1 This paper describes about
neural machine translation which is new approach for translating. it creates a
neural network which increase the translation performance. Here the neural
network model uses encoder -decoder architecture, which encode the source
sentence and decode it to destination sentence by predicting the target words.
With this approach machine translation can be achieved.
2 DNN is powerful model used in
many field it is suitable for performing difficult learning tasks.in this paper
we discuss about seq2seq approach including LSTM to map the input sequence to a
vector fixed dimension. it uses encoder-decoder RNN which encode the source
sentence and decode the sentence to provide target words.
3 Conservation model is important task in machine learning
as it understands the question and reply for those question. In this paper the
conservation model is built by using sequence to sequence framework. Replies
are generated by training large datasets. conservation model can be open domain
or closed domain. the conservation model can vary based on the datasets example
IT helpdesk datasets can be used to provide solutions for technical problem.
4 Here open domain conservation system is build based on
large dialogue corpora using generative model.it produces system response by
understanding the questions with help of RNN encoder- decoder. this paper discusses
about the limitation of this approaches and show how the maximize the
5 In some cases seq2seq neural networks can produce
responses which does not suites to the responses so usage of Maximum Mutual
information is increased for producing more interesting and more appropriate
6 This paper proposes that in seq2seq neural networks
attention and intention plays an important role. so, in order to focus on
intention neural network should consist three recurrent networks. The encoder
layer for word level model represent score side sentence. The intention network
layer for modeling the dynamics of the intension process. The decoder layer for
produce the response to the input. Neural network improves the efficiency of
7 Here persona-based model is created in order to reduce
human work. it replaces the human and provide human like responses it captures
individual characteristics such as background info and speaking style.
8 In seq2seq learning (copying) i.e. certain contents are
selectively repeated in the output sequence. A similar way in humans, replicate
certain content. Here we incorporate copying in the neural network and build a
new model called Copy net, involved in encoder and in decoder. The Copy net is
designed in such a way that it generates a response by copying mechanism and
use them in proper place in the output sequence.
9 This paper describes about the chat- bot used in twitter
social network for entertainment and for advertisement. these bots use twitter
datasets and provide a realistic conversation like human
10 this paper describe application that incorporate human appearance
and simulate human dialog.
the knowledge of bot is stored in a database by the developers. This is just an
initial stage for AI bot start to interact with humans later bots can be built
in human appearance and are used for substitute for men.
11 this paper discusses about the understanding and analyses
the intelligence of chatbot. The analysis is done in order to check the intelligence
of the chatbot. There are various parameters to consider like text
categorization, entity extraction, frequency analysis and model the vocabulary
using word to vector system. Here this analysis provide metric called bot
intelligence score to evaluate.
12 This paper proposes a hybrid neural network model which
contains of some important network model. here the datasets are trained and
experimented results show that the best accuracy belong to different hybrid
13 here the human machine interface demand for a clean
communication that are applied on various tasks. here we develop an intelligent
chat machine which generate conservation sentence using RNN and single neural
network model that process a conversation sentence by connecting the words.