ABSTRACT to create virtual assistant like Siri (Apple), Cortana

ABSTRACT

Natural language processing (NLP) is a recent developing
area of research and innovation which is        
gaining a lot of popularity these days. Natural language processing is a
part of AI which focuses on human like natural language processing and is
extensively used in developing virtual assistants and chat bots .Several
companies has already applied NLP techniques to create virtual assistant like
Siri (Apple), Cortana (Microsoft), Alexa (Amazon) etc. This paper represents
ideas for further developing virtual assistants so they can be used in daily
life. Some deep learning can be introduced in natural language processing which
may help us build a more advanced virtual assistant.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

INTRODUCTION

Machine learning and deep learning  fueled by big data are helping mankind to
solve intelligence and companies like DeepMind and OpenAI are pushing the
boundaries of AI to make it more practical and usable in our day to day life.
One Such application is creating virtual assistants that requires concepts from
disciplines like NLP and Deep Learning. One such concept
is recursive neural network as mentioned in 1,it takes and understands in
information by the previous stored information in its hidden layer which is of
fixed size.

LITERATURE SURVEY

Linguistic and computational knowledge is needed  for NLP.while extracting the information
there are many problems like paraphrasing,understanding idioms and metaphors.By
introduncing the concepts of deep learning in NLP the machine will be able to
understand the language better like use of adjectives and sarcastic sentences.The
2 major classifications of approaches for NLP 1 are:

1. Statistical learning approach     2. Rule-based approach

Since textual tree construction can be time consuming for
long sentences, it is inefficient. Recurrent Neural Network can extract
contextual information by utilizing stored previous text in the form of fixed
sized hidden layer.This deeper learning in language processing will be
beneficial for further developments of virtual assistants.The idea of  use virtual assistants for self driven cars
have also been proposed2,embedding autonomy,deep learning and adaptability
can help improve and drive better than humans.The human brain of drivers is
emotionally aware and developed and it takes decisions according to the gut
feeling and past experiences but this is not so in machines and virtual
assistants.The combination of artificial neural networks and deep learning will
simplify the task of replicating the functioning of human brain in machines and
make the artificial brains emotionally aware.The artificial brain of virtual
assistants should be able to take independent decisions,obey traffic
laws,optimize fuel consumption,reduce pollution and ensure better safety
features for the passengers.In the workshop “talking with conversational agents
in collaborative action” as mentioned in 2 in mobile phones the evolution
from touch to speech interface idea was proposed with smart T.V.s,smart watches
etc. which are now have been successfully implemented.In virtual assistant
there is typical use of speech based queries.The idea of”virtual Buttler” was
also proposed in a workshop which will act like a real person and practices long
term  use of a companion.Text to speech
and speech to text are very popular applications of NLP used in creating
chatbots.The idea of expressive virtual text to speech system was presented
using deep neural network 3.In this a text sentence will be given with a
expressive tag which will be used to produce a photorealistic talking head.The
speaker adaption technique uses small training data from a novel speaker to
modify a system that has already been trained.The system can be modified in
three ways3:

1.Speaker specific code is appended to input 2.Speaker
specific reweighing of hidden contributions is learnt  3.Speaker dependent mapping is learnt from
each speaker.

In the visual model of virtual assistant image of face is
constructed by separating modes into semantically meaningful actions and
regions.One Mode is used for model blinking,two modes are used for 3D head
rotation,eight modes are used to model lower half of the face and six modes are
used to model upper half of the face. There are expression space final
expression dependent layers for mapping to specific expressions already coded
called “Expression Adaptation”3.Some new expressions can also be added to the
model with small amount of adaptation data and new output layer can be added
for  the same.The linguistic feature is
sent to the penultimate layer of previously trained Deep Neural Network for
producing final expressions in the virtual head model.Regularised form of
linear least square algorithm should be used to avoid overfitting on small
amount of adaptation data.Different layers are formed for different expressions
like anger,fear.happy,sad and tender.The vehicles use sensors for vehicle to
vehicle communication.When a lead vehicle suddenly deaccelerates a MATLAB
simulink model varies vehicle dynamics,following vehicle’s distance and initial
speeds.Artificial Intelligence can improve cases of escaping injury,vehicle
damage and traffic jams.The driving assistant uses telematics,global
positioning and sensors ,it can avoid potential dangers.The warning methodology
is constructed through  “spatio temporal
safety zones”4,the three zones are:

1.Accident Mitigation (AM)  
2.Accident Avoidance(AA)  3.
Accident Free(AF)

After examining these three zones the system increases the
radius of the vehicle concentrically from the vehicle’s center. Each vehicle
monitors other vehicles constantly in case the vehicle assistant needs to
react.Feed forward neural network approach is used  if any vehicle suddenly applies brakes in
front of the vehicle following it. Another application of virtual assistant has
been proposed in the field of medicine “The Mindbot” 5 for mental health
problems.Psychotherapies have been developed to identify facts  that cause mental illness.This virtual agent
will behave as a real human therapist and will ask mental health related
questions from the patients and the answers will be evaluated according to the
standard psychological scales and results will show the mental fitness level.The
concept of “Listen while speaking” 6 for virtual assistant in all the devices
can soon be introduced by the use of peech chain in deep learning.The speech
chain is a process that helps replicating the exact human speech using speech
chain algorithm as referred in 6.The technology of automatic speech
recognition (ASR)and text to speech recognition(TTS) enables machine to process
and respond to human speech.

FUTURE SCOPE

The use of virtual assistants is gaining a lot of popularity
in many fields and the interest in deep learning.In future the virtual
assistants will help humans in their day to day life.We will be able to embed
virtual  assistants in many public places
like railway stations,hospitals, banks.Installing virtual assistants in ATMS will
help  visually challenged people to
access the facility easily without being dependent on theIrs ans we can further
widen their usability by introducing virtual assistants in different languages
including the local languages of different states of India.