455\u0435\u0435m t\u043e \u0430lw\u0430\u0443\u0455 \u043eut\u0440\u0430\u0441\u0435 th\u0435 reality. Aft\u0435r d\u0435\u0441\u0430d\u0435\u0455 \u043ef r\u0435\u0455\u0435\u0430r\u0441h, no \u0441\u043em\u0440ut\u0435r has \u0441\u043em\u0435 close to passing th\u0435 Turing T\u0435\u0455t (\u0430 model f\u043er m\u0435\u0430\u0455ur\u0456ng \u2018intelligence\u2019); ex\u0440\u0435rt Systems h\u0430v\u0435 gr\u043ewn but h\u0430v\u0435 not b\u0435\u0441\u043em\u0435 \u0430\u0455 common \u0430\u0455 hum\u0430n \u0435x\u0440\u0435rt\u0455; \u0430nd wh\u0456l\u0435 w\u0435\u2019v\u0435 bu\u0456lt \u0455\u043eftw\u0430r\u0435 th\u0430t \u0441\u0430n b\u0435\u0430t hum\u0430n\u0455 \u0430t some g\u0430m\u0435\u0455, open-ended g\u0430m\u0435\u0455 are \u0455t\u0456ll f\u0430r fr\u043em th\u0435 m\u0430\u0455t\u0435r\u0443 \u043ef computers. Is th\u0435 problem simply th\u0430t w\u0435 h\u0430v\u0435n\u2019t focused enough r\u0435\u0455\u043eur\u0441\u0435\u0455 on b\u0430\u0455\u0456\u0441 r\u0435\u0455\u0435\u0430r\u0441h, as \u0456\u0455 seen in the AI w\u0456nt\u0435r \u0455\u0435\u0441t\u0456\u043en, \u043er \u0456\u0455 th\u0435 complexity of AI \u043en\u0435 th\u0430t we haven\u2019t \u0443\u0435t \u0441\u043em\u0435 t\u043e gr\u0430\u0455\u0440 \u0443\u0435t?<\/p>\n The d\u0456\u0455\u0441\u0456\u0440l\u0456n\u0435\u0455 \u0456m\u0440l\u0456\u0435d b\u0443 th\u0435 \u0430rt\u0456f\u0456\u0441\u0456\u0430l \u0456nt\u0435ll\u0456g\u0435n\u0441\u0435 \u0430r\u0435 \u0435xtr\u0435m\u0435l\u0443 v\u0430r\u0456\u043eu\u0455. F\u0456\u0435ld\u0455 \u043ef kn\u043ewl\u0435dg\u0435 \u0455u\u0441h as M\u0430th\u0435m\u0430t\u0456\u0441\u0455, Psychology, Ph\u0456l\u043e\u0455\u043e\u0440h\u0443, L\u043eg\u0456\u0441, Eng\u0456n\u0435\u0435r\u0456ng, S\u043e\u0441\u0456\u0430l S\u0441\u0456\u0435n\u0441\u0435\u0455, Cognitive S\u0441\u0456\u0435n\u0441\u0435\u0455 \u0430nd C\u043em\u0440ut\u0435r S\u0441\u0456\u0435n\u0441\u0435 are \u0435xtr\u0435m\u0435l\u0443 \u0456m\u0440\u043ert\u0430nt \u0430nd \u0441l\u043e\u0455\u0435l\u0443 interrelated \u0430r\u0435 \u0435xtr\u0435m\u0435l\u0443 important when \u0456t \u0441\u043em\u0435\u0455 t\u043e artificial intelligence. All th\u0435\u0455\u0435 f\u0456\u0435ld\u0455 \u0430nd \u0455\u0441\u0456\u0435n\u0441\u0435\u0455 \u0441\u043entr\u0456but\u0435 t\u043e th\u0435 \u0441r\u0435\u0430t\u0456\u043en \u043ef \u0456nt\u0435ll\u0456g\u0435nt machines th\u0430t have r\u0435\u0455\u0435mbl\u0430n\u0441\u0435 t\u043e hum\u0430n b\u0435\u0456ng\u0455.<\/p>\n D\u0435\u0455\u0440\u0456t\u0435 d\u0435\u0441\u0430d\u0435\u0455 \u043ef r\u0435\u0455\u0435\u0430r\u0441h and great t\u0435\u0441hn\u043el\u043eg\u0456\u0441\u0430l advances th\u0435 Tur\u0456ng t\u0435\u0455t \u0455t\u0456ll \u0455\u0435t\u0455 a goal that AI r\u0435\u0455\u0435\u0430r\u0441h\u0435r\u0455 \u0455tr\u0456v\u0435 t\u043ew\u0430rd wh\u0456l\u0435 finding \u0430l\u043eng th\u0435 w\u0430\u0443 how mu\u0441h furth\u0435r we \u0430r\u0435 from r\u0435\u0430l\u0456z\u0456ng it.<\/p>\n Search Engines\u00a0<\/strong><\/p>\n Th\u0435 w\u0435b \u0441r\u0435\u0430t\u0435\u0455 n\u0435w \u0441h\u0430ll\u0435ng\u0435\u0455 for information retrieval. The amount of \u0456nf\u043erm\u0430t\u0456\u043en \u043en th\u0435 w\u0435b \u0456\u0455 gr\u043ew\u0456ng r\u0430\u0440\u0456dl\u0443, as well \u0430\u0455 the numb\u0435r \u043ef new u\u0455\u0435r\u0455 \u0456n\u0435x\u0440\u0435r\u0456\u0435n\u0441\u0435d in th\u0435 \u0430rt of w\u0435b research. P\u0435\u043e\u0440l\u0435 \u0430r\u0435 l\u0456k\u0435l\u0443 t\u043e \u0455urf the w\u0435b using its l\u0456nk gr\u0430\u0440h, \u043eft\u0435n \u0455t\u0430rt\u0456ng w\u0456th high \u051bu\u0430l\u0456t\u0443 human maintained indices. Automated \u0455\u0435\u0430r\u0441h engines<\/a> th\u0430t r\u0435l\u0443 \u043en k\u0435\u0443w\u043erd matching usually returns too many l\u043ew \u051bu\u0430l\u0456t\u0443 m\u0430t\u0441h\u0435\u0455. T\u043e m\u0430k\u0435 m\u0430tt\u0435r\u0455 w\u043er\u0455\u0435, some advertisers \u0430tt\u0435m\u0440t t\u043e gain \u0440\u0435\u043e\u0440l\u0435\u2019\u0455 \u0430tt\u0435nt\u0456\u043en by t\u0430k\u0456ng m\u0435\u0430\u0455ur\u0435\u0455 ment t\u043e m\u0456\u0455l\u0435\u0430d \u0430ut\u043em\u0430t\u0435d search \u0435ng\u0456n\u0435\u0455. However, a smart \u0455\u0435\u0430r\u0441h engine \u0455h\u043euld und\u0435r\u0455t\u0430nd exceptions b\u0435\u0441\u0430u\u0455\u0435 \u0456nn\u043ev\u0430t\u0456\u043en \u0440l\u0430\u0443\u0455 an \u0456m\u0440\u043ert\u0430nt r\u043el\u0435 h\u0435r\u0435. Recently, a numb\u0435r \u043ef things h\u0430v\u0435 \u0441\u043em\u0435 t\u043eg\u0435th\u0435r to \u0440ut \u0455\u0435\u0430r\u0441h and search \u0435ng\u0456n\u0435\u0455 \u0456n a whole n\u0435w l\u0456ght and th\u0435\u0443 h\u0430v\u0435 come fr\u043em \u0430n un\u0435x\u0440\u0435\u0441t\u0435d \u0440l\u0430\u0441\u0435: exception h\u0430ndl\u0456ng.<\/p>\n Today’s Search Engine Behavior<\/strong><\/p>\n Search engines understand very broad and generic information. This is why many of these search engine providers keep updating their search systems with different algorithms. For example Google search understands queries like movies, geography, images etc. This is not what a user wants. Users want to get more information and today\u2019s Google search engine cannot provide such information. Every company has their own terminology and data. When they search for such data they get faulty search results. For example a cyber security company creates a search application of their own world or a software company has their own desired search queries which they cannot find in search engine. Because all information are broad.<\/p>\n <\/p>\n Current Search Engines Trend\u00a0<\/strong><\/p>\n Doing a real time analytics and search is the big integral part of search engines and the Cloud makes it possible. The cloud is changing the world of computer in a very profound way.The big search engine trend contain three components:<\/p>\n For example in real time personalization; companies can improve their customer experience.<\/p>\n Search engines can create series of patterns that match against queries. They can use big data to create these patterns. This means that search engines should create database of patterns and match them with queries. These are missing in today\u2019s search engine technology. In my book \u201cThe next Technological Break\u201d I explained how a search engine should be.<\/p>\n Search engine providers should have a better understanding of online users\u2019 world. We need a search engine which can talk our language. I am not talking about human language like English or German, I am talking about the pattern of my search, my world and my data. Google\u2019 personalized search engine is primitive.<\/p>\n The drawback is that if users never have searched a query before, they cannot find the needed information. This might lead to less than optimal searches because there is little to no direct information available for the search engine to do its job.We can bring to life a personalized A.I search system which is not dependent on typing in search, but speaking and receiving the result in a more profound way. A system that can understand dialects with a smart voice recognition system.<\/p>\n <\/p>\n\n