অন্টলজিঃ খায় না মাথায় দেয়

ওয়ার্ড ওয়াইড ওয়েভ (WWW), এই সময়ে সবচেয়ে বড় এবং সবচেয়ে ব্যবহৃত তথ্য ভান্ডার। সমস্যা হচ্ছে, ওয়েবের অধিকাংশ ডকুমেন্ট শুধু মানুষ (পড়ে) বুঝতে পারে কিন্তু মেশিন বুঝতে পারে না। ভবিষ্যতের ওয়েব হবে এমন যা মানুষের পাশাপাশি মেশিনও বুঝতে পারবে। এতে করে, আমরা যেসব কাজ নিজে করি, যেমন-ওয়েব থেকে কোন তথ্য খুজে বের করা, প্রয়োজনীয় তথ্য একসাথ করা, বিভিন্ন তথ্যের মধ্যে তুলনা করা, সিদ্ধান্ত গ্রহণ করা, ইত্যাদি-মেশিন আমাদের হয়ে নিজে করে দিবে। এই উদ্দেশ্যে, ওয়েবে তথ্যকে (web content) অন্টলজি দিয়ে উপস্থাপন করার জন্য ওয়ার্ড ওয়াইড ওয়েভ কন্সর্টিয়াম (WWW consortium) সুপারিশ করেছে।

অন্টলজি প্রথম ফিলোসফিতে আলোচিত হলেও, বর্তমানে আর্টিফিশিয়াল ইন্টিলিজেন্স এবং নলেজ ইঞ্জিনিয়ারিং তা অভিযোজন করা হয়েছে এবং ব্যাপকহারে ব্যবহৃত হচ্ছে। এই ক্ষেত্রে, অন্টলজি হচ্ছে কোন ডোমেইন নলেজ উপস্থাপন করার একটি ফরমাল মডল যা ঐ ডোমেইনের অনুসঙ্গিক শব্দতালিকা, শব্দার্থের বিবরণী, ও ডোমেইনের একটি বিমূর্ত ধারণা প্রকাশ করে। বাস্তব জীবনে, ই-সায়েন্স, ই-কমার্স, মেডিকেল ইনফরমেটিক্স, বায়ো-ইনফরমেটিক্স এবং সেমান্টিক ওয়েবে অন্টলজি ব্যবহৃত হয়। বিশেষ করে, ওয়েব সার্স, ইন্টিলিজেন্ট সফটওয়ার এজেন্ট, নলেজ ম্যানেজমেন্টে অন্টলজি ব্যবহৃত হয়।

যত বেশি ওন্টলোজি তৈরি হবে, সেসব ওন্টলোজি যত বেশি বড় হবে, এবং তাদের যত বেশি বোধগম্য হিসেবে উপস্থাপন করা হবে, তত বেশি স্বয়ংক্রিয় ওন্টলোজি তৈরি ও ব্যবস্থাপনা পদ্ধতি প্রয়োজন হবে।  আমাদের ল্যাবে, ডাটা মাইনিং ব্যবহার করে ডকুমেন্ট থেকে স্বয়ংক্রিয় ওন্টলোজি তৈরি ও ব্যবস্থাপনা পদ্ধতি নিয়ে আমরা কাজ করছি।

অডিও রেকর্ডিং: ব্রেইল প্রিন্টারের বিকল্প

অন্ধদের জন্য ব্রেইল প্রিন্টার একটা চমৎকার জিনিস। তবে ব্রেইল প্রিন্ট সাধারণ প্রিন্টের মত এত সহজ লভ্য নয়। যারফলে সাহিত্য জগতের খুব অল্প পরিসরেই দৃষ্টিপ্রতিবন্ধিদের বিচরণ সিমিত থাকে। এই সমস্যা অনেকাংশে দূর করা সম্ভব অডিও রেকর্ডিং এর মাধ্যমে। প্রকাশিত বই গুলো কেউ একজন (বা একাধিকজন) পড়ে রেকর্ডিং করে দিবে। সেটা বাজারে বিক্রি হবে দৃষ্টি প্রতিবন্ধিদের জন্য বইয়ের বিকল্প হিসেবে। বাহিরে এসে দেখলাম এই পদ্ধতিটা দৃষ্টি প্রতিবন্ধিদের কাছে বেশ জনপ্রিয়। আমাদের দেশের প্রকাশকরা যদি এই ব্যপারে একটু দৃষ্টি দেন তবে মনে হয় দৃষ্টি প্রতিবন্ধিরা নিজেদের সাহিত্যের আলোয় আরো বেশি আলোকিত করতে পারবে।

Project Management

Risk Management

Assumptions:  Assumptions are factors that we believe to be true, although these factors are not confirmed to be true. Assumptions add risk to a project since it is possible that they will turn out to be false. Assumptions can impact any part of your project life cycle and resulting solution implementation, so it is important to document and analyze them.


People make assumptions about things that they believe to be true every day. When was the last time that you left the house headed out to do your errands and calculated the time it would take to get those errands done based upon your assumptions for what you needed to do, where you were stopping and what traffic would be like along the way? Of course, there was always the risk that you would get those errands done in more time than you planned if the order of stops, what you did at each stop and the traffic did not behave as you assumed it would behave [LTIB].

মুকুল ভাইয়ের জন্ম দিন

[শানে নজুলঃ মুকুল ভাইয়ের জন্ম দিন উপলক্ষে আমরা তাকে সারপ্রাইজ দিতে অনেক রাতে তার বাসায় হানা দিয়েছিলাম। উনার যে জন্ম দিন ছিলা সেটা উনি ভুলে গিয়েছিলেন। অনেক রাতে উনার ঘরের দরজা নাড়ানোতে ভয় পেয়েছিলেন। ঘটনা ক্রমে আগের দিন উনি আমাদেরকে নিয়ে মাছ ধরেছিলেন। সে মাছ রাজ আমাদেরকে না বলে রান্না করে খেয়ে ফেলছিল।]

রাতের বেলা, বাহিরে যেন কারা ডাকে

শুনেছিলাম, এখানে নাকি ভুতও থাকে।

একা নয়, অনেক আছে
ঘাড় না মটকায়, পাইলে কাছে।

রাজের বুদ্ধি, না জানিয়ে মাছ খেয়েছে
সেটাই কি তবে কাল হয়েছে?

Topic Modelling

Introduction

Following its publication in 2003, Blei et al.’s Latent Dirichlet Allocation (LDA) [3] has made topic modeling – a subfield of machine learning applied to everything from computational linguistics [4] to bioinformatics [8] and political science [2] – one of the most popular and most successful paradigms for both supervised and unsupervised learning.

Required Basic Definitions

Joint (combined) probability distribution: In the study of probability, given at least two random variables X, Y, ..., that are defined on a probability space, the joint probability distribution for X, Y, ... is a probability distribution that gives the probability that each of X, Y, ... falls in any particular range or discrete set of values specified for that variable. In the case of only two random variables, this is called a bivariate distribution, but the concept generalizes to any number of random variables, giving a multivariate distribution [Wikipedia].

Random Set

Related Definitions

Random variable: It is not a variable in the traditional sense of the world.  It is actually a function. The outcome of an experiment need not be a number, for example, the outcome when a coin is tossed can be 'heads' or 'tails'. However, we often want to represent outcomes as numbers [STEPS]. A random variable is a function that associates a unique numerical value with every outcome of an experiment. Random variable is written as capital letter, usually X.

Probability distribution and probability density are similar, where distribution is used for discrete  variable and density is used for continuous variable.

Expected Value: The expected value (or population mean) of a random variable indicates its average or central value. It is a useful summary value (a number) of the variable's distribution [STEPS].

The expected value of a random variable X is symbolised by E(X) or µ.

If X is a discrete random variable with possible values x1, x2, x3, ..., xn, and p(xi) denotes P(X = xi), then the expected value of X is defined by:

where the elements are summed over all values of the random variable X.

Prestigious Rankers of Journals and Conferences

Institute for Scientific Information (ISI)
http://en.wikipedia.org/wiki/Institute_for_Scientific_Information

Data Mining

Introduction: Knowledge Discovery and Data Mining (KDD) is an interdisciplinary area focusing upon methodologies for extracting useful knowledge from data. The ongoing rapid growth of online data due to the Internet and the widespread use of databases have created an immense need for KDD methodologies. The challenge of extracting knowledge from data draws upon research in statistics, databases, pattern recognition, machine learning, data visualization, optimization, and high-performance computing, to deliver advanced business intelligence and web discovery solutions [IBM Research].

"The basic task of KDD is to extract knowledge (or information) from lower level data (databases)." [1] Data, in its raw form, is simply a collection of elements, from which little knowledge can be gleaned. With the development of data discovery techniques the value of the data is significantly improved [UF].

A variety of methods are available to assist in extracting patterns that when interpreted provide valuable, possibly previously unknown, insight into the stored data. This information can be predictive or descriptive in nature. Data mining, the pattern extraction phase of KDD, can take on many forms, the choice dependent on the desired results. KDD is a multi-step process that facilitates the conversion of data to useful information.

Endnote সফটওয়ার

http://www.somewhereinblog.net/blog/shiyak/29664683

DISTRIBUTED DATABASE SYSTEM

What is DDMS
http://en.wikipedia.org/wiki/Distributed_database
A distributed database is a database that is under the control of a central database management system (DBMS) in which storage devices are not all attached to a common CPU. It may be stored in multiple computers located in the same physical location, or may be dispersed over a network of interconnected computers.

Collections of data (eg. in a database) can be distributed across multiple physical locations. A distributed database can reside on network servers on the Internet, on corporate intranets or extranets, or on other company networks. Replication and distribution of databases improve database performance at end-user worksites.

In short a distributed database is a collection of databases that can be stored at different computer network sites. Each database may involve different database management systems and different architectures that distribute the execution of transactions . The objective of a distributed database management system (DDBMS) is to control the management of a distributed database (DDB) in such a way that it appears to the user as a centralized database.

Providing the appearance of a centralized database system is one of the many objectives of a distributed database system. Such an image is accomplished by using the following transparencies: Location Transparency, Performance Transparency, Copy Transparency, Transaction Transparency, Transaction Transparency, Fragment Transparency, Schema Change Transparency, and Local DBMS Transparency. These eight transparencies are believed to incorporate the desired functions of a distributed database system.

Limitations
Lack of DDBMS Standards
Of C.J. Date's twelve commandments for a distributed database, four refer to open standards issues required for DDBMSs to reach their full potential. These are hardware independence, operating system independence, network independence, and database independence. Today's DDBMS products are still do not meet these four standards. Burleson described the causes of lack of open standards in the DDBMS market and the impact this has had on DDBMS growth (Burleson, 1994, pp. 72-73). Five years later, this is still a fact. DDBMS technology is relatively new, and is still suffering from vendors fighting to develop and hold on to proprietary features. Today the situation is improving, but cross-vendor connectivity is sometimes limited, especially for legacy systems that do not implement newer standards.

How to achieve vendor independence
Ref: FUTURE TRENDS IN DATA BASE SYSTEMS (paper)


Research Issues in Distributed DBMSs:
There has been a mountain of research on algorithms to support distributed data bases in the areas of query processing [SELI80], concurrency control [BERN81], crash recovery [SKEE82] and update of multiple copies [DAVI85]. In this section, I indicate two important problems which require further investigation. First, users are contemplating very large distributed data base systems consisting of hundreds or even thousands of nodes. In a large network, it becomes unreasonable to assume that each relation has a unique name. Moreover, having the query optimizer inspect all possible processing sites as candidate locations to perform a distributed join will result in unreasonably long optimizer running times. In short, the problems of ‘‘scale’’ in distributed data bases merit investigation by the research community.

Second, current techniques for updating multiple copies of objects require additional investigation. Consider the simple case of a second copy of a person’s checking account at a remote location. When that person cashes a check, both copies must be updated to ensure consistency in case of failure. Hence, at least two round trip messages must be paid to the remote location to perform this reliably. If the remote account is in Hong Kong, one can expect to wait an unreasonable amount of time for this message traffic to occur. Hence, there will be no sub-second response times to updates of a replicated object. To a user of DBMS services, this delay is unreasonable, and algorithms that address this issue efficiently must be developed. Either a lesser guarantee than consistency must be considered, or alternatively algorithms that work only on special case updates (e.g, ones guaranteed to be commutative) must be investigated. The work reported in [KUMA88] is a step in this direction.


An important aspect related to databases is integrity. Preserving the data integrity is a much more complicated issue in the heterogeneous distributed databases than in homogeneous databases. If the nodes in the distributed database are heterogeneous, there may come up troubles that could threaten the integrity of the distributed data, among which we can mention:
• inconsistencies among local integrity integration constraints.
• difficulties in specifying global integrity constraints.
• inconsistencies among local and global integrity constraints.
If the control coming from the center is powerful, priority is given to global integrity constraints. In case of a poorer central control, local integrity constraints are given priority.


To date, little work has been reported on security for data warehouses as well as the inference problem for the warehouse. This is an area that needs much research.

My Plans
I plan to examine more security issues, concurrent transaction and update-anywhere asynchronous and synchronous system for distributed databases.

It may be good to separate development environment from execution environment of a distributed heterogeneous database system considering integrity constraints. A query is verified in development environment for integrity constraints then executed in execution environment.

Conclusion:
Distributed database systems are a reality. Many organizations are now deploying distributed database systems. Therefore, We have no choice but to ensure that these systems operate in a secure environment. We believe that as more and more technologies emerge, the impact of secure distributed database systems on these technologies will be significant.