»õ·Î¿î ¹Ì·¡¸¦ ±×¸®´Â ±â¼ú, ÀΰøÁö´É ¼¿µÁÖ Æ÷½ºÅØ ±³¼ö
¼¿µÁÖ Æ÷½ºÅØ ÀΰøÁö´É´ëÇпø ±³¼ö°¡ 18ÀÏ °æºÏµµÃ»¿¡¼ '»õ·Î¿î ¹Ì·¡¸¦ ±×¸®´Â ±â¼ú, ÀΰøÁö´É'À» ÁÖÁ¦·Î °¿¬À» Çß´Ù.
¼±³¼ö´Â ¡ã 4Â÷»ê¾÷Çõ¸íÀÌ ¹«¾ùÀΰ¡? ¡ã Big Data ¡ãArtificial Intelligence(¡âAI¶õ ¹«¾ùÀΰ¡? ¡â AI°¡ ÇÏ´Â °ÍÀº ¹«¾ùÀΰ¡?) ¡ãAI½Ã´ëÀÇ ¸ð½À µîÀÇ ¼øÀ¸·Î °¿¬ÇÑ °¡¿îµ¥
¡ß »ê¾÷Çõ¸í °ú 4Â÷»ê¾÷Çõ¸í?
¡ã »ê¾÷Çõ¸íÀÇ µµ·¡? (1Â÷»ê¾÷Çõ¸í, 1700³â´ë)Áõ±â±â°üÀÇ ¹ß¸í ¡æ (2Â÷ " , 1800³â´ë ÈĹÝ)Àü±â¸¦ ±â¹ÝÀ¸·Î ÄÁº£À̾Ʈ¿¡ ÀÇÇÑ ´ë·®»ý»êüÁ¦¡æ (3Â÷ " , 1900³â´ë ÈĹÝ)ÄÄÇ»ÅÍ&ÀÎÅÍ³Ý ±â¹ÝÀÇ Áö½ÄÁ¤º¸Çõ¸í¡æ (Áö´ÉÁ¤º¸±â¼ú) = Áö´É(AISW) + Á¤º¸(ºòµ¥ÀÌÅÍ+IOT+Ŭ¶ó¿ìµå) (4Â÷»ê¾÷Çõ¸í/2Â÷Á¤º¸Çõ¸í, 21¼¼±â ÃʹÝ)
¡ã 4Â÷»ê¾÷Çõ¸íÀÇ ÃÖÃÊ ¾ð±ÞÀÚ ¤Ñ 2016³â World Economic Forum¿¡¼ SchwabÀÇÀåÀÌ Ã³À½ »ç¿ë ¤Ñ Á¤ÀÇ : - ongoing automation of traditional manufacturing and industrial practices, using modern smart technology. (À§Å°Çǵð¾Æ) - Àΰø Áö´É, »ç¹° ÀÎÅͳÝ, ºòµ¥ÀÌÅÍ, ¸ð¹ÙÀÏ µî ÷´Ü Á¤º¸Åë½Å±â¼úÀÌ °æÁ¦, »çȸ Àü¹Ý¿¡ À¶ÇյǾî Çõ½ÅÀûÀκ¯È°¡ ³ªÅ¸³ª´Â Â÷¼¼´ë »ê¾÷Çõ¸í (IT¿ë¾î»çÀü) ¡ã Ư¡ ¹× ¿ä¼Ò±â¼ú ¤Ñ Hyper-connectivity, Super-intelligence¿¡ ÀÇÇÑ Hyper-convergence ¤Ñ4IRÀ» À§ÇÑ Çٽɱâ¼ú : ICBM(IoT, Cloud, Big Data, Mobile)+AI ¤ÑÁ¦4Â÷»ê¾÷Çõ¸íÀÇ Àü·É : 2016³â º½ À̼¼µ¹°ú ¾ËÆÄ°íÀÇ ´ë°á ¡ã 4Â÷»ê¾÷Çõ¸í=SWÇõ¸í ¡â GE -1892³â ¼³¸³µÈ ÀüÅëÀÇ Á¦Á¶¾÷ü -Ç×°ø±â ¿£ÁøºÐ¾ß ¼¼°è 1À§±â¾÷(M/S 60% Àå¾Ç) -Ç×°ø±â ¿îÇ× µ¥ÀÌÅÍ ¼öÁý ¹× ¼öÁý µ¥ÀÌÅÍ ºÐ¼®
-ºòµ¥¾îÅÍ È°¿ë ¿£Áø À¯Áöº¸¼ö, ¿¹Áö º¸Á¸, ¿¬ºñ ±Ø´ëÈ, ÃÖÀûÇ×·Î ºÐ¼® µî “SÃãÇü ¼ºñ½º Á¦°ø
¤Ñ 2010³âºÎÅÍ ¿¬°£ 5¾ï´Þ·¯ ÀÌ»óÀ» SWºÎ¹®¿¡ ÅõÀÚ -¼ÒÇÁÆ®¿þ¾î °³¹ßÀÚ 1¸¸5000¿©¸í È®º¸ -¼¼°è ÃÖ´ë ¼ÒÇÁÆ®¿þ¾î±â¾÷ ±¸±Û(2¸¸3000¿©¸í)ÀÇ 65% ¼öÁØ
¤Ñ2014³â °¡Àü/±ÝÀ¶ºÎ¹® ¸Å°¢ ¤Ñ2015³â "¿ì¸®´Â ´õ ÀÌ»ó Á¦Á¶È¸»ç°¡ ¾Æ´Ï´Ù" ¼±¾ð ¤Ñ Jeff Immelt, Chairman and CEO of GE "If you went to bed last night as an industrial company, you're going to wake up today as a software and analistics company"
"We believe that every industrial company will become a software company....
¡â 4Â÷»ê¾÷Çõ¸íÀÇ ½Ç»óÀº "SW Çõ¸í" ¤Ñ "ÀÚµ¿Â÷´Â ÀÌÁ¦ ±â¸§ÀÌ ¾Æ´Ï¶ó ¼ÒÇÁÆ®¿þ¾î·Î ´Þ¸°´Ù" - µðÅÍ Ã¼Á¦(¸Þ¸£¼¼µ¥½º º¥Ã÷ ±×·ì ȸÀå) 2012³â CES±âÁ¶ ¿¬¼³
¡â Software is Eating the World ¤ÑMarc Andreessen - American entrepreneur, investor, and software engineer. Co-author of Mosaic (the first widely used Web browser), co-founder of Netscape, co-founder and general partner of Silicon Valley venture capital firm Andreessen Horowitz. - In the future every company will become a software company
¡â Software Àη ¼ö±Þ ¤ÑBy 2020, an estimated one million computer programming jobs in the U.S. will go unfilled (The Wall Street Journal, 2011) - ÇÏÁö¸¸ ½ÇÁ¦ ÄÄÇ»ÅÍÀü°øÇлýÀº 40¸¸¿© ¸í¿¡ ºÒ°úÇÏ´Ù.
¡ã AI Is Going to Eat Software ¤Ñ Nvidia CEO Jensen Huang: Software Is Eating the World, but AI Is Going to Eat Software -Jensen Huang predicts that health care and autos are going to be transformed by artificial intelligence. - MIT Technology Review, 2017(Nvidia: ÄÄÇ»ÅÍ¿ë GPU¿Í SoC ÀåÄ¡ Á¦Á¶»ç)
¡ã AI is eating SW, and thus the World ¡ã 4IR±¸Çö Çٽɱâ¼ú
AI(Data-based learning) ¡è Big Data(Capture, strage, analysis of data) ¡è IoT(Data Collection through IoT)
-¼ö¸¹Àº ¼¾¼ µîÀÇ µð¹ÙÀ̽ºµéÀÌ ´ë±Ô¸ð µ¥ÀÌÅ͸¦ »ý¼º ¡æ H/W ¹× IoT Sensor ÀÇ ¹ß´Þ - »ý¼ºµÈ µ¥ÀÌÅÍ´Â ´Ù¾çÇÑ ³×Æ®¿öÅ©¸¦ ÅëÇØ Ã¤Áý ¹× Àü¼ÛµÊ¡æ Wi-Fi /Bluetooth /Zigbee/LoRa/5G -Ŭ¶ó¿ìµå µîÀÇ ÀúÀå¼Ò¿¡ ¼öÁýµÈ ¼ö ¸¹Àº µ¥ÀÌÅÍ´Â ºòµ¥ÀÌÅ͸¦ Çü¼º¡æCloud Storage/Data Center -¼öÁýµÈ µ¥ÀÌÅ͸¦ ºÐ¼®, ó¸®ÇÏ¿©Áö´ÉÈµÈ ¼ºñ½º¸¦ Á¦°ø¡æ Big Data ºÐ¼® ¡æ AI
¡ß Big Data
¡ã What is Big Data? ¤ÑExtremely large data sets that may be analyzed computationally to reveal patterns, trends, and associations, especially relating to human behavior and interactions (Dictionary) ¤Ñ High-volume, high-velocity, and high-variety information assets that demand cost-effective, innovative forms of processing that enable enhanced insight, decision making, and process automation (Gartner) ¤Ñ µ¥ÀÌÅͷκÎÅÍ °¡Ä¡¸¦ ÃßÃâÇÏ°í °á°ú¸¦ ºÐ¼®ÇÏ´Â ±â¼ú (À§Å°¹é°ú) ¤Ñ »õ·Î¿î µ¥ÀÌÅÍ ¼Ò½º¿¡¼ ³ª¿Â ´õ Å©°í ´õ º¹ÀâÇÑ µ¥ÀÌÅÍ ¼¼Æ® (¿À¶óŬ)
¡ã Big Data Characteristics (5 V¡¯s fo Big Data)
¡ã How Big? ¤Ñ 2.5 Quintillion (1018) bytes of data is created every day by 2020 it will equal 40 Zetta (1021) bytes
¡ã Key Enablers of Big Data ¤Ñ Increase in storage capacity Low cost storage to store data ¤Ñ Increase in processing power Powerful multi-core processors ¤Ñ Availability of data From everywhere (SNS, mobile devices, IoT sensors, ¡¦.)
¡ã The First Big Data Challenge – Census ¤Ñ 1880 census 50 million people(1880³â´ë 5000¸¸¸í ¹Ì±¹ ±¹¹Î¿¡ ´ëÇÑ Àα¸ Åë°è Á¶»ç¿¡ 8³â ¼Ò¿ä) Age, gender, occupation, education level, # of insane people in household Requires 8 years.
¤Ñ Herman Hollerith Tabulating Machine Used punched cards for 1890 census 6 months instead of 8 years Saved more than $5 million budget ¤Ñ In 1896 Hollerith organized the Tabulating Machine Company In 1911, TMC merged with 4 companies to form Computing-TabulatingRecording Company (CTR) In 1924, CTR was renamed International Business Machines Corporation under the presidency of Thomas J. Watson
¡ã Google Flu Trends ¤Ñ First launched in 2008 by Google to help predict outbreaks of flu By aggregating Google Search queries, it attempted to make accurate predictions about flu activity Investigates frequency of search query, location, popularity change over time, related search Detects flu activity two weeks before CDC (Centers for Disease Control and Prevention)
¡ã Data is the New Oil ¤Ñ Oil wells ¡æ Oil Production ¡æ Rfinery ¡æ Transportation ¡æ Pump
¤Ñ Data Sensing and generation ¡æ Data collection ¡æ Data Analytics ¡æ Data center & Storage ¡æ Intelligence of Things ¡ã Big data analytics ¤Ñ The process of inspecting, cleaning, transforming, and modeling big data with the goal of discovering useful information, suggesting conclusions, and supporting decision making ¤Ñ Big data analytics examines large and different types data to uncover hidden patterns, correlations and other insights.
¡ß Artificial Intelligence
¡ã What is AI ¡â Turing Test -Alan Turing, "Computing Machinery and Intelligence¡±, 1950. I propose to consider the question, ¡°Can machines think?¡± A test of a machine's ability to exhibit intelligent behavior equivalent to that of a human - Imitation Game Does not check the ability to give correct answers to questions, only how closely answers resemble those a human would give. If the evaluator cannot tell the machine from the human, the machine is said to have passed the test.
¡â Definition - The capacity given by humans to machines to memorize and learn from experience, to think and create, to speak, to judge and make decision. - An area of computer science that aims to create intelligent machines that work and react like humans.
¡â Learning A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, improves with experience E. - Tom Mitchell, "Machine Learning¡±, Tom Mitchell, McGraw Hill, 1997.
Learning is nothing but updating confidence from prior to posterior - Key factor is the likelihood based on data
¡â Programming vs. ML ¤Ñ<Traditional Programming> - programs are created manually by humans - by providing input data and based on the programming logic, computer generates the output.
¤Ñ<Machine Learning> - data and answers (or labels) go in as input - the learned rules (models) come out as output - it lets computer to learn new rules (that predicts future outcomes)
¡â Machine Learning Workflow • Data Acquisition - data is gathered/collected from various sources • Data Preparation - data is cleaned, preprocessed, and eventually becomes dataset - removing errors, mistakes, duplicates, and inconsistencies in data - data integration - combining data from different sources • Model development & Training- data is patternized and generalized as models
¡â Supervised Learning
Classification - process of finding a function which helps in dividing the dataset into classes based on different parameters. - used to predict/Classify the discrete values such as Male or Female, True or False, Spam or Not Spam, etc.
Regression - process of finding correlations between dependent and independent variables. - helps in predicting continuous variables such as prediction of Market Trends, prediction of House prices, etc.
¡â Unsupervised Learning Clustering - points with high similarity are grouped together in clusters to identify patterns in data. Dimensionality reduction - tries to transfer highdimensional data into efficient low-dimensional representations by finding appropriate transformations ϕ.
¡â Reinforcement Learning ¤Ñ Learns through repeated trial-and-error interactions with a dynamic environment. ¤Ñ To make a series of decisions that maximize a reward metric for the task without human intervention
¡â Machine Learning vs. Deep Learning
Information is learnt step by step - First layers focus on learning more specific concepts while deeper layers use the information already learnt to soak in more abstract concepts. This procedure of constructing representations of the data is known as feature extraction......(Ý»ÝÂó¢ùê)
¡ã What can AI do?
¡â So many Applications
¡âObject Recognition ¤Ñ Computer vision technique for identifying objects in images or videos
¡âHandwriting Recognition ¤Ñ Examples of digits from the MNIST database(a collection of 70000 handwritten digits)
¡â Sentiment Analysis (°¨Á¤ ºÐ¼®) ¤Ñ A natural language processing technique used to determine whether data is positive, negative or neutral.
¤Ñ ¡°This movie should have NEVER been made. From the poorly done animation, to the beyond bad acting. I am not sure at what point the people behind this movie said "Ok, looks good! Lets do it!" I was in awe of how truly horrid this movie was.¡± Positive or Negativ
¡âGo Playing - Google AlphaGo (2016) ¤Ñ A computer program that plays the board game Go. It was developed by DeepMind Technologies, which was later acquired by Google. Won 4-1 victory over Go grandmaster Lee Sedol
¡âLanguage Translation Services & Search by Image ¤Ñ Google Translate & Google Image
¡â Visual Search ¤Ñ Google Lens ¡°Search What You See¡±
¡â Music Recognition ¤Ñ Shazam identifies songs based on an audio fingerprint based on a time-frequency graph called a spectrogram.
¡â Personal (voice, virtual) Assistant ¤ÑApple Siri, Amazon Alexa, Google Assistant, Samsung Bixby A software agent that can perform tasks or services for an individual based on commands or questions Speech recognition and language understanding
¡âSelf-driving (Driverless) Cars ¤Ñ A vehicle that is capable of sensing its environment and moving safely with little or no human input Employ a wide range of technologies like radar, cameras, ultrasounds, and radio antennas to navigate safely on the road, are used in conjunction with one another. Tesla¡¯s driverless car technology (known as ¡°Autopilot¡±) uses 8 cameras to provide 360-degree visibility, while 12 ultrasonic sensors and a front-facing radar work to analyze the vehicle¡¯s surroundings for potential hazards.
¤Ñ However, one key component still in development that will ultimately make autonomous cars more reliable is the implementation of 5G cellular networks.
¡â¹Ì·¡ ÀΰøÁö´É ½Ã´ëÀÇ Àΰ£ÀÇ ¸ð½ÀÀº? ÀÎÁ¶¿©¼º, Å×ÀÌÅ;çÀÌ ¾Æ³»º¸´Ù ´õ .................
¡ß Survival of the fittest!
AI, Àηù¸¦ ¸ê¸Á½Ãų ±â¼úÀΰ¡? - ¼±ÇÑ AI or ¾ÇÇÑ AI? • To do or not to do? • ¿ì¸®³ª¶ó ±×¸®°í °æºÏÀÇ ¹Ì·¡?
|