We live in an ocean of data. Big Data is characterized by vast amounts of data sized in the order of petabytes or even exabytes. Though Big Data has great potential, Big data by itself has no value unless one can derive meaningful results from it. That is where Artificial Intelligence pitches in. Artificial Intelligence’s most common application is about finding patterns in enormous quantities of data. The confluence of Big Data and Artificial Intelligence allows companies to automate and improve complex descriptive, predictive and prescriptive analytical tasks. In other words, Big Data can offer great insights with the help of Artificial Intelligence (AI). Artificial Intelligence can act as a catalyst to derive tangible value from Big data and serve as key to unlocking Big data. This review article focuses on applications of artificial intelligence to Big Data, its Limitations and issues.
In today’s digital world, data has grown ‘big’ – steering in the era of the petabytes and exabytes. Big Data is characterized by astronomical amounts of data being generated continuously by interconnected systems of people, transactions, media, devices (sensors, smartphones, smart meters, cameras and tablet computers) -- click data, audio/speech data, natural language text (in multiple languages), images/video data. The growth of Big data is a result of the wide variety of data and growing channels in today’s world. The internet way of things has a significant contribution in the growth of Big data. By 2015, research firm IDC predicts there will be more than 5,300 exabytes of unstructured digital consumer data stored in databases, and we expect a large share of that to be generated by social networks. Facebook ingests approximately 500 times more data each day than the New York Stock Exchange (NYSE). Twitter stores at least 12 times more data each day than the NYSE [Smith, 2014]. The challenge is to analyze the information content in these vast, continuous data streams, use them for descriptive and predictive analytics in various domains and build more robust and intelligent learning systems. With big data benefitting from improved and increased storage capabilities at extremely reasonable prices - the cost of a gigabyte of storage has dropped from approximately $16 in February
2000 to less than $0.07 today [SAS, 2012], and with processing technologies specifically designed to handle huge data volumes, thinking moves away from what data/records to keep and store, to muse over the problem of how to make sense and derive logic from these increasing large volumes of data Yvonne Hofstetter, managing director of Teramark Technologies GmbH, a Germany-based provider of big data technologies and artificial intelligence for the industrial Internet states that the core of big data is the analysis of big data and the inference, which is provided by artificial intelligence (AI) and not storage or retrieval of raw data [O'Dwyer, 2014]. Big Data can offer great insights with the help of Artificial Intelligence (AI).
Artificial Intelligence deals with the study and development of software and machines that can imitate human-like intelligence and it is a branch of computer science that is extremely technical. Artificial intelligence is used in a variety of ways and can be found across a large number of industry sectors-manufacturing, life sciences and healthcare, transportation, and healthcare, finance to name a few. Some examples of its usage are in assembly line robots, advanced toys, online search engines, speech recognition systems, medical research, and marketing. Artificial Intelligence’s most common application is about finding patterns in enormous quantities of data. Smaller more homogenous fixed data sets will not serve the purpose as the patterns may not be evident in them. This allows companies to automate and improve complex descriptive, predictive and prescriptive analytical tasks, which would be tremendously labor intensive and time consuming if carried out by humans beings.
The aim of this paper is to explore the opportunities of Big Data focusing on applications of artificial intelligence to Big Data problems. The paper begins with a brief overview of Big data and the characteristics of Big data followed by a sections discussing the application of Artificial Intelligence to Big data and limitations and issues of Big data and Artificial Intelligence.
Big Data: Definition
The use of the term “big data” can be traced back to discussions of handling huge groups of datasets in both academia and industry during the 1980s [Yan, 2013]. Michael Cox and David Ellsworth were among the first to use the term big data literally, referring to using larger volumes of scientific data for visualization (the term large data also has been used) [ Cox and Ellsworth, 1997].
The first formal, academic definition appears in a paper submitted in July 2000 by Francis Diebold of University of Pennsylvania in his work of econometrics and statistics (2000):
Big Data refers to the explosion in the quantity (and sometimes, quality) of available and potentially relevant data, largely the result of recent and unprecedented advancements in data recording and storage technology. In this new and exciting world, sample sizes are no longer fruitfully measured in “number of observations,” but rather in, say, megabytes. Even data accruing at the rate of several gigabytes per day are not uncommon.
Wikipedia defines big data as a collection of data sets so enormous and complex that it becomes challenging to process using on-hand database management tools or traditional data processing applications [Press, 2014].
Gartner defines big data as follows:
Big data is high-volume, high-velocity and high-variety information assets that demand cost-effective, innovative forms of information processing for enhanced insight and decision making. [“Gartner IT Glossary,” n. d.; Lapkin 2012]
SAS defines big data as follows:
Big data is a relative term describing a situation where the volume, velocity and variety of data surpass an organization’s storage or compute capacity for accurate and timely decision making. [SAS, 2012]
Tech-America Foundation defines big data as follows:
Big data is a term that describes large volumes of high velocity, complex and variable data that require advanced techniques and technologies to enable the capture, storage, distribution, management, and analysis of the information. [Tech-America Foundation’s Federal Big Data Commission, 2012]
Big Data Characteristics
| Figure 1. Big data characteristics |
|---|
|
IBM suggested that big data could be characterized by any or all of three “V” words to investigate situations, events, and so on: volume, variety, and velocity (Zikopoulous et al., 2013).
Perhaps it is best to think of Big Data in multidimensional terms, in which five dimensions or characteristics relate to the primary aspects of Big Data. Figure 1 depicts the characteristics, that is, five V’s of Big Data.
The big data characteristics can be elaborated as follows:
Definitions of big data volumes are relative and fluctuate by factors, such as time and the type of data. What may be considered big data today may not meet the threshold in the future because storage capacities will increase, allowing even bigger data sets to be captured and stored. In addition, the type of data, discussed under variety, delineates what is meant by ‘big’. Two datasets of the same size may necessitate different data management technologies based on their type, e.g., tabular versus video data [Gandomi & Haider, 2015].
Often time sensitive, Big Data must be used as it is streaming into the enterprise in order to maximize its value to the business, but it must also still be available from the archival sources as well [Ohlhorst, 2012]. For example, there are more than 250 million tweets per day [Pingdom, 2012]. Tweets lead to decisions about other Tweets, intensifying the velocity [O'Leary, 2013].
McKinsey (2011) projected in an industry report that five new kinds of value might come from big data (McKinsey Global Institute, 2011):
A data environment can become extreme [SAS, 2012] along any of the dimensions/characteristics or with a combination of two or more of them at the same time. Also, the relativity of big data volumes discussed formerly applies to all dimensions. Thus, universal benchmarks do not exist for volume, variety, and velocity that describe big data. The defining limits are influenced by the size, sector, and location of the firm and these limits evolve over time [Gandomi & Haider, 2015].
Also, the dimensions are not independent of one another. As one dimension changes, the probability increases that another dimension will also change as a result. However, a ‘three-V tipping point’ exists for every organization beyond which traditional data management and analysis technologies are no longer for deriving timely intelligence. The Three-V tipping point is the threshold beyond which organizations start dealing with big data. The organizations should then trade-off the future value expected from big data technologies against their implementation costs [Gandomi & Haider, 2015].
Applying Artificial Intelligence (AI) to Big Data
Unleashing AI on big data can have a noteworthy influence on the role data plays in conducting business, analytics and decision making. Big data is not component of Artificial intelligence [Umbler Corp., 2015]. However, the two are entwined: AI provides the large-scale analytics necessary to extract meaning and value from big data, while big data provides the knowledge required for AI to continue to learn and evolve — or to become more intelligent [Umbler Corp., 2015]. In other words, AI offers the technology and methodology for better understanding of the ever-growing amounts of data. Artificial intelligence solutions are by nature multi-disciplinary, encompassing computer science, mathematics, statistics, philosophical thinking and the industry sector the problem corresponds to. Some of the applications of artificial intelligence have origins in academia, while others have their origins in the research divisions of private companies or even individuals that managed to market their products.
The big data market has been maturing for years now. There is a plethora of technology that can crunch the numbers and spit them out in a spreadsheet or chart. Now, entrepreneurs are beginning to fill this gap with technology that not only synthesizes the data, but interprets it, too [Lapowsky, 2014]. Some of them are as follows [Lapowsky, 2014, Rijmenam, 2014]:
Like Big Data, AI is about exploding volumes, velocities and variety of data. Under situations of astronomical volumes of data, AI permits delegation of difficult pattern recognition, learning, and other tasks to computer-based approaches. Effective machine language translation is statistics-based, benefitting from the availability of huge data sets. For example, more than one-half of the world’s stock trades are done using AI-based systems. In addition, AI lends to the velocity of data, by assisting rapid computerized decisions that lead to other decisions. For example, since so many stock trades are made by AI-based systems rather than people, the velocity of the trades can increase, and one trade could lead to others. Finally, variety issues aren’t solved simply by parallelizing and distributing the problem. Instead, variety is mitigated by capturing, structuring, and understanding unstructured data using AI and other analytics [O'Leary, 2013].
Social media big data has a huge potential locked in unstructured data, which comprises of the billions of user-generated written posts, pictures, and videos that circulate on social media. However, only a small fraction of its potential is currently being realized. Artificial intelligence is quickly changing the way social big data is mined for insights and used in emerging marketing applications. Keeping up with the massive volume of messages, photos, and videos that consumers upload and share on social networks every day is a task that only automated intelligent machines can deal with. The task would simply be beyond scope and capacity of human-directed, manual systems. Artificial intelligence, and in particular, a subset of AI known as “deep learning,” is key to social media’s future as an industry and as a force in society. AI systems are capable of not just computing, but actually learning — machine learning, along with a subset of machine learning called deep learning. Machine learning is aspect of AI and refers to the capability of machines to learn and progress through exposure to new data [Umbel Corp., 2015]. With deep learning, machines themselves figure out which rules to follow based on data researchers feed them. Deep learning is only possible with big data, because an enormous quantity of data is needed to “teach” AI systems. The other component necessary for deep learning is the algorithmic power to make sense of all that data [Umbel Corp., 2015]. The ability of Deep Learning to extract high-level, complicated abstractions and data representations from large sizes of data, specifically unsupervised data, makes it attractive as a valuable tool for Big Data Analytics. More specifically, Big Data problems such as semantic indexing, data tagging, fast information retrieval and discriminative modeling can be better addressed with the help of Deep Learning. More traditional machine learning and feature engineering algorithms are not competent enough to extract the complex and non-linear patterns normally witnessed in Big Data. By extracting such features, Deep Learning facilitates the use of comparatively simpler linear models for Big Data analysis tasks, such as classification and prediction, which is significant when developing models to deal with the scale of Big Data [Najafabadi et al., 2015].
Machine-learning algorithms can be applied to streaming data. High volume data streams (high speed continuous flow of data) arise in numerous settings, like IT operations, sensors, and social media to name a few. Medical domains include many settings where data is produced in a streaming fashion, such as anatomical and physiological sensors, or incidence records and health information systems. Xmp from Argyle Data consist of algorithms for online learning and real time pattern detection. Feature Stream is a new web service for applying machine-learning to data streams. Yahoo! recently released SAMOA—a distributed streaming machine-learning framework. SAMOA lets developers code algorithms once and execute them in multiple stream processing environments [O’Reilly, 2015].
Consumer Internet companies are in a hurry to build out their AI talent and acquire the most advanced machine-learning systems. Some of the major acquisitions from the AI field that occurred in recent months [Cooper, 2014]:
Artificial Intelligence and Big data can be used for fraud management. Neural networks are analytics that learn to distinguish complex patterns of behavior (in customer transactions, network activity, etc.). This analytic technique imitates how neurons in the brain store knowledge in their connections to other neurons and how the strengths of those synapses change with different kinds of mental activity. Neural networks have been widely deployed in fraud management because they excel at swiftly spotting abnormal data patterns within large amounts of transaction data. Features can be thought of as super-variables, as they are variables which have been engineered (selected, combined, calculated) by an analytics expert to make them highly predictive. In fraud detection, a feature might be the number of card transactions in the last hour, or the number of transactions over a specific amount during the last 12 hours between the hours of 3am and 6am compared to overall dollar volume [Najafabadi et al., 2015].
In development, a neural network undergoes supervised training, in which it evaluates enormous volumes of historical data labeled with an outcome. By analyzing months of cardholder transaction data, for example, a fraud model’s hidden layer incrementally learns the sometimes complex and understated feature relationships connected with the outcome of a fraudulent transaction. It changes the “weights” (relative strengths) of these connections to recognize the weighted feature relationships that best foresees that outcome. A deployed neural network model trained this way can instantaneously detect predictive patterns in new data as huge volumes of it stream in (e.g., from cardholder transactions, network activity and sensor readings). When deployed with dynamic profiles, an associated analytic technique, the model also learns the typical patterns for individual entities (e.g., a specific cardholder, merchant or ATM), and so can spot suspicious deviations [Najafabadi et al., 2015].
At Griffith’s School of ICT, Australia, interdisciplinary research using Artificial Intelligence technology in the area of Coastal management has generated camera-based technology for analyzing enormous amounts of digital video images of beach scenes to count and comprehend the behavior of persons on the shore along Australia’s coast. This technology is accurate to 80-90% and can discover, from very low-resolution video, whether a person is walking or running along the beach or entering the ocean. The application of this work can help in beach safety and support the work of lifeguards to detect swimmers in danger. Another research work being undertaken involves merging the fields of IT and Engineering for forecasting the deterioration of bridges in Queensland (10,000 in total) using Artificial Neural Network (ANN) technology. This research work, undertaken in collaboration between the City of Gold Coast and the Queensland State Government, has created predictive models trained using historical bridge data to guarantee the safety and effective maintenance of important State assets. The research in this area has generated effective models to forecast the deterioration of bridges up to 20-30 years into the future with the capability to save millions of dollars through prioritization of asset maintenance regimes. Another interdisciplinary research, which intersects information technology and the environment, has fashioned accurate and sophisticated AI models for the prediction of flood events on the Gold Coast. Using a assortment of data from sensors located at several catchments around the Gold Coast, rainfall data and other complementary information, ANNs have been trained to understand historical trends and are able to perform predictive analytics for forecasting flood events up to three hours ahead, which provides adequate time for undertaking effective disaster management. This work being undertaken at Griffith’s School of ICT is scalable and can be extended across the State and indeed the Nation [Griffith Sciences].
Social networking experiences are becoming increasingly centered around photos and video [Cooper, 2014].
However, it is extremely difficult to extract information from visual content. Because of this, image and video recognition are two of the more exciting disciplines being worked on in the field of AI and deep learning. Advances in “deep learning,” cutting-edge AI research that tries to program machines to perform high-level thought and abstractions, are enabling marketers to extract information from the billions of photos, videos, and messages uploaded and shared on social networks each day. Image recognition technology is now advanced enough to identify brand logos in photos. Web giants such as Google and Baidu (a China-based search engine) are using a deep learning-derived technique known as “back-propagation,” in order to classify images in user photo collections. Back-propagation is a method for training computer systems to match images with labels or tags that were pre-defined by users. For example, if enough people upload photos tagged “cat,” then a system has a large enough sample size that it can reference to identify new photos of cats, and tag them appropriately. This is one of the reasons why services such as Facebook and Instagram encourage users to tag objects and people in photos. However, there is very little user-generated data identifying the contents of online video, which so far means that back-propagation is a poor method for video recognition [Smith, 2014].
Speech recognition is another field where machine learning can be applied to big data. Speech recognition is the transformation of spoken words into text. At Baidu, the Chinese-based search giant, the objective is to have mobile phone software accurately transcribe words in languages such as English or Mandarin and understand the request. With its Deep Speech system, which first came out in December 2014, Baidu trained it on more than 100,000 hours of voice recordings, first getting people to read to the machine and then adding synthesized noise over the top to make it sound like they are talking in a noisy room, cafeteria, or car. Then it let the system learn to recognize speech even amid all that noise. Baidu’s Deep Speech system uses deep learning algorithms that are trained on a recurrent neural network, or simulation of connected neurons [Hof, 2014; Merrett, 2015]
Text mining is another field that is quickly evolving. A team of Belgian computer science researchers developed an opinion mining algorithm that can identify positive, negative, and neutral sentiment in Web content with 83% accuracy for English text. Accuracy vacillates depending on the language of the text because of the diversity of linguistic expressions. The more complex a language is, the more training a machine-learning system needs [Smith, 2014].
Particle Swarm Optimization (PSO) is a computational procedure used in data mining to competently harvest useful information from big data, by repetitively improving candidate solutions to optimize a given problem. These candidate solutions (also called particles) can be moved around a search space through simple mathematical formula. These particles follow flocking rules to form swarms, moving the swarm towards solutions, in due course allowing a particle to find a position that surpasses minimal requirements given to a solution [Dervojeda, 2013].
Agent-based computational economics studies dynamic systems of interacting agents and can be used to model economic processes and even entire economies, through agent-based models. In these models, the interaction of agents is modelled according to rules that model behaviour and social interactions, which permits for behavioural forecasting and price movements. This methodology necessitates involvement of experts in order for the rules and models to correctly reflect reality, which then could tap into big data to cater for dynamic, real time analysis and forecasting [Dervojeda, 2013].
From digital advertisements to landing pages, marketing content has become progressively more challenging to create, fine-tune and manage so customers receive the most appealing and effective messages. What makes content creation and management challenging is the increasing number of channels and the volume of content that companies must deal with today. AI-based solutions are available to help the creation and deployment of the most effective, targeted content possible, often in near real time [Umbel Corp., 2015]. Some examples are:
Artificial Intelligence systems are helping marketers and advertisers garner insights from the vast ocean of unstructured consumer data collected by the world’s largest social networks. Audience targeting and personalized predictive marketing using social data are expected to be some of the business areas that benefit the most from mining big data. IBM Watson Engagement Advisor, a cognitive computing assistant, can help a business to better serve its consumers. IBM Watson Engagement Advisor permits brands to crunch big data and transform the fashion that they engage clients in key functions, such as customer service, marketing, and sales. IBM Watson Engagement Advisor can learn, adapt, and gain an understanding of a company’s data, enabling customer-facing personnel to assist consumers with deeper insights faster than previously possible [Takeda & Onodera, 2013]. Watson uses natural language processing to comprehend words and their relationships. It also depends on cognitive computing techniques to sift through more than 200 million-plus pages of structured and unstructured data that consume approximately 4 terabytes of disk space. Among other things, Watson can mine posts on Twitter to better recognize market trends and brand sentiment [Greengard, 2014]. According to Booz & Company, 61% of data professionals say big data will renovate marketing for the better. However, although machine-learning systems are being tasked with analyzing users on an individual basis, the value to marketers is grouping like-minded consumers together, so that they can target people at scale [Smith, 2014].
Systems based on Artificial Intelligence have an ability to evaluate massive and growing stores of customer data and apply the results to enhance customer experiences. The ability to suggest a product that specifically fits a customer’s need right at the moment of decision is the holy grail of marketing for businesses ranging from fashion and beauty companies to media and entertainment providers — and from retailers to etailers. To achieve this ability, necessitates moving away from traditional systems that filter products based on previous purchases [Umbel Corp., 2015]. Using data from multiple sources, AI can construct a store of knowledge that will in due course of time, enable accurate predictions about you as a customer which are based not just on what you purchase, but on how much time you spend in a specific part of an site or store, what you look at while you are there, what you do purchase compared with what you don’t — and a multitude of other bits of data that AI can synthesize and add to, ultimately getting to know you and what you want extremely well [Umbel Corp., 2015]. For instance, if a woman accesses a retailer's Web site and looks at a dress, the Web site might record the length of time she spent at the page, what color options she eyed, and compare the clicks to other Web surfing activity--as well as other variables ranging from her overall purchase patterns to past purchases--to identify exactly what triggers a purchase. If she leaves an item in her shopping cart, she might receive a promotion a couple of days later with a free shipping coupon or a certain percentage off coupon code or, perhaps, a gift-with-purchase offer. Moreover, the data might be used to customize pricing in the future and adjust promotions so that she is more likely to close the purchase immediately. Of course, different customers would receive diverse experiences and offers based on their exclusive data fingerprints [Greengard, 2014]. All of this is done by AI software. In short, companies are using AI based recommendation engines that make suggestions based on everything they know about a shopper, not just what they bought before. Companies that are already using AI and big data for product recommendations include [Umbel Corp., 2015]:
Limitations and Issues of Big Data and Artificial Intelligence
Irrespective of how powerful and how much value big data brings, it has its limitations. Big data has its distinctive characteristics that can provide decision-makers with more timely, rich information, but without the specific context, data on its own can never tell the whole story. Like any data, big data is not a panacea to solve all questions for all organizations. An organization should consider at least the “Three V’s” of its data and its practical capacities before implementing big data technologies. Also big data cannot replace traditional analytics. With its complexity and its requirements for technology and talent, big data is more appropriate to organizations with large-scale, multi-structured datasets [Yan, 2013].
With big data there will also be dirty data, with potential errors, incompleteness, or differential precision. AI can be used to identify and clean dirty data or use dirty data as a means of establishing context knowledge for the data. For example, “consistent” dirty data might indicate a different context than the one assumed—for example, data in a different language [O'Leary 2013]. In short, the adage Garbage in, garbage out (GIGO) applies when AI is applied to dirty data.
Current AI algorithm sets are often non-standard and primarily research-based. Algorithms might lack documentation, support, and clear examples. Further, historically the focus of AI has largely been on single-machine implementations. With big data, we now need AI that’s scalable to clusters of machines or that can be logically set on a MapReduce structure such as Hadoop. MapReduce allows the development of approaches that can handle larger volumes of data using larger numbers of processors. As a result, some of the issues caused by increasing volumes and velocities of data can be addressed using parallel-based approaches. As a result, effectively using current AI algorithms in big data enterprise settings might be limited. First, unfortunately, the nature of some machine-learning algorithms—for example, iterative approaches such as genetic algorithms—can make their use in a MapReduce environment more difficult. However, recently, MapReduce has been used to develop parallel processing approaches to AI algorithms. Hadoop (http://hadoop.apache.org), named after a boy’s toy elephant, is an open source version of MapReduce [O'Leary 2013].
Artificial Intelligence and Big data mesh well. Artificial Intelligence and Big data when put together opens up innumerable opportunities for resolving the problems faced in modern society in 21st century and beyond. Big data is about the stunningly increasing growth of data, and companies can gain competitive advantage by better understanding the ever-growing amounts of data. Artificial intelligence offers the technology and methodology to do so. The combination of Artificial Intelligence based analytics and Big Data is stimulating because it increases the likelihood of better and faster information extraction from large-scale data [FICO, 2014].
Since big data are noisy, greatly interrelated, and unreliable, it will likely lead to the development of statistical techniques more readily apt for mining big data while remaining sensitive to the unique characteristics [Gandomi & Haider, 2015]. The quality issues in the data, needs to be either tackled at the data pre-processing stage or by the learning algorithm [Zhou, 2014]. Going beyond samples, additional valuable insights could be obtained from the colossal volumes of less ‘trustworthy’ data [Gandomi & Haider, 2015].
Continued innovations in AI are certainly promising and merit renewed interest in the discipline. While none of these techniques are a “silver bullet”—and human analytic and business domain expertise remain indispensable—they can improve the return on Big Data analytic investments in numerous important ways [FICO, 2014]. To make the best out of AI and big data, big data needs to be aligned with a specific business goal or a well-defined problem statement. Gains from Big Data and AI will be governed by right analytic practices, clear understanding of business requirements that is- what is AI aiming to achieve from the data, relevant data and active involvement and collaboration amongst stakeholders - data scientist, domain/subject matter experts, business and data analysts) and top management support for the endeavor.
This research was previously published in the Handbook of Research on Computational Intelligence Applications in Bioinformatics edited by Sujata Dash and Bidyadhar Subudhi; pages 1-16, copyright year 2016 by Medical Information Science Reference (an imprint of IGI Global).
CoxM.EllsworthD. (1997). Managing Big Data for Scientific Visualization. Proc. of ACM Siggraph (pp. 5-1–5-17).
Cukier, K. (2010). Data, data everywhere: A special report on managing information. The Economist. Retrieved from http://www.economist.com/node/15557443
Dervojeda, K., Verzijl, D., Nagtegaal, F., Lengton, M., & Rouwmaat, E. (2013, September). Big Data Artificial Intelligence, Business Innovation Observatory Contract No 190/PP/ENT/CIP/12/C/N03C01. European Union. Retrieved from http://ec.europa.eu/enterprise/policies/innovation/policy/business-innovation-observatory/files/case-studies/09-bid-artificial-intelligence_en.pdf
Diebold, F.X. (2000, August). Big data dynamic factor models for macroeconomic measurement and forecasting. Presented at the 8th World Congress of the Econometric Society, Seattle. Retrieved from http://www.upenn.edu/~fdiebold/papers107/ABCD_HOED.pdf
Dumbill, E. (2012, January 11). What is big data? An introduction to the big data landscape. O'Reilly Radar.
FICO. (2014). Does AI + Big Data = Business Gain? Insights White Paper. Retrieved from http://www.fico.com/en/latest-thinking/white-papers/insightsdoes-ai--big-data--business-gain
Fortuna's Corner. (2014). Ten Big Data Start-Ups To Watch. Retrieved from http://fortunascorner.com/2014/04/17/ten-big-data-start-ups-to-watch/
Gandomi, A., & Murtaza, H. (2015). Beyond the hype: Big data concepts, methods, and analytics International Journal of Information Management , 35(2), 137–144. doi:10.1016/j.ijinfomgt.2014.10.007
Big Data. (n. d.). Gartner IT Glossary. Retrieved from http://www.gartner.com/it-glossary/big-data/
Gens, F. (2011). IDC 2012 Predictions: Competing for 2020. Retrieved from http://cdn.idc.com/research/Predictions12/Main/downloads/IDCTOP10Predictions2012.pdf
Greengard, S. (2014). Artificial Intelligence: Ready To Live Up To Its Hype? Insight/ Market Research, CMO by Adobe. Retrieved from http://www.cmo.com/articles/2014/11/5/artifical_intelligence.html
Sciences, G. (n. d.). Deep Learning from Big data. Impact @ Griffith Sciences. Retrieved from http://app.griffith.edu.au/sciencesimpact/deep-learning-big-data/
Hof, R. (2014). Baidu Announces Breakthrough In Speech Recognition, Claiming To Top Google And Apple. Forbes. Retrieved from http://www.forbes.com/sites/roberthof/2014/12/18/baidu-announces-breakthrough-in-speech-recognition-claiming-to-top-google-and-apple/2/
Holland, P. (2015). Characteristics of Big Data – Part One. Retrieved from http://makingdatameaningful.com/2015/05/26/characteristics-of-big-data-part-one/
IS0/IEC 2015 ISO/IEC JTC 1. (n. d.). Information technology 2015, Big Data Preliminary Report 2014. Retrieved from http://www.iso.org/iso/big_data_report-jtc1.pdf
Lapowsky, I. (2014). 4 Big Opportunities in Artificial Intelligence. Inc.com. Retrieved from http://www.inc.com/issie-lapowsky/4-big-opportunities-artificial-intelligence.html
Lapkin, A. (2012). Hype Cycle for Big Data . Gartner.
Levine, B. (2015). Persado scores $21M to become the Moneyball of marketing. VB News. Retrieved from http://venturebeat.com/2015/01/22/persado-scores-21m-to-become-the-moneyball-of-marketing/
McKinsey Global Institute. (2011, May). Big data: The next frontier for innovation, competition, and productivity.
Merrett, R. (2015). Intelligent machines part 1: Big data, machine learning and the future. CIO. Retrieved from http://www.cio.com.au/article/576664/intelligent-machines-part-1-big-data-machine-learning-future/
Mitchell, I., & Wilson, M. (2012). Linked data Connecting and exploiting big data (White paper). Retrieved from from www.fujitsu.com/.../Linked-data-connecting-and-exploiting-big-data-(v1.0). pdf
Mongardi, S. (2015). ‘What is the Grid and why it's a revolution. The Web Mate. Retrieved from http://www.thewebmate.com/2015/05/18/what-is-the-grid-and-why-its-a-revolution/
Najafabadi, M., Villanustre, F., Khoshgoftaar, T. M., Seliya, N., Wald, R., & Muharemagic, E. (2015). Deep learning applications and challenges in big data analytics. Journal of Big Data. Retrieved from http://www.journalofbigdata.com/content/2/1/1
O'Dwyer, M. (2014). How companies can make the most of big data. Dell.com. Retrieved from http://techpageone.dell.com/technology/companies-can-make-big-data/
Ohlhorst, F. J. (2012). Big Data Analytics: Turning Big Data into Big Money . Cary, North Carolina, USA: SAS Institute Inc. doi:10.1002/9781119205005
Big Data Now: 2014 Edition. (2015). O’Reilly Media, Inc. United States of America.
Pingdom. (2012). Internet 2011 in Numbers. Pingdom.com. Retrieved from http://royal. pingdom.com/2012/01/17/internet-2011-in-numbers
Press, G. (2014). 12 Big Data Definitions: What's Yours?” Forbes.com. Retrieved from http://www.forbes.com/sites/gilpress/2014/09/03/12-big-data-definitions-whats-yours/
van Rijmenam, M. (2014). Is Artificial Intelligence About To Change Doing Business Forever? Retrieved from http://www.bigdata-startups.com/artificial-intelligence-change-business-forever/
SAS. (2012). Big Data Meets Big Data Analytics (White Paper). SAS.com. Retrieved from http://www.sas.com/content/dam/SAS/en_us/doc/whitepaper1/big-data-meets-big-data-analytics-105777.pdf
Smith, C. (2014). Social Media's New Big Data Frontiers -- Artificial Intelligence, Deep Learning, And Predictive Marketing. Business Insider.com. Retrieved from: http://www.businessinsider.com.au/social-medias-new-big-data-frontiers-artificial-intelligence-deep-learning-and-predictive-marketing-2014-2
Soubra, D. (2012). The 3Vs that define Big Data. Data Science Central.com. http://www.datasciencecentral.com/forum/topics/the-3vs-that-define-big-data
Takeda, K., & Onodera, T. (2013). Artificial Intelligence: Learning Through Interactions and Big Data. IBM Redbooks Point-of-View publication. Retrieved from http://www.redbooks.ibm.com/abstracts/redp4974.html?Open
Demystifying big data: A practical guide to transforming the business of Government. (2012). TechAmerica. Retrieved from http://www.techamerica.org/Docs/fileManager.cfm?f=techamerica-bigdatareport-final.pdf
Corp, U. (2015). AI meets Big Data (White paper). Retrieved from http://etailwest.wbresearch.com/ai-meets-big-data-ml
Yan, J. (2013). Big Data, Bigger Opportunities. Meritalk. Retrieved from http://www.meritalk.com/pdfs/bdx/bdx-whitepaper-090413.pdf
Zhou, Z.-H., Chawla, N. V., Jin, Y., & Williams, G. J. (2014). Big Data Opportunities and Challenges: Discussions from Data Analytics Perspectives. IEEE Computational Intelligence Magazine, 9(4). 62-74.
Zikopoulous, P., Deroos, D., Parasuraman, K., Deutsch, T., Corrigan, D., & Giles, J. (2013). Harness the Power of Big Data . McGraw-Hill.