| viXra.org > Artificial Intelligence > 2308 |
 |
 |
| viXra.org > Artificial Intelligence > 2308 |
|
|
[8] viXra:2308.0137 [pdf] replaced on 2023-09-30 22:42:32
Authors: Victor V. Senkevich
Comments: 16 Pages.
All magic and mystery disappear as soon as an obscure mysterious concept gets a rigorous formal definition. In order to provide an opportunity to talk about the applicability of philosophical / cognitive concepts to the subject area of AI, it is necessary to "ground" these concepts by formulating rigorous formal definitions for them. The fundamental importance of such formal definitions is quite obvious, since any concepts applied to the field of Information Technology must be "codable", i.e. potentially implementable in program code. Thus, the "codable" formal definitions of cognitive terms are the necessary basis on which alone it is possible to build the architecture of AI technology that has the ability to embody these concepts in a real software. The question of the adequacy of such definitions of "reality" and their compliance with existing generally accepted philosophical theories is also very important and quite discussable, but this does not affect the priority and fundamental nature of the requirement for the formulation of "codable" formal definitions. The formulation of "codable" definitions for the concept of "consciousness" and related cognitive concepts and, based on them, statements about their applicability to the subject area of AI is the topic of this publication. Covering questions:Can AI have a Personality / Motivations / Free Will?
Category: Artificial Intelligence
[7] viXra:2308.0116 [pdf] replaced on 2023-11-12 21:54:59
Authors: Youming Zhao
Comments: 10 pages, fixed two mistakes
We present an alternating direction method of multipliers (ADMM) for a generic overlapping group lasso problem, where the groups can be overlapping in an arbitrary way. Meanwhile, we prove the lower bounds and upper bounds for both the $ell_1$ sparse group lasso problem and the $ell_0$ sparse group lasso problem. Also, we propose the algorithms for computing these bounds.
Category: Artificial Intelligence
[6] viXra:2308.0112 [pdf] submitted on 2023-08-17 22:48:28
Authors: Nana Abeka Otoo
Comments: 12 Pages.
Mutation validation as a complement to existing applied machine learning validation schemes hasbeen explored in recent times. Exploratory work for Learning vector quantization (LVQ) based onthis model-validation scheme remains to be discovered. This paper proposes mutation validation as an extension to existing cross-validation and holdout schemes for Generalized LVQ and its advanced variants. The mutation validation scheme provides a responsive, interpretable, intuitive and easily comprehensible score that complements existing validation schemes employed in the performance evaluation of the prototype-based LVQ family of classification algorithms. This paper establishes a relation between the mutation validation scheme and the goodness of fit evaluation for four LVQ models: Generalized LVQ, Generalized Matrix LVQ, Generalized Tangent LVQ and Robust Soft LVQ models. Numerical evaluation regarding these models complexity and effects on test outcomes,pitches mutation validation scheme above cross-validation and holdout schemes.
Category: Artificial Intelligence
[5] viXra:2308.0077 [pdf] submitted on 2023-08-12 12:07:43
Authors: Ahmed Taha Hassina
Comments: 10 Pages.
Mapping the universe has always been a salient endeavor in astronomy and astrophysics. Advancements in observational astronomy have generated vast amounts of data containing various features of celestial objects. Inducing a growing need for accurate and detailed classification and localization of stellar objects in the cosmos. In this paper, we present a comprehensive study that combines machine learning techniques to classify celestial objects into distinct categories and predict their precise locations in the sky. This study is divided into two parts: a classification task, where the stellar objects are classified into galaxies, stars, or quasars (quasi-stellar radio sources). The resulting model exhibits exceptional performance in differentiating these objects, as demonstrated by high classification accuracy. We extend our analysis to predict the location of stellar objects using regression techniques. By employing multi-target regression, we model the right ascension and declination coordinates, enabling accurate localization of celestial objects on the celestial sphere. The practical implications of our research lie in producing comprehensive celestial catalogs, facilitating targeted observations, and contributing to the broader field of observational astronomy. The ability to accurately classify and localize stellar objects lays the groundwork for mapping the cosmos and advancing our understanding of the universe's intricate structure.
Category: Artificial Intelligence
[4] viXra:2308.0075 [pdf] submitted on 2023-08-12 13:44:31
Authors: Xie Lei
Comments: 2 Pages.
Deep learning techniques have shown remarkable success in various tasks, including feature learning, representation learning, and data reconstruction. Autoencoders, a subset of neural networks, are particularly powerful in capturing data patterns and generating meaningful representations. This paper presents an investigation into the use of combination with Deep SVDD and memory modules.
Category: Artificial Intelligence
[3] viXra:2308.0062 [pdf] submitted on 2023-08-11 16:35:06
Authors: Satish Gajawada, Hassan Mustafa
Comments: 61 Pages.
Preface: In 20th and 21st Centuries the global optimization algorithms were created by taking inspiration from birds (Particle Swarm Optimization), ants (Ant Colony Optimization), chromosomes (Genetic Algorithms) etc. In "Twenty Second Century Artificial Intelligence" book global optimization algorithms are created by taking inspiration from Humans, Souls, Gods, Satisfied Beings, Mothers, Children, Particular Human Beings and Stories.In 20th and 21st Centuries research scientists focused mainly on Brain Inspired Computing. In "Twenty Second Century Artificial Intelligence" book a new path is shown where algorithms are created by taking inspiration from both heart and brain.In 20th and 21st Centuries the path of "Artificial Intelligence" is the main focus of research. In "Twenty Second Century Artificial Intelligence" book we defined "Artificial Satisfaction".In 20th and 21st Centuries researchers created many algorithms by taking inspiration from Nature (Nature Inspired Computing). In "Twenty Second Century Artificial Intelligence" book we created "Nature Plus Plus Inspired Computing".Abstract: The book defines various new paths as nine different chapters. First, second and third chapters deal with "Artificial Human Optimization", "Artificial Soul Optimization" and "Artificial God Optimization" respectively.Three new branches titled "Artificial Satisfaction", "Deep Loving" and "Nature Plus Plus Inspired Computing" are shown in fourth, fifth and sixth chapters respectively.The seventh chapter describes "Artificial Heart Neural Networks" where algorithms are created by taking inspiration from both Heart and Brain.Two new branches "Artificial Excellence" and "Stories Inspired Optimization Algorithms" are created in last two chapters of this book.
Category: Artificial Intelligence
[2] viXra:2308.0061 [pdf] submitted on 2023-08-11 16:41:42
Authors: Satish Gajawada
Comments: 2 Pages.
The primary purpose of writing this letter is to invent and define a new area called "Stories Inspired Optimization Algorithms (SIOA)".
Category: Artificial Intelligence
[1] viXra:2308.0048 [pdf] submitted on 2023-08-10 00:02:53
Authors: Vitaly Pilkin
Comments: 11 Pages.
To understand the degree of danger of AI for human civilization and the existence of humanity as a whole is possible only through understanding the Universe, the place of humans in the Universe and understanding the nature of thinking, consciousness and mentality.
Category: Artificial Intelligence
| Third-Party Links: |
|