See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/332876721

Applications of artificial intelligence in academic libraries

Conference Paper · April 2019

CITATION READS
1 10,503

2 authors:

Vijay Kumar S Kn Sheshadri

Presidency University Presidency University, Bangalore
12 PUBLICATIONS   35 CITATIONS    11 PUBLICATIONS   36 CITATIONS   

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Kn Sheshadri on 23 August 2019.

The user has requested enhancement of the downloaded file.

 International Journal of Computer Sciences and Engineering Open Access
Review Paper Vol.-7, Special Issue-16, May 2019 E-ISSN: 2347-2693

Applications of Artificial Intelligence in Academic Libraries

S. Vijayakumar1*, K.N. Sheshadri2

1,2
Centre for Research in Library and Information Science, Presidency University, Bangalore, India

Corresponding Author: vijaykumars@presidencyuniversity.in, Mob: +91-9980352646

DOI: https://doi.org/10.26438/ijcse/v7si16.136140 | Available online at: www.ijcseonline.org

Abstract— The application of artificial intelligence involves the areas such as artificial intelligence, expert system, artificial
neural network, fuzzy logic, image processing, natural language processing, speech recognition, robotics etc. Though these
areas are not separate, at times two or more applications are contributes to enrich the library services. In this article, the authors
have explored the various possible applications of artificial intelligence as mentioned above. In addition, authors explain the
possible areas where few of these applications can be implemented which enhances the quality of services and thereby create
the potential impact of AI on library services.

Keywords— Artificial Intelligence, Academic Libraries, Expert systems, Robotics, Machine Learning.

I. INTRODUCTION useful shortcut to apply this knowledge and produce better

outcomes. Libraries focus on enhancing the access to content
An attempt to replace human power with the machines was with the application of AI. We have been watching the
the creation of the first industrial revolution. The impact of evidence of this transformation toward AI application with
artificial intelligence and advanced computer technology on many libraries initiating and providing Makerspace
the nature of future libraries will be enormous, and the competences. The libraries are positioning themselves to take
quality differences will be different from what our current advantage of the application of cognitive computing in
work expects. Most library-oriented artificial intelligence general and artificial intelligence in particular for their
applications developed until today or currently under potential utility as a tool for refining the quality of library
development are basic business aids of the runtime because services. Liu (2010) in her articles provide a comprehensive
they built today. Potential applications include systems that literature review on the utilization of intelligent agent
help perform the different tasks for the library such as technology in the library environment. The researcher here
people, budget, collection development, scheduling etc. expressed that both AI and librarians reinforcing each other
These applications include systems for enhancing user in the interest of providing the best service to patrons. Hence,
services, such as ready references and information storage & application of AI can never be a threat for librarians, rather
retrieval. supplementary.

II. REVIEW OF LITERATURE III. ARTIFICIAL INTELLIGENCE

Norman Jacknis (2017) expressed in his blog that though Artificial Intelligence focuses on non-algorithmic methods
librarians have acquired many skills to organize the for solving problems and symbols. AI depends on the skill of
information and making it accessible anywhere, libraries can mapping the symbols. New applications have created great
ensure the application of the tools for the new generation of opportunities for informational researchers, such as
knowledge, which surpasses Google search has been multimedia systems, digital libraries, GISs and e-commerce.
developed for academic purposes. It is explained in the As the application becomes increasingly powerful,
Merriam-Webster (2019) that, artificial intelligence is “a diversified, pressing, several known problems in finding
part of computer science that deals with giving ability to the information became even more important in this
machines to look like they have natural human intelligence.” technological era
Artificial intelligence (AI) is perhaps most familiar to the
public in many ways today. According to Kristin Whitehair The basic technique in the IR includes the identification of
(2016), with many AI applications, it gives libraries the key characteristics in the object. For example, natural
opportunity to change the emphasis and attention. The way language processing and automatic indexing used to
we navigate the information is kept altering. AI gives a very distinguish meaningful words. In general, image

© 2019, IJCSE All Rights Reserved 136

International Journal of Computer Sciences and Engineering Vol. 7(16), May 2019, E-ISSN: 2347-2693

identification uses texture, color or shape - based indexing A. Expert System: The Expert System is a computerized
and segmentation techniques. They used to identify knowledge system that serves as a gateway or interface
meaningful descriptions in their streams for applications such for providing access to the database and obtaining
as audio, video, speech recognition and scene segmentation. relevant information. It ranges from simple regulatory
In order to perform semantic analysis on multimedia objects systems with flat data to very large-scale, integrated
or text, several types of techniques are used. development that takes many years to develop. An
expert system is a computer program that offers expert
Machine learning, graph-based clustering and classification, advice, decisions or solutions to a particular situation.
statistical-based multivariate analysis, artificial neural Knowledge base, inference engine and user interface are
networks, and evolution-based programming are popular the various components of the expert systems.
techniques. These technologies are good alternative for
summarizing, analysing and processing a large number of B. NATURAL Language Processing: One of the long-
multimedia messages, which is different and rapidly standing goals of CS is to teach computers to understand
changing. The results of the semantic analysis process the language we are talking about today. The ultimate
expressed in the form of semantic networks, decisions, rules, generation of computer language is Natural Language.
or predicate logic. Activation-based, Propagation-based Artificial intelligence scientists have been able to build a
reasoning methods commonly used to negotiate various natural language interface using a limited vocabulary
structures of large-scale knowledge. and syntax. The computer can understand the key
language concepts within a question and solution
All search engines for text, images and videos increase user through the natural language process. It aims to design
expectations for presenting and manipulating information. and create a computer that analyzes the language that a
Recent advances in language and platform development, for person uses, understands, and generates. Speech
example VRML, Java, OpenGL, and the provision of synthesis, machine translation, linguistic approaches,
affordable high quality graphic workstations have also made information recovery, information extraction and speech
visualization of information perspective in the field of recognition are the various elements of natural language
research. processing.

Although artificial intelligence is a young discipline, it C. Pattern Recognition: The new stimulus and the pre-
makes society beyond imagination. stored stimulus coincide closely by this process. This
process takes place continuously through the lives of all
living beings. Pattern recognition is being studied in
many areas, including psychology, ethology, cognitive
science, and informatics. Pattern recognition based on
prior knowledge or on data from the patterns. Classified
patterns typically consist of groups of dimensions or
observations that define points in a multi-dimensional
space. Components for pattern recognition are data
collection, pre-processing, selection of characters,
selection of models and training, and evaluation.

D. Robotics: The field robotics is frequently describes as an

AI subfield, which deals with motor and perceptive
tasks. Robot is a mechanical device, which carries out
automation tasks using artificial intelligence techniques,
either directly human control or a predetermined
program.
Figure 1. Pictorial diagram of AI Components E. Machine learning: Arthur Samuel, an American pioneer
in computer gaming and artificial intelligence, invented
AI sub-areas, namely expert systems, natural language the term 'machine learning' in 1959 and defined it as “it
processing, pattern recognition and robotics, aim to simulate gives computer the ability to learn without explicit
human intelligence with computers. Some recent computing programming”. Depending on the nature of the "signal"
techniques and areas for artificial intelligence development or "responses" to the learning system, machine-learning
discussed below: applications divided into the three primary categories,
i.e., (a) Supervised learning (b) Unsupervised learning

© 2019, IJCSE All Rights Reserved 137

International Journal of Computer Sciences and Engineering Vol. 7(16), May 2019, E-ISSN: 2347-2693

(c) Reinforcement learning (d) Semi-supervised 4) Indexing of periodicals is another area where expert
learning. systems are developed. Indexing a periodical article
involves identification of concepts, to translate the
F. HAMLET: The system is HAMLET (How about concepts into verbal descriptions, & selecting and
Machine Learning Enhanced Theses) currently a assigning controlled vocabulary terms that are
developer at the Berkman Klein Center for Internet and conceptually equivalent to verbal descriptions. The
Society at Harvard. HAMLET uses the doc2vec reason for automating the intellectual aspects of
algorithm. This is an algorithm for estimating the indexing is to improve indexing consistency and quality.
similarity in meaning between different documents, Based on the information provided by the indexer, the
based on a widely used algorithm word2vec, which systems can arrive at appropriate preferred terms
estimates the similarity between words. It explores the automatically to assign relevant subdivisions. The
results online at the URL in the gray box. HAMLET has system can make inferences & based on the inference, it
three prototype interfaces: a recommendation engine, an can take appropriate action. The 'Med Index' is the best
uploaded file oracle, and a literature review buddy. example of the library indexing system. As there is a
lack of exposure to these expert system oriented services
IV. APPLICATION OF AI COMPONENTS IN in many libraries, very few library users have interacted
LIBRARY SERVICES with knowledge-based systems. In addition, most of
these expert systems oriented services are evolving over
A. Expert System in Library Services the period and undergoing many improvements to suit
Library activities related to reading materials, users and staff. the needs of the library patron.
Application of expert systems where dialogue between staff
and users, users and databases is promising. The expert 5) Acquisition: The users of the library have a significant
system will help the librarian to understand the need for role to play in building library collection and online
improvement in productivity. A well-programmed expert resources in particular. Several systems have been
system will also improve quality. incorporates for the acquisition of these resources.
Monograph Selection Advisor, a pioneering effort in
1) Reference service is the foremost activity of any library applying this emerging technology is another area of
and the expert system will serve as a substitute for building library collection. Specifically, the task
reference librarians. REFSEARCH, POINTER, Online modeled is the item-by-item decision that a subject
Reference Assistance (ORA), AMSWERMAN, bibliographer makes in selecting monographic. The
PLEXUS all of these systems are advisory systems for prerequisite is that the knowledge base has to be broad
locating reference resources and factual data. enough and the interfacing aspect must be easy enough
for the library to get the desired information from the
2) Cataloging is one of the oldest library techniques. machine.
Recent attempts to automate cataloging through expert
systems have focused on descriptive cataloging because B. Natural Language Processing in Library Services
it is rule-based (AACR2). There are two ways to apply When we think of the term NLP, the first thing that comes to
artificial intelligence techniques in cataloging: (a) mind is the ability to speak or write a complete sentence and
Human-machine interfaces, where intellectual work is have a machine process of requesting and speaking. NLP can
divides between the intermediary and the support applied to many disciplines, including libraries. When apply
system. (b) An expert system with full cataloging to the field of library and information science, more
capabilities associated with electronic publishing specifically, to search databases such as the Online Public
systems. Since the cataloging text is generates online, it Access Catalogue (OPAC), indexing is the basis of document
can be passed through a knowledge-based system, and retrieval. The purpose of the index is to improve the
the intermediary does the cataloging process without any precision of retrieving parts of the relevant documents; and
intellectual input. to reduce the proportion of recalls and related files retrieved.

3) Classification is the basic activity of a knowledge C. Machine learning in Library Services

organization. Therefore, it is prominent in all systems One specific challenge that is ripe is the improvement of
that organize knowledge in libraries and information library metadata generation. Libraries, through various
centers. The application of expert systems in the field of vendors as part of the purchasing and acquisitions process,
library classification includes Coal SORT, EP-X, and acquire thousands of pieces of metadata for print and digital
BIOSIS. resources made available to their library users. In cases
where an e-book platform does not include metadata,
libraries they generate their own. For the increasing majority

© 2019, IJCSE All Rights Reserved 138

International Journal of Computer Sciences and Engineering Vol. 7(16), May 2019, E-ISSN: 2347-2693

of born-digital resources, machine learning provides an array V. CONCLUSION

of possible tools to help libraries generate metadata for
digital resources, allowing cataloguing to not only increase There are a number of possible applications of Artificial
the speed of metadata generation but also vastly improve the Intelligence implemented and they have been creating a
depth and breadth of subject terms. positive impact on libraries. This has proved that applications
of AI saves time and money in almost all sectors in the
D. Robotics in the Library Services society. The application of AI in the academic libraries have
The robot is "A reprogrammable, multipurpose manipulator, been increasing in very high speed. As authors of this paper
automatically controlled, programmable in three or more discussed, implementation of AI in libraries has triggered the
axes, which can be fixed on the location or portable for use discovery of many new ideas. The development of expert
in automation applications". Libraries providing an increased system libraries greatly benefited, sometimes it appears like
variety of services and resources for digital libraries, they are “Librarianship is at stake” and now it is challenging to ensure
still acquiring a great number of printed documents. This the values of librarianship. Artificial intelligence (AI)
combined pressure to provide electronic and printed systematically tops popular lists of the most imperative
resources and services has caused serious space constraints emerging technologies. With a mixed feeling of fear and
for many libraries, especially academic libraries and eagerness, readers seem to agree that the AI shapes the future
research. The objective of CAPM (comprehensive approach libraries.
to printed material) is to build a personalized robotic
scanning system based on a series, which allows the REFERENCES
browsing of imprints in real time via the web interface. The
user includes a CAPM system that, in turn, starts a robot that [1] A. A. Asefeh Asemi, "Artificial Intelligence (AI) application in
recovers the item requested. This item iss delivered to Library Systems in Iran: A taxonomy study," Library Philosophy
and Practice, vol. 2, March-April 2018.
another robotic system, which opens the item and rotates
[2] S. S. Mogali, "Artificial Intelligence and its applications in
pages automatically. Libraries," 2014.
[3] J. Charles W. Bailey, "Intelligent library systems: artificial
E. Intelligent Interfaces to Online Databases intelligence technology and library automation systems,"
Online access to databases is still difficult for many potential Advances in Library Automation and Networking, vol. 4, pp. 1-23,
users. The user may need to know different communication 1991.
[4] W. Kristin, "Libraries in an Artificially Intelligent World," Public
protocols, master language control, search techniques, Libraries, 2016. [Online]. Available: http://publiclibrariesonli
database file structures, and terminological terminology. The ne.org/2016/02/libraries-in-an-artificially-intelligent-world/
aim of the intelligent interface is to facilitate the access to the [Accessed Feb 2016].
construction of some of the necessary knowledge in the [5] B. Johnson, "Libraries in the age of artificial intelligence,"
front-end software used to test the online search system. This Computers in Libraries, vol. 38, 2018.
goal does not coincide with the goal of creating an intelligent [6] G. Liu, "The application of intelligent agents in libraries: a
survey," Electronic library and information systems, vol. 45, pp.
search system. The interface of access to existing online 78-97, 2011.
systems, with all their limitations and disadvantages, so it [7] N. Jacknis, "The AI-Enhanced Library," [Online]. Available:
can be equally successful as an on-line search system. The https://medium.com/@ NormanJacknis/the-ai-enhanced-library-
interface does not solve the problem of restructuring the a34d96fffdfe [Accessed 2019 Jun 2017].
database, but rather allows the search system itself to make [8] Merriam-webster, "Definition of Artificial Intelligence," 2019.
[Online]. Available: https://www.merriam-
the approach more intelligent.
webster.com/dictionary/artificial+intelligence. [Accessed Feb
2019].
Searching online databases can be helpful in these ways:
 select the appropriate hosts and databases; Authors Profile
 allow the seeker to state an information want in their Mr. S Vijayakumar working as Assistant
own terminology; Librarian in Presidency University,
 determine the level and access to the information Bangalore since Aug 2017. He pursed
requested; Bachelor of Arts from University of
 adapt the extent of the information to be retrieved; Mysore in 2014 and Master of Library and
 formulate the vocabulary query used in the selected Information Science from University of
databases; Mysore in the year 2016. He is currently
 express a search query in the format required; and pursuing Ph.D. in Department of Library & Information
 Present search results in a helpful way, e.g. ranking in Science, University of Mysore. He received 4 Gold Medals
the order of probability of relevance. and One Cash Prize from University of Mysore for securing
First Rank in MLISc in 2016. He is qualified NET & KSET.

© 2019, IJCSE All Rights Reserved 139

 International Journal of Computer Sciences and Engineering Vol. 7(16), May 2019, E-ISSN: 2347-2693

He has published 8 research papers in national &

international journals, conferences and edited books.

Dr. K N Sheshadri is the Senior Librarian at

Presidency University, Bengaluru since Jan
2016. He has over 16 years of professional
experiences. He worked as Librarian in
BITS Pilani Dubai Campus in Dubai, UAE
for 12 years. Prior to join BITS, he worked
as information officer for one year at
Jubilant Biosys (Bioinformatics Company) at Bengaluru. He
received his doctoral degree from Mangalore University in
the year 2013 under the supervision of Prof. Dr. D.
Shivalingaiah. His topic of research was “Networking of
Technical Libraries in UAE: A Consortium Model”. He has
more than 20 research articles in national & international
journals and international conferences. He has guided 4
MPhil dissertations in Library and Information Sciences.
Also he has served as examiner for 2 PhD candidates in
Library and Information Sciences. Presently guiding two
PhD candidates in Library and Information Science,
Presidency University.

© 2019, IJCSE All Rights Reserved 140

View publication stats