Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

Bigtable is a distributed storage system for managing structured data that is designed to scale to a very large size: petabytes of data across thousands of commodity servers. Many projects at Google store data in Bigtable, including web indexing, Google Earth, and Google Finance. These applications place very different demands on Bigtable, both in terms of data size (from URLs to web pages to satellite imagery) and latency requirements (from backend bulk processing to real-time data serving). Despite these varied demands, Bigtable has successfully provided a flexible, high-performance solution for all of these Google products. In this article, we describe the simple data model provided by Bigtable, which gives clients dynamic control over data layout and format, and we describe the design and implementation of Bigtable.

Bigtable: A Distributed Storage System for Structured Data (Awarded Best Paper!).

Bigtable: A Distributed Storage System for Structured Data

/pdf/the-theory-of-industrial-organization-19bejph0ej.pdf

The Theory of Industrial Organization

The objective of this paper is to give a comprehensive introduction to applied cryptography with an engineer or computer scientist in mind. The emphasis is on the knowledge needed to create practical systems which supports integrity, confidentiality, or authenticity. Topics covered includes an introduction to the concepts in cryptography, attacks against cryptographic systems, key use and handling, random bit generation, encryption modes, and message authentication codes. Recommendations on algorithms and further reading is given in the end of the paper. This paper should make the reader able to build, understand and evaluate system descriptions and designs based on the cryptographic components described in the paper.

[서평]「Applied Cryptography」

This paper describes Seeker, a platform for large-scale text analytics, and SemTag, an application written on the platform to perform automated semantic tagging of large corpora. We apply SemTag to a collection of approximately 264 million web pages, and generate approximately 434 million automatically disambiguated semantic tags, published to the web as a label bureau providing metadata regarding the 434 million annotations. To our knowledge, this is the largest scale semantic tagging effort to date.We describe the Seeker platform, discuss the architecture of the SemTag application, describe a new disambiguation algorithm specialized to support ontological disambiguation of large-scale data, evaluate the algorithm, and present our final results with information about acquiring and making use of the semantic tags. We argue that automated large scale semantic tagging of ambiguous content can bootstrap and accelerate the creation of the semantic web.

/pdf/semtag-and-seeker-bootstrapping-the-semantic-web-via-4azq3axspb.pdf

SemTag and seeker: bootstrapping the semantic web via automated semantic annotation

There is a fundamental transformation that is taking place on the web around information composition through mashups. We first describe this transformation and then assert that this will also affect enterprise architectures. Currently the state-of-the-art in enterprises around information composition is federation and other integration technologies. These scale well, and are well worth the upfront investment for enterprise class, long-lived applications. However, there are many information composition tasks that are not currently well served by these architectures. The needs of Situational Applications (i.e. applications that come together for solving some immediate business problems) are one such set of tasks. Augmenting structured data with unstructured information is another such task. Our hypothesis is that a new class of integration technologies will emerge to serve these tasks, and we call it an enterprise information mashup fabric. In the talk, we discuss the information management primitives that are needed for this fabric, the various options that exist for implementation, and pose several, currently unanswered, research questions.

https://www.vldb.org/conf/2006/p3-jhingran.pdf

Enterprise information mashups: integrating information, simply

Methods and apparatus for providing a multi-tier object-relational database architecture are disclosed. In one illustrative embodiment of the present invention, a multi-tier database architecture comprises an object-relational database engine as a top tier, one or more domain-specific extension modules as a bottom tier, and one or more universal extension modules as a middle tier. The individual extension modules of the bottom tier operationally connect with the one or more universal extension modules which, themselves, operationally connect with the database engine. The domain-specific extension modules preferably provide such functions as search, index, and retrieval services of images, video, audio, time series, web pages, text, XML, spatial data, etc. The domain-specific extension modules may include one or more IBM DB2 extenders, Oracle data cartridges and/or Informix datablades, although other domain-specific extension modules may be used.

/pdf/methods-and-apparatus-for-constructing-and-implementing-a-230wkk3yi9.pdf

Methods and apparatus for constructing and implementing a universal extension module for processing objects in a database

An apparatus and method for approximating the data stored in a databases by generating multiple projections and representations from the database such that the OLAP queries for the original database (such as aggregation and histogram operations) may be applied to the approximated version of the database, which can be much smaller than the original databases. Other aspects optimize a mapping, via a mapping (or dimension) table, of non-numeric or numeric attributes to other numeric attributes such that the error incurred on applying queries to the approximated version of the database is minimized. Still further aspects define boundaries of approximations so that the boundaries are preserved when approximated versions of the databases are generated.

System and method for generating multi-representations of a data cube

In a partitioned database system of the Shared Nothing type, one or more secondary replicas of each partition are maintained by spooling (i.e., asynchronously sending) modified (usually called dirty) pages from the primary replica to the secondary replica(s) rather than by using a synchronous page update or by sending log entries instead of entire pages. A Write-Ahead Log protocol is used so that a dirty page is not forced to non-volatile storage until a log record of the modification is created and written to non-volatile storage. Replica updating does not delay the committing of transactions because replica updating is done asynchronously with respect to transaction processing. Since dirty pages are sent rather than only log entries, disk accesses and processing at the secondary replica(s) arising from the maintaining of the replicas are minimized as well. Only one centrally accessible log is maintained for all replicas of the same partition.

Anant Jhingran

Papers

SemTag and seeker: bootstrapping the semantic web via automated semantic annotation

Enterprise information mashups: integrating information, simply

Methods and apparatus for constructing and implementing a universal extension module for processing objects in a database

System and method for generating multi-representations of a data cube

Asynchronous replica management in shared nothing architectures