| viXra.org > Artificial Intelligence > 2306 |
 |
 |
| viXra.org > Artificial Intelligence > 2306 |
|
|
[6] viXra:2306.0168 [pdf] submitted on 2023-06-30 16:21:18
Authors: Roman V. Yampolskiy
Comments: 30 Pages.
Artificially Intelligent (AI) systems have ushered in a transformative era across various domains, yet their inherent traits of unpredictability, unexplainability, and uncontrollability have given rise to concerns surrounding AI safety. This paper aims to demonstrate the infeasibility of accurately monitoring advanced AI systems to predict the emergence of certain capabilities prior to their manifestation. Through an analysis of the intricacies of AI systems, the boundaries of human comprehension, and the elusive nature of emergent behaviors, we argue for the impossibility of reliably foreseeing some capabilities. By investigating these impossibility results, we shed light on their potential implications for AI safety research and propose potential strategies to overcome these limitations.
Category: Artificial Intelligence
[5] viXra:2306.0099 [pdf] submitted on 2023-06-17 01:24:43
Authors: Sing Kuang Tan
Comments: 11 Pages.
In this paper, I am going to propose a new Boolean Structured Autoencoder Convolutional Deep Learning Network (BSautoconvnet) built on top of BSconvnet, based on the concept of monotone multi-layer Boolean algebra. I have shown that this network has achieved significant improvement in accuracy over an ordinary Relu Autoencoder Convolutional Deep Learning Network with much lesser number of parameters on the CIFAR10 dataset. The model is evaluated by visual inspection of the quality of the reconstructed images against groundtruth with reconstructed images by models in the internet.
Category: Artificial Intelligence
[4] viXra:2306.0055 [pdf] replaced on 2023-10-10 01:20:34
Authors: Shaun Stoltz
Comments: 10 Pages.
There have been significant improvements in directing large language models (LLM) to answer logic-based question such as mathematical reasoning tasks. This has resulted in near perfect performance on these types of problems with accuracy levels in the mid ninety percentile level using state of the art models (GPT-4). The achievement of this level of accuracy has previously needed a multi-prompt approach to elicit better performances from LLM’s. This paper introduces a new prompt paradigm termed "Mega prompt" and further introduces Proteus, a state of the art mega prompt, that has been used to achieve a new level of accuracy on the GSM8K math data set of 97%.
Category: Artificial Intelligence
[3] viXra:2306.0052 [pdf] submitted on 2023-06-10 12:16:23
Authors: Rodrigo F. Calhau, João Paulo A. Almeida, Giancarlo Guizzardi
Comments: 27 Pages. Preprint submitted to the International Journal on Software and Systems Modeling (SoSyM), Trends in Enterprise Architecture Management Research
Competence-based approaches have received increased attention, as the demand for qualified people with the right combination of competences establishes itself as a major factor of organizational performance. This paper examines how competences can be incorporated into Enterprise Architecture modeling: (i) we identify a key set of competence-related concepts such as skills, knowledge, and attitudes, (ii) analyze and relate them using a reference ontology (grounded on the Unified Foundational Ontology), and (iii) propose a representation strategy for modeling competences and their constituent elements leveraging the ArchiMate language, discussing how the proposed models can fit in enterprise competence-based practices. Our approach is intended to cover two tasks relevant to the combined application of Enterprise Architecture and Competence Modeling: `zooming in' on competences, revealing the relations between competences, knowledge, skills, attitudes and other personal characteristics that matter in organizational performance, and `zooming out' of competences, placing them in the wider context of other personal competences and overall organizational capabilities.
Category: Artificial Intelligence
[2] viXra:2306.0037 [pdf] submitted on 2023-06-09 01:04:04
Authors: Maksym Oleksandrovich Stavratii
Comments: 7 Pages.
Classification of electroencephalography (EEG) signals has important applications in the diagnosis and treatment of various neurological disorders. In this paper, we propose a methodology for classifying EEG signals based on signal processing using wavelet transform and superlet transform. The wavelet transform is used to decompose the EEG signal into frequency components, which are then used as features for classification. The proposed approach is evaluated using the publicly available "GAMEEMO" EEG dataset, which has been annotated by valence and emotional arousal. We use a Convolutional Neural Network (CNN) for classification at the waveform level. The results of this study suggest that the wavelet transform and its modifications, such as the superlet transform, can be valuable tools for analyzing and classifying EEG signals
Category: Artificial Intelligence
[1] viXra:2306.0003 [pdf] replaced on 2023-06-05 10:32:44
Authors: Essam El-Tobgi
Comments: 10 Pages.
Deep learning has become a powerful tool for solving a wide variety of problems, including those in physics. In this paper, we explore the use of deep learning for the detection of continuous gravitational waves. We propose two different approaches: one based on time-domain analysis and the other based on frequency-domain analysis. Both approaches achieve nearly the same performance, suggesting that deep learning is a promising technique for this task. The main purpose of this paper is to provide an overview of the potential of deep learning for physics problems. We do not provide a performance-measured solution, as this is beyond the scope of this paper. However, we believe that the results presented here are encouraging and suggest that deep learning is a valuable tool for physicists.
Category: Artificial Intelligence
| Third-Party Links: |
|