Artificial Intelligence

2101 Submissions

[6] viXra:2101.0168 [pdf] submitted on 2021-01-27 06:10:38

Recent Trends in Named Entity Recognition (NER)

Authors: Arya Roy
Comments: 27 Pages.

The availability of large amounts of computer-readable textual data and hardware that can process the data has shifted the focus of knowledge projects towards deep learning architec- ture. Natural Language Processing, particularly the task of Named Entity Recognition is no exception. The bulk of the learning methods that have produced state-of-the-art results have changed the deep learning model, the training method used, the training data itself or the encoding of the output of the NER system. In this paper, we review significant learning methods that have been employed for NER in the recent past and how they came about from the linear learning methods of the past. We also cover the progress of related tasks that are upstream or downstream to NER eg. sequence tagging, entity linking etc. wherever the processes in question have also improved NER results.
Category: Artificial Intelligence

[5] viXra:2101.0163 [pdf] submitted on 2021-01-26 20:22:30

Forecasting Stock Market Price Using Multiple Machine Learning Technique

Authors: Tanvir Rahman, Rafia Akhter
Comments: 5 Pages.

The stock market is an emerging sector in any country in the world. Many people are directly related to this sector. Stock market prediction is the act of trying to determine the future value of company stock or another financial instrument. When publicly traded, companies issue shares of stock to investors, every one of those shares is assigned monetary value or price. Stock prices can go up or down depending on different factors. Stock prices can be affected by several things including volatility in the market, current economic conditions, and the popularity of the company. The successful prediction of a stock's future price could yield a significant profit. Along with the development of the stock market, forecasting has become an important topic. Since the finance market has become more and more competitive, stock price prediction has been a hot research topic in the past few decades. Predicting stock price is regarded as a challenging task because the stock market is essentially nonlinear, on-parametric, noisy, and a chaotic system. The trend of a market depends on many things like liquid money human behavior, news related to the stock market, etc. All this together controls the behavior of trends in a stock market with the advancement of the computing technology we use machine learning techniques, like Support Vector Regression, K-nearest neighbor, liner Regression, Random Forest Regression, for analyzing time-series data to predict stock price. In this paper, we try to develop a forecasting model by stacking multiple methods to find the best forecast of the stock price.
Category: Artificial Intelligence

[4] viXra:2101.0122 [pdf] replaced on 2021-07-14 12:55:15

Simplifying Object Segmentation with PixelLib Library

Authors: Ayoola Olafenwa
Comments: 8 Pages.

PixelLib is a library created to allow easy implementation of object segmentation in real life applications. In this paper we discussed in detail how PixelLib makes it possible for developers to implement semantic segmentation, instance segmentation, extraction of objects and background editing in images and videos with great simplification.
Category: Artificial Intelligence

[3] viXra:2101.0115 [pdf] submitted on 2021-01-18 04:51:58

CNN Based Common Approach to Handwritten Character Recognition of Multiple Scripts

Authors: Durjoy Sen Maitra, Ujjwal Bhattacharya, SK Parui
Comments: 5 Pages. Paper published in ICDAR 2015

There are many scripts in the world, several of which are used by hundreds of millions of people. Handwrittencharacter recognition studies of several of these scripts arefound in the literature. Different hand-crafted feature sets havebeen used in these recognition studies. However, convolutionalneural network (CNN) has recently been used as an efficientunsupervised feature vector extractor. Although such a networkcan be used as a unified framework for both feature extractionand classification, it is more efficient as a feature extractor than asa classifier. In the present study, we performed certain amount of training of a 5-layer CNN for a moderately large class characterrecognition problem. We used this CNN trained for a larger classrecognition problem towards feature extraction of samples of several smaller class recognition problems. In each case, a distinctSupport Vector Machine (SVM) was used as the correspondingclassifier. In particular, the CNN of the present study is trainedusing samples of a standard 50-class Bangla basic characterdatabase and features have been extracted for 5 different 10-classnumeral recognition problems of English, Devanagari, Bangla,Telugu and Oriya each of which is an official Indian script.Recognition accuracies are comparable with the state-of-the-art
Category: Artificial Intelligence

[2] viXra:2101.0089 [pdf] submitted on 2021-01-14 12:47:14

Introduction to CAT4: Part 1. Axioms

Authors: Andrew Holster
Comments: 36 Pages. [Corrections made by viXra Admin to conform with scholarly norm]

CAT4 is proposed as a general method for representing information, enabling a powerful programming method for large-scale information systems. It enables generalised machine learning, software automation and novel AI capabilities. It is based on a special type of relation called CAT4, which is interpreted to provide a semantic representation. This is Part 1 of a five-part introduction. The focus here is on defining the key mathematical structures first, and presenting the semantic-database application in subsequent Parts. We focus in Part 1 on general axioms for the structures, and introduce key concepts. Part 2 analyses the CAT2 sub-relation of CAT4 in more detail. The interpretation of fact networks is introduced in Part 3, where we turn to interpreting semantics. We start with examples of relational and graph databases, with methods to translate them into CAT3 networks, with the aim of retaining the meaning of information. The full application to semantic theory comes in Part 4, where we introduce general functions, including the language interpretation or linguistic functions. The representation of linear symbolic languages, including natural languages and formal symbolic languages, is a function that CAT4 is uniquely suited to. In Part 5, we turn to software design considerations, to show how files, indexes, functions and screens can be defined to implement a CAT4 system efficiently.
Category: Artificial Intelligence

[1] viXra:2101.0088 [pdf] submitted on 2021-01-14 12:53:01

Introduction to Cat4: Part 2. Cat2

Authors: Andrew Holster
Comments: 56 Pages. [Corrections made by viXra Admin to conform with scholarly norm]

CAT4 is proposed as a general method for representing information, enabling a powerful programming method for large-scale information systems. It enables generalised machine learning, software automation and novel AI capabilities. It is based on a special type of relation called CAT4, which is interpreted to provide a semantic representation. This is Part 2 of a five-part introduction. The focus here is on defining key mathematical properties of CAT2, identifying the topology and defining essential functions over a coordinate system. The analysis is from first principles. This develops on from the axioms introduced in Part 1. The interpretation of fact networks is introduced in Part 3, and the full application to semantic theory comes in Part 4, where we introduce general functions, including the language interpretation or linguistic functions. In Part 5, we turn to software design considerations, to show how files, indexes, functions and screens can be defined to implement a CAT4 system efficiently.
Category: Artificial Intelligence