Tag Archives: IBM

IBM Watson Analytics – Powerful Analytics for Everyone

Update – Watson Analytics Beta is available now. Signup and try now @ http://www.ibm.com/analytics/watson-analytics/

On16 Sep 2014 IBM announced Watson Analytics, a breakthrough natural language-based cognitive service that can provide instant access to powerful predictive and visual analytic tools for businesses. Watson Analytics is designed to make advanced and predictive analytics easy to acquire and use.

Watson Analytics Sample 1

Watson Analytics Sample 1

Watson Analytics offers a full range of self-service analytics, including access to easy-to-use data refinement and data warehousing services that make it easier for business users to acquire and prepare data – beyond simple spreadsheets – for analysis and visualization that can be acted upon and interacted with. Watson Analytics also incorporates natural language processing so business users can ask the right questions and get results in terms familiar to their business that can be read and understood or interacted with e.g.Who are my most profitable customers and why?  or Which customers are showing signs that they might be considering defecting to a competitor? or Which sales opportunities are likely to turn into wins?  or Why are these products being returned?

I think the most amazing aspect is that Watson Analytics not only can show what is likely to happen but also what you can do about it e.g. What actions could I take that would improve my chances of closing specific deals in the pipeline? Where should we be focusing our loyalty programs?

Surely Watson Analytics is a powerful tool for any business. To make it even easier for businesses to use this technology, its being made available over cloud.  It will be hosted on SoftLayer and available through the IBM Cloud marketplace. IBM also intends to make Watson Analytics services available through IBM Bluemix to enable developers and ISVs to leverage its capabilities in their applications. Certain Watson Analytics capabilities will be available for beta test users within 30 days, and offered in a variety of freemium and premium packages starting later this year.

This is where Watson Analytics distinguishes itself from the crowded analytics marketplace. IBM makes really advanced business analytics accessible to any business user anywhere and that too without a cost! In the freemium business model a product or service is provided free to a large group of users, but a premium is charged for access to more data sources, data storage and enterprise capabilities.

Beta for the service is expected to launch September, with availability later this year. Don’t wait! Check out IBM Watson Analytics!

 

Elements of Watson – Natural Language Processing, an introduction

I would say the core component of IBM Watson is the Natural Language Processing, without which, surely Watson will not even be able to understand the question, let alone answer. NLP allows Watson to really derive the meaning of the question. How many parts are there in the question and how to interpret meaning and relation between each part.  This semantics of the question, obviously, will have direct implication on the correctness of answer for the question.

Natural Language Processing or NLP as its called, is a huge area in artificial intelligence field concerned with Human Computer interactions (HCI). I do not think of myself as knowledgeable on this subject matter but lately I did try understand its depth in respect to my interests in Watson. This blog post simply documents that study to some extent.

ELIZA written at MIT by Joseph Weizenbaum around 1964-66, named after the protagonist in George Bernard Shaw’s Pygmalion, is one of the earliest examples of NLP. ELIZA is in some sorts great grandmother of Apple’s Siri J

The first thing to know to begin with NLP will be the linguistic concepts. Grammar is at the core of any language. Grammar of a language is how sentences are constructed. It is interesting to note history of grammar that, the first systematic grammars originated in Iron Age India, with Yaska (6th century BC), Pāṇini (4th century BC) and his commentators Pingala (c. 200 BC), Katyayana, and Patanjali (2nd century BC). Tolkāppiyam is the earliest Tamil grammar is mostly dated to before the 5th century AD.

In NLP, multiple text processing components are used in sort of a pipeline of tasks performed, in order to provide value from text. Text processing components are tokenization, sentence segmentation, part-of-speech tagging, named entity extraction, chunking, parsing, and co-reference resolution.

Tokenization is breaking up a stream of text into smaller meaningful parts called tokens. Tokens are usually words, phrases or sentences demarcated by punctuation. In English, words are mostly separated from each other by blanks or white space as they are called in IT world. Thing to remember thought is that not all white space is equal e.g. “San Francisco” or “fast food” is supposed to taken as a single word. On other hand “I’m” should be two words “I am”. Further challenges to consider are abbreviations, acronyms, hyphenated words and most importantly numerical and special expressions such as telephone numbers, date format, vehicle license number.

Sentence segmentation is simply put dividing the stream of text into sentences. Let us not jump into conclusion that its basically looking for a period or other end of sentence punctuation. Remember, period is used in many more context and not only end of sentence. A period may denote an abbreviation, decimal point, or an email address. Question marks and exclamation marks may appear in embedded quotations, emoticons, computer code, and slang. Thus, we need specialized algorithms to really identify the end of sentence. In fact, there is special name for such algorithms. Its called Sentence Boundary Detection. I read this very interesting paper by Dan Gillick of UC-Berkley on this. https://code.google.com/p/splitta/ He claims this includes proper tokenization and models for very high accuracy sentence boundary detection (English only for now). The models are trained from Wall Street Journal news combined with the Brown Corpus which is intended to be widely representative of written English. Error rates on test news data are near 0.25%.

Another very interesting component of text processing is part of speech tagging. The basics of a language are the words. Linguist classify the words into various classes or ‘parts of speech’ (POS). In English language, they noun, verb, article, adjective, preposition, pronoun, adverb, conjunction, and interjection. The process of assigning one of the parts of speech to the given word, rather token, is called Parts Of Speech tagging. POS Tagger is an algorithm or program which does POS tagging to tokens. POS Tagger marks tokens with their corresponding word type based on the token itself and the context of the token. The very important part of context of the token. Wiki has a very interesting example in sentence “The sailor dogs the hatch.”. “dogs” is usually thought of as just a plural noun, in this context is a verb. There are variety of algorithms which are used to do POS tagging. Some of them are Viterbi algorithm, Brill Tagger, Constraint Grammar, Baum-Welch algorithm. Methods like Hidden Markov Models or visible Markov model have been used. Many machine learning methods have also been applied to the problem of POS tagging e.g. SVM, Maximum entropy classifier, maximum-entropy Markov model (MEMM), or conditional Markov model (CMM), Perceptron, and Nearest-neighbor.

Further classification by labeling sequences of words in text processing as names of persons, organizations, locations, expressions of times, quantities, monetary values, percentages, etc. is named entity extraction. There are many approaches, linguistic grammar-based techniques as well as statistical models, for Named entity extraction or Named entity recognition (NER), as it is some times called. Conditional random fields (CRFs) are a class of statistical modeling method often applied for NER.

Parsing is the task of analyzing the grammatical structure of natural language. Parser finds which groups of words go together (as “phrases”) and which words are the subject or object of a verb. Parser generates a parse tree of a given sentence. Parse trees are something like the “sentence diagram” we learnt when we were learning English grammar in school. Parse trees can be Constituency-based or Dependency-based. A constituent is a word or a group of words that functions as a single unit within a hierarchical structure in a phrase structure grammar; on other hand, the dependency relation views the (finite) verb as the structural center of all clause structure. All other syntactic units (e.g. words) are either directly or indirectly dependent on the verb.

Chunking is also called shallow parsing and it’s basically the identification of parts of speech and short phrases e.g. noun phrase or verb phrase. Full parsing is expensive, Partial or shallow parsing can be much faster, may be sufficient for many applications and can also serve as a possible first step for full parsing.

Last but not the least, we discuss co-reference resolution. Co-reference resolution is the task of finding all expressions that refer to the same entity in a discourse. E.g. ‘Sheila said she fell down’. Here She is referring to Sheila is the co-reference resolution. There are many types of co-references, like anaphora, cataphora, split antecedents, co-referring noun phrases, etc. Anaphora is when proform follows the expression to which it refers e.g. ‘Sheila said she fell down’ She follows Sheila, to whom it refers to. Cataphora is when proform precedes the expression to which it refers e.g. e.g. ‘she fell down, Sheila said’, She precedes Sheila, to whom it refers to. Algorithms intended to resolve co-references commonly look first for the nearest preceding individual that is compatible with the referring expression. Some are deterministic or multi-pass sieve algorithms.

So these were some of the very basic components of Natural language processing which forms the core of IBM Watson.

Monetizing Telco Big Data: Location, Location, Location

This post is based on article with same name by Dr Sambit Sahu, a manager and senior research scientist at the IBM T.J. Watson Research Center. His current research focuses on Cloud and Big Data analytics for Telco and Smarter Cities.

As I read the above mentioned article, I felt that future has come to present in case of broadcasting!

Broadcasting is more than a century old,  starting towards end of 19th century, with data services offered by stock telegraph companies and “the theatre phone”, a telephonic distribution system that allowed the subscribers to listen to opera and theatre performances over the telephone lines.   The business model for broadcasting has been limited to very few ways or a combination of these, like funding, subscription, paid programming or the most popular one, advertising. There has been obviously tremendous growth in Radio and Television broadcasting industry in the past few decades. From the typical broadcast, with advent of internet and mobile, Radio and TV has option to become interactive too; yet, the traditional plain broadcast is still the most popular.

In the advertising world, contextual advertising is gaining ground. Though I am not aware of any studies which suggest effectiveness of this mode of advertising but logically it does seem that this will be likely the most effective. Contextual advertising is a form of targeted advertising where, the advertisements themselves are selected and served by automated systems based on the context of the user. For example, per Wikipedia, if the user is viewing a website pertaining to sports and that website uses contextual advertising, the user may see advertisements for sports-related companies, such as memorabilia dealers or ticket sellers. Contextual advertising is also used by search engines to display advertisements on their search results pages based on the keywords in the user’s query.

That was a lot of ‘context’ 🙂 to the article; coming to Big Data and monetizing Telco Big Data; Telcos have been the one industry which has bloomed exponentially in the last 2 decades; growing their susbscriber base through landlines, mobile telephony, internet and broadcasting services. This kind of provisioning in Telco world is called Triple play. Per Wikipedia, In telecommunications, triple play service is a marketing term for the provisioning, over a single broadband connection, of two bandwidth-intensive services, high-speed Internet access and television, and the latency-sensitive telephone.  This implies telcos have huge volume of data which can give ‘context’ to acitvities of their subscriber. This could lead to some indications of ‘behavior’ of the subscribers. This information along with location of the subscriber can potentially open lock to a huge treasure of knowledge.

Before we move forward on the monetizing, we need to also keep in mind the security and privacy issues related to usage of this data. The good news is that there exists tested and proven IBM technologies, which can surely give solutions in these areas. That, I am not covering in this blog post; but surely something to not just ponder but even decisive criteria.

In order to monetize the huge treasure of data in context of location; it needs to be converted into information which can be judiciously used to not only derive contexual advertising but also ensure that the efficiency of the advertisement peaks. Dr Sambit Sahu mentions two use-cases which is team has prepared and are going to demonstrate them in the ongoing Mobile World Congress. Below are the use-cases he mentions in his article

 In the first, the analytics platform is being used to create targeted Internet Protocol Television (IPTV) advertisements based on a customer’s profile. So, instead of every TV watcher seeing the same ad at the same time, during the same program, opt-in participants would see tailored advertisements that best match their profiles. Even individual family members would see different advertising based on knowing who is at home via cell phone location data, in conjunction with what programming they’re watching and their specific profile attributes.

The second use case, seeks to explore hyper-local targeted advertisements that will be delivered to mobile phones. In this use case, targeted advertisements and coupons will be delivered to a customer based on a better understanding of their profile, as well as current and predicted locations.

The play of the analytics technology is extremely core to these use-cases. Its the big data analytics technology which can from understanding of customer’s profile, infer intent, current and predicted locations or even predict future behaviors! This move from infering intent to predicting future behavior is which can lead to much better levels of effectiveness of the advertisement.

The model of Clow and Baack speaks of the six steps of objectives of an advertising campaign; Awareness, Knowledge, Liking, Preference, Conviction, Purchase. With real contextual advertising, telcos can move forward from branding methods of Awareness, Knowledge, Liking & Preference, closer to actual selling and sales revenues by Conviction, Purchase!

Creating Cognitive Apps Powered By IBM Watson

There has been a lot of buzz around IBM Watson ever since, it defeated Brad Rutter and Ken Jennings in the Jeopardy! Challenge of February 2011. Recently again there has been much activity surrounding IBM Watson, when IBM did a press release stating that its forming a new Watson group to meet growing demand of cognitive innovations. In addition to launching the IBM Watson Group, IBM also announced three new services based on Watson’s cognitive intelligence, IBM Watson Discovery Advisor, IBM Watson Analytics & IBM Watson Explorer.

In this blog post, I want to share how one can create cognitive applications using IBM Watson. To start with, we need to first understand what is cognitive computing. Originally referred to as Artificial Intelligence, the term ‘Cognitive computing’ is not just a newer term but also an re-engineering to make the computer systems model like a human brain. The key difference is that artifical intelligent system lack ability to learn from its experience. I like what I read somewhere tha, Cognitive computing aims to reverse-Engineer the mind! Cognitive computing systems learn and interact naturally with people to extend what either humans or machine could do on their own. These are machines of tomorrow and will completely change the way people interact with computing systems. IBM Watson represents a first step into cognitive systems. Following three form the core to IBM Watson –
1. Natural language processing by helping to understand the complexities of unstructured data
2. Hypothesis generation and evaluation by applying advanced analytics to weigh and evaluate a panel of responses based on only relevant evidence
3. Dynamic learning by helping to improve learning based on outcomes to get smarter with each iteration and interaction

This redguide will help you understand insides of IBM Watson and how it works.

Companies can embed Watson’s cognitive capabilities in their application without the need for building deep natural language, machine learning and ranking algorithms or other core technology skills. They could accomplish this by being able to embed the Watson platform capabilities as a service, through the use of API and tools on the Watson Developer Cloud.

Application Programming Interface (API) : The Watson Question and Answer API (QAAPI) is a Representational State Transfer (REST) service interface that allows applications to interact with Watson. Using this API, one can pose questions to Watson, retrieve responses, and submit feedback on those responses. In addition to simple question and responses, Watson can provide transparency into how it reached its conclusions through the REST services. There are other functions like ingesting content in the Watson platform that will also be exposed as tools and APIs and can be accessed from within an application.

Watson supports two ways of using the QAAPI – Asynchronous and synchronous mode. In the asynchronous mode, the question is posted to Watson and the response returns a link to retrieve the answer when it is ready. In the synchronous mode, a POST operation sends the question to Watson. The answer is received synchronously. The output from Watson contains not just the response, but it also contains the confidence score, the evidence and other metadata about the response. If there is an error condition with the pipeline, additional information is returned in the “errorNotifications” and is helpful in working with IBM Support.

The key steps involved in using QAAPI are:
1. Configure parameters, including authentication : First the user configures various question parameters on the question, including the Basic Authentication, Content type, Accept type, and for a request in a synchronousmode, the sync-timeout, etc. The QAAPI uses Basic Authentication over SSL to provide security. The QAAPI supports JSON as the content-type and accept type. XML is also supported.

2. Post question : The question needs to be in a JSON format. The output can be customized by passing in the number of responses to return with the question itself, along with additional insights into how Watson handled the question.

3. Receive response : The output from Watson contains not just the response, but it also contains the confidence score, the evidence and other metadata about the response. Some of this information provides insight into how the question was interpreted by Watson. If there is an error condition with the pipeline, additional information is returned in the “errorNotifications” and is helpful in working with IBM Support.

4. Process response : The response can be shown to a user asking the question as-is, or it could passed on to another component for additional processing before it is displayed to a user. It could be used to visually represent the response(s), create a way to ask additional simulated questions or process the answer to obtain additional insights on the interactions. After Watson provides the response, the application developer is responsible to integrate that response seamlessly into their application.

In this manner by creating applications to embed Watson’s cognitive skills, Watson can enable businesses and consumers to have a dynamic exchange, leading to insights and new, synergistic ways of thinking. This can enable companies to
• Gain insights from a hitherto largely untapped sources of unstructured content
• Converse with clients or users in a way that humans would – using natural language
• Enhance the knowledge in the company’s domain with a system that learns over time.

Reference : Whitepaper: Creating Cognitive Apps Powered By IBM Watson – Getting Started with the API

The answer is blowin’ in the wind

This post is on news release about how IBM Cloud lifts Bharat Light and Power to build Smarter Operations.

I came across this news release about Bharat Light and Power and IBM Collaborate to Drive Business Growth and Build Smarter Operations. The Indian renewable energy producer will use IBM cloud, analytics, and mobile technologies to increase power generation. This news item pleases me much more than some other multi-million dollar deals since this truely is what IBM has been striving to achieve in its campaign for Smarter Planet.
Since the development of Wind Power in early 90s, India has installed capacity of 19,000+ MW and accounts for about 8.5% of India’s total power capacity.
Bharat Light owns about 200 megawatts of operating wind farms, including 150 megawatts of capacity acquired from DLF in July. It plans to diversify into solar, biomass and hydropower to reach 1,000 megawatts in five years through acquisitions and by building projects.

So what is IBM, an IT major doing for Bharat Light and Power? Earlier this year, India’s Central Electricity Regulatory Commission, gave a directive ordering wind farms with a capacity of 10 megawatts or more to forecast their generation in 15-minute blocks for the following day. Missing estimates by more than 30 percent will incur penalties. The grid infrastructure had one of largest backouts last year and it cannot cope with volatility of wind and sun power generation. Though the directive is quite needed but it will put additional burden on wind and sun power generation companies in adding  the complex forecasting systems for this purpose. IBM technologies will help Bharat Light in achieving this directive.

Another big contribution IBM will do is to increase the operational efficiencies. “The focus of power generation globally, and especially in India, in the next five years is going to be about operational efficiency and productivity enhancements. This requires an intelligent and innovative collaboration between the power and the IT sector. BLP has ambitious growth plans for providing this service and we needed a strategic partner who understands, and helps us realize, our vision,” said Balki G. Iyer, chief development officer, BLP. “A significant factor in selecting IBM was its rich global experience in the renewable energy sector and scalable technology capabilities that suits our requirement.”

Per the news release, IBM will deploy its SoftLayer cloud infrastructure as a service to centrally monitor and manage BLP’s existing and future generation plants as well as store and manage the data on cloud. Also the IBM technologies will help Bharat Light manage the huge amount of data and leverage the crucial insights using advanced analytics including predictive analytics. IBM and BLP’s big data analytics capabilities will help gather valuable insights from the data generated, which will ensure that BLP has an integrated view of its operations and is equipped to take pro-active measures. Using IBM’s mobile technology the company will be able to provide all the information, analyzed on the cloud platform, to its ground staff on their handsets and alert them well in advance. This will enable BLP to build smarter operations with higher efficiency and greater utilization.

The news release further informs that the solution is expected to significantly improve BLP’s Plant Load Factor (PLF), which is essentially the generation efficiency of the installed capacity. IBM will also bring its subject matter experts to help BLP take effective business decisions and make efficient planning, based on the data analysis.

“By making clean energy widely available and affordable, BLP is solving a very important challenge for India and we are excited about collaborating with BLP on this transformative journey. We see tremendous potential for smarter operations, using cloud and analytics, in achieving scale and efficiencies in this business. As a strategic partner, we will deliver measurable business outcomes to BLP by bringing together expertise and capabilities across IBM,” said Ajoy Menon, director of strategic outsourcing, Global Technology Services, IBM India South Asia.

I congratulate IBM and Bharat Light and Power on this agreement and hope that this boosts multifold, the electricity output of wind farms in India.

Migrating from IBM SmartCloud Enterprise to the IBM SoftLayer cloud environment

Further to my earlier post, I was questioned on the migration from IBM SmartCloud Enterprise to the IBM SoftLayer cloud environment. I thought surely it is important to have this information readily available. Yes, it indeed is!
I am blogging my understanding on this migration activity just to verify from my side that customers would be comfortable with the process. I find the process to be quite straight forward and uncomplicated, nothing which should bring concern to customers.
There are strong reasons for recommending this migration to SCE customers. Customers will have One portal, one API, one platform with the stability, support and trust you have always received through your relationship with IBM. As a SoftLayer customer, gains are :
Unmatched flexibility, customization and control for Internet-scale applications with 1600+ functional calls and 200+ services
Unparalleled speed with integrated high-performance bare metal and virtualized servers in public, private or hybrid cloud configurations
Flexibility to take advantage of a wide range of storage, network and security options
Ability to create Internet-scale distributed compute environments in minutes and control through a single management system via API, browser or mobile app

Per the migration web link, ‘On 8 July 2013, IBM® announced the acquisition of SoftLayer, the industry’s only seamlessly unified global cloud computing infrastructure. Security, scalability, and on-demand access to virtual server and storage resources are important to IBM SmartCloud® Enterprise (SCE) customers. SoftLayer delivers global, high-performance cloud computing infrastructure able to support applications at Internet-scale. With SoftLayer’s global cloud environment, you can have it all: One portal, one API, and one platform with the stability, support, and trust you have always received through your partnership with IBM. IBM is encouraging all SmartCloud Enterprise customers and partners to migrate eligible workloads to the SoftLayer cloud environment. To assist in this migration, IBM is providing the use of a fast and easy migration tool to streamline the process. In addition, the IBM team has created a set of information and education resources to help you with your transition to SoftLayer. If you need additional support, contact an IBM representative at Migr8Me@us.ibm.com.’

There a testing or sandbox area where customers can validate their workload under SoftLayer. SoftLayer offers a one month free trial on one CloudLayer Computing public cloud instance. Remember that an instance migration from SCE to SoftLayer may cause service disruption. Although the original SCE instance should remain active during the entire migration process, we recommend that you take appropriate precautions and backup methods.

One option for the migration is to use Racemi CloudPath. Racemi Cloud Path for IBM is a no charge [ See T&C ], Software as a Service (SaaS) tool that enables customers to quickly and easily migrate their existing physical, virtual, or SmartCloud Enterprise cloud servers to IBM’s SoftLayer CloudLayer platform in an automated, unattended way. With Racemi Cloud Path for IBM, you can now automatically move your existing IBM SmartCloud Enterprise (SCE) servers to the new SoftLayer platform and experience minimal downtime of your existing SCE workload. Cloud Path for IBM provides an automated cloud instances migration process that moves the entire server stack to the new SoftLayer environment in a few easy-to-follow steps, without the need to make additional setting changes on the SCE servers.
The article on getting started on the migration, documents crystal clear and easy 10 steps –

1) Analyze your current SCE usage, including items such as operating systems, VM sizes, persistent storage requirements, reserved IP addresses, applications and services used.
2) Read & Understand Migrating VM instances from SCE to SoftLayer using Racemi Cloud Path service.
3) Read Mapping SCE VM sizes to equivalent configurations in SoftLayer and Mapping IBM SmartCloud Enterprise service names, terms, and concepts to SoftLayer to help you understand differences between SCE and SoftLayer.
4) Identify which of your workload VMs you wish to migrate.
5) Identify your workloads that can be migrated using the Racemi tool and those that must use another method.
6) Create accounts in SoftLayer and Racemi.
7) Migrate your workloads.
8 ) Test and validate your migrated workloads on SoftLayer. Verify that your migration was successful using your normal testing methods.
9) Close your SCE account. This will avoid any unnecessary billing for SCE resources.
10) Send an email to migr8me@us.ibm.com if you have questions regarding migration to SoftLayer.

Details of these steps are available on this IBM DeveloperWorks article http://www.ibm.com/developerworks/cloud/library/cl-sce-migration-get-started/index.html

Migration FAQs & Resources
http://www.ibm.com/developerworks/cloud/library/cl-sce-migration-FAQ/
http://www.ibm.com/developerworks/cloud/library/cl-sce-migration-get-started/index.html
http://www.racemi.com/ibm/

http://www.ibm.com/developerworks/cloud/library/?series_title_by=migrate+to+softlayer

NOTE: Please do note that this post is only for informational purpose and NOT an advice or consultation for the migration activity. Please contact IBM representative at Migr8Me@us.ibm.com for any further queries, details or request for engagement for migration.