Accelerat ing t he world's research.

Advances in Natural Language

Processing – A Survey of Current
Research Trends, Development Tools
and Industry Ap...
Krishna Prakash

International Journal of Recent Technology and Engineering (IJRTE)

Cite this paper Downloaded from Academia.edu 

Get the citation in MLA, APA, or Chicago styles

Related papers Download a PDF Pack of t he best relat ed papers 

Conversat ional Art ificial Int elligence (AI) -demyst ifying st at ist ical vs linguist ic NLP Solut ions
Dr Kulvinder Panesar

Underst anding EFL Linguist ic Models t hrough Relat ionship bet ween Nat ural Language Processing and…
Arab World English Journal (AWEJ)

Comput at ional Int elligence in Chat Bot s.docx

anil kumar bheemaiah
 International Journal of Recent Technology and Engineering (IJRTE)
ISSN: 2277-3878, Volume-7, Issue-5C, February 2019

Advances in Natural Language Processing –

A Survey of Current Research Trends,
Development Tools and Industry
Applications
Krishna Prakash Kalyanathaya, D. Akila and P. Rajesh 

Abstract--- Natural Language Processing (NLP) is a communication. The machine learning and deep learning
subfield of Artificial Intelligence and getting lot of focus on algorithms are widely is used to develop frameworks for
research and development due to emergence of its NLP and efficiently perform common NLP tasks.
applications. The research areas in focus are conversation
systems, Language processing, Machine Translation, Deep
learning. The researches in these areas lead to development of II. BRIEF HISTORY
many tools to build industrial applications. Combining Deep Although the work on NLP dates back to around 1950
Learning techniques with Natural Language Processing is
with the development of what is called as “Turing Test”
finding lot of applications in domains such as Healthcare,
Finance, Manufacturing, Education, Retail and customer and in 1957 a rule based system of syntactic structures.
service. This paper provides bird’s view of advancement in The progress was slow until 1990 due to limited
research, development and application areas of Natural computational power and systems were based on
Language Processing. This paper captures21research focus complex sets of hand written rules and limited
areas, 22 development tools and 6 domains where Natural vocabulary. With the introduction of machine learning
Language Processing are making rapid advancements.
and steady increase in computational power, recently
Key terms used in this text: NLP, Natural Language
Processing, Deep Learning, Sentiment Analysis, Question interest on research and applications are growing. The
Answering, Dialogue Systems, Parsing, Named-Entity recent major breakthrough areas of NLP are: speech
Recognition, POS Tagging, Chatbots, Human-Computer- recognition, language processing, dialog systems and
Interface. applying deep learning techniques.
While NLP is still facing lot of challenges (like human
I. INTRODUCTION computer interfaces), there has been lot of research
Natural Language processing (NLP) is a subfield of interests and it has opened to many opportunities for
artificial intelligence dealing with computational using the techniques in robotics, automation and digital
algorithms to automatically represent and process various transformation.
forms of human (natural) language inputs and
communicate with Human-Computer-Interface (HCI). It III. RESEARCH WORKS ON NLP
is also known by a name “Computational Linguistics”. Until 1990, most of the research work was done on the
Natural Language Processing involves following NLP concepts and machine translation. Most recent
stages of processing namely, lexical (structure) analysis, research work on NLP have harnessed the power of
parsing, semantic analysis, discourse integration, and statistical models, machine learning, Deep learning
pragmatic analysis. Some well-known application areas technologies that are using data driven approach.
of NLP are Speech Recognition, Optical Character The research topics on Natural Language Processing
Recognition (OCR), Machine Translation, and Chatbots. sometimes overlap with some artificial intelligence and
Recently, Machine Learning algorithms are used to Deep Learning topics. These approaches generally
process Natural Language input by studying millions of adopted recently to perform NLP tasks in most efficient
examples of text — words, sentences, and paragraphs — way. The ACL 2018 Main Conference invited papers in
written by humans. By studying these samples, training 21areas which are Dialogue and Interactive Systems,
algorithms gain an understanding of the “context” of Discourse and Pragmatics, Document Analysis,
human speech, writing, and other modes of Generation, Information Extraction and Text Mining,
Linguistic Theories, Cognitive Modeling and
Revised Version Manuscript Received on 22 February, 2019 Psycholinguistics, Machine Learning, Machine
Krishna Prakash Kalyanathaya, M.Phil, Research Scholar,
Department of Computer Science, VELS Institute of Science,
Translation, Multidisciplinary, Multilinguality,
Technology & Advanced Studies, Chennai, Tamilnadu, India. Phonology, Morphology and Word Segmentation,
(e-mail:krishna.prakash.kk@ gmail.com) Question Answering, Resources and Evaluation,
D. Akila, Associate Professor, Department of Information Sentence-level Semantics, Sentiment Analysis and
Technology, School of Computing Sciences, VELS Institute of Science,
Technology & Advanced Studies, Chennai, Tamilnadu, India. Argument Mining, Social Media, Summarization,
(e-mail:[email protected])
P. Rajesh, Assistant Professor, Department of Information
Technology, School of Computing Sciences, VELS Institute of Science,
Technology & Advanced Studies, Chennai, Tamilnadu, India.
(e-mail:[email protected])

Published By:
Blue Eyes Intelligence Engineering
Retrieval Number: E10480275C19/19©BEIESP 199 & Sciences Publication
International Conference on Advances in Signal Processing, Power, Embedded, Soft Computing, Communication and
Control Systems (ICSPECS-2019) | 11th & 12th January 2019 | GPREC, Kurnool, A.P. India

Tagging, Chunking, Syntax and Parsing, Textual D. Healthcare

Inference and Other Areas of Semantics, Vision, Hospitals are deploying Virtual Assistants developed
Robotics, Multimodal, Grounding and Speech, Word- with combination of Natural Language Processing,
level Semantics [3][4]. The conference accepted 258 long Computer Vision and Machine learning which will
papers and 156 short papers on the above areas. automatically create and retrieve patient history by
interacting with the Patients. Virtual Assistant handle
IV. INDUSTRIAL APPLICATIONS OF routine tasks such as scheduling appointments and
NLP registration of patients.
NLP aims to dominate human-to-machine interaction E. Automotive
to the point where talking to a machine is as easy as Self driving cars are one of most significant
talking to a human. NLP still continues to harness innovations in this field. NLP enables human computer
unstructured data and make it meaningful to a machine. interface in what is now called as “in-car assistants”
IDC recently forecasted that the quantity of analyzed data which is gaining popularity in the industry.
by cognitive systems will grow by a factor of 100 to 1.4
ZB by 2025 impacting thousands of industries and F. Finance
companies around the globe [2]. Robotics, health care, NLP based solutions are developed in applications
financial services, connected auto and smart homes are such as credit scoring, sentiment analysis and document
some of the sectors that will continue to be advanced by search. Credit scoring application help the banks and
NLP. financial institutions assess an individual’s
One of initial applications of NLP in early years of creditworthiness and provide credit score using NLP and
2000 was on machine translation to work as translator machine learning. In sentiment analysis applications
from one human language to another. But it rapidly found automate the tasks of text mining for real time
its acceptance in customer service industry. Most popular information on market from news site and social media
application of NLP in customer service is called as and then perform document classification and named
“Chatbots” or Virtual assistant. entity recognition to filter out most relevant information
Industrial applications of NLP can be broadly to the investors needs.In document search applications,
classified into 3 categories:Conversational systems, Text the banks or financial institutions uses chatbot interface
Analytics, Machine translation that enables its customers to search for information and
get answers to basic transactional queries.
A. Conversational Systems
Conversational system allows us to make the G. Manufacturing
conversation with the automated system in a natural Robotics and Process Automation are two very
language via voice or text interface. They help to promising applications areas that will employ Natural
automate the complex workflows in an organization with Language Processing capabilities. Using NLP, Robot on
24X7 support to its users. a manufacturing floor can interact with human operator at
Most common type of conversation devices are a remote place and process instructions for assembly and
Chatbots and Virtual Assistants. Today, these two movement of machines and products. The industries have
devices are employed by banks, e-commerce, social already begun to use NLP based solution to automate
media and other self service point of sales systems to their process workflow for improving efficiency of
provide various services to their customers. operations.
B. Text Analytics H. Retail
Text Analytics also called as Text mining which aims Using Natural Language, Computer Vision and
to extract meaningful content from text, either in Machine learning technologies, a Virtual Assistant placed
documents, emails or short-form communications such as in front of the Retail shop can identify and know what
tweets and SMS texts [2]. customer wants and provide them quick information and
Most common use cases of text analytics on social promotional offers. The specialized virtual assistant also
media analytics. understand the shoppers comments and complaints
regarding the products and provide a quick direction for
C. Machine Translation
resolution. Many e-commerce platforms are currently
Machine translation is the task of automatically using conversational agents for customer support.
translating one natural language into another, preserving
the meaning of the input text [4]. I. Education
Most popular application of machine translation is A combination of Natural Language Processing and
Google translator. Other machine translation software’s Computer vision enable a platform to deliver virtual
are also used in speech translation and teaching. classroom for students. We have already seen the digital
Now, we will look at some industrial applications in assistants helping students solving the problems from
following domain areas: Healthcare, Automotive,
Finance, Manufacturing, Retail, Education and customer
service.

Published By:
Blue Eyes Intelligence Engineering
Retrieval Number: E10480275C19/19©BEIESP 200 & Sciences Publication
 International Journal of Recent Technology and Engineering (IJRTE)
ISSN: 2277-3878, Volume-7, Issue-5C, February 2019

expert knowledge available in digital libraries. Language transformation will consists of suite of various
The top vendors delivering NLP solutions are IBM, learning, extraction algorithms to extract meaningful
Google, Amazon, Cerner, Nuance Communications and actions from the NLP tasks. Natural Language
Microsoft. communication is a presentation of the desired actions as
an outcome of the NLP tasks. The generated outcome
V. DEVELOPMENT FRAMEWORKS may be either in Natural Language or in the form of
AND TOOLS FOR NLP computer action such as robot arm movement.
The following table gives an overview of the popular
The development frameworks and tools will help to tools currently available in the market.
build industrial applications discussed in the previous
section. There has been numerous development tools
available today due to significant interest shown by open
Abbreviations
1
source communities around the world. These frameworks Common NLP tasks: Lexical analysis, Parsing,
and tools provide built in libraries and also customizable semantic analysis, POS tags, Chunking, Name Entity
to adapt specific needs of the industry. Recognition (NER), Semantic Role Labeling
(SRL)and Syntactic parsing.
2
NLTK: Natural Language Toolkit
3
UIMA: Unstructured Information Management
Architecture
4
GATE: General Architecture for Text Processing
5
GPL: General Public License
6
Dynet: Dynamic Neural Network
7
MALLET: Machine learning for Language Toolkit
Fig 1: Block representation of stages in the 8
LUIS: Language Understanding Intelligent Service
development of NLP tools
Fig 1 shows the block representation of various stages
in NLP application development. The Natural Language REFERENCES
acquisition block built with speech processing, Computer 1. Daniel W. Otter, Julian R. Medina, and Jugal K. Kalita.
vision or any data acquisition tools to inject Natural 2018. A Survey of the Usages of Deep Learning inNatural
Language Processing. 1, 1 (July 2018), 35 pages.
language text into the system. 2. ROBERT DALE. "The commercial NLP Landscape in
Table 1: Natural Language Processing Tools 2017", Article in Natural LanguageEngineering, July2017
3. ACL 2018: 56th Annual Meeting of Association for
Computational Linguistics https://acl2018.org
4. Predictive Analytics Today:
www.predictiveanalyticstoday.com[accessed in Dec 2018]
5. Ali Shatnawi, Ghadeer Al-Bdour, Raffi Al-Qurran and
Mahmoud Al-Ayyoub 2018. A Comparative Study of Open
Source Deep Learning Frameworks. 2018 9th International
Conference on Information and Communication Systems
(ICICS)
6. Intelligent automation: Making cognitive real Knowledge
Series I Chapter 2. 2018, EY report.
7. Jacques Bughin, Eric Hazan, SreeRamaswamy, Michael
Chui , TeraAllas, Peter Dahlström, Nicolaus Henke, Monica
Trench, 2017. MGI ARTIFICIAL INTELLIGENCE THE
NEXT DIGITAL FRONTIER? McKinsey & Company
McKinsey & Company report July 2017
8. Svetlana Sicular, Kenneth Brant 2018, Hype Cycle for
ArtificialIntelligence, 2018 Gartner report July 2018.
9. McCallum, Andrew Kachites. "MALLET: A Machine
Learning for Language Toolkit." http://mallet.cs.umass.edu.
2002.
10. Quarteroni, Silvia. (2018). Natural Language Processing for
Industry: ELCA’s experience. Informatik-Spektrum. 41.
10.1007/s00287-018-1094-1.
11. Young, Tom &Hazarika, Devamanyu&Poria, Soujanya&
Cambria, Erik. (2018). Recent Trends in Deep Learning
Based Natural Language Processing [Review Article]. IEEE
Computational Intelligence Magazine. 13. 55-75.
10.1109/MCI.2018.2840738.
12. Amirhosseini, Mohammad Hossein, Kazemian, Hassan,
Ouazzane, Karim and Chandler, Chris (2018) Natural
Natural Language representation block uses structured, language processing approach to NLP meta model
tree or graph models to represent the Natural Language automation. In: International Joint Conference on Neural
understanding. Natural Language database is a repository Networks (IJCNN), 8-13 July 2018, Rio de Janeiro, Brazil.
of Natural Language data like MNIST or similar
databases which are then used by machine learning
algorithms to perform other NLP tasks.
This database is accessed by representation and
transformation blocks to perform their tasks. Natural

Published By:
Blue Eyes Intelligence Engineering
Retrieval Number: E10480275C19/19©BEIESP 201 & Sciences Publication