Add Thesis

How will Artificial Intelligence impact the labour market, which jobs will be replaced and what will it mean for society within the next decade?

Written by L. Adolfsson

Paper category

Master Thesis

Subject

Economics

Year

2020

Abstract

Master Thesis: What is artificial intelligence? To fully grasp artificial intelligence, we must first clarify what intelligence is. Artificial intelligence researcher Max Tegmark gave a broad definition in his book "Life 3.0": "Intelligence is the computational part of the ability to achieve complex goals. This includes logical ability, understanding ability, planning ability, emotional insight, self-awareness, Creativity, problem solving and learning" 4. Among many people known as the father of artificial intelligence, Alan Turing defines this subject as: "Artificial intelligence is the science and engineering of making intelligent machines, especially intelligent computer programs." 56 Artificial intelligence is the display of machines. Of intelligence. Today, the discipline has created weak artificial intelligence, which means that machines can fight or surpass human intelligence in specific tasks or professional fields, but cannot fight human intelligence in other fields. The goal of many researchers is to achieve general artificial intelligence, which means that machines can complete any task that humans can complete, or exceed human performance. 7 In the book Introduction to Machine Learning, the author, Professor of Computer Engineering at the University of Özkin and member of the Academy of Sciences Ethem Alpaydin wrote: "To be smart, a system in a constantly changing environment should have the ability to learn. If the system can learn And to adapt to such changes, system designers do not need to foresee and provide solutions for all possible situations.” 8 Machines (ML) with deep learning (DL) as the leading method in the past ten years are currently the most recognized technology in Artificial intelligence research. 1.1.1 Machine learning and deep learning Machine learning is a subset of artificial intelligence, a calculation method that uses experience to optimize performance standards or make accurate predictions. 9 This is equivalent to an algorithm that creates a mathematical model based on sample data. It was explicitly programmed to do this from the beginning. 10 "Learning" is actually a purely statistical mechanism in most applications, executing computer programs to optimize model parameters, using initial data or past experience. Professor Alpaydin wrote that the model can be predictive, predictive, descriptive to gain knowledge, or both. 8 Due to the fruitful results of deep learning, it has become the main method for many ongoing work in the field of machine learning. 8Deep learning is a statistical technique and can be reduced to a subset of machine learning. It is used to classify patterns from some sample data. Just like machine learning algorithms and functions are created, but instead of a "simple" layer, there are many neural networks (NN) with multiple layers. Each layer contains hidden units and neurons, which process the input and perform its specific purpose to produce the desired output. Artificial intelligence is replicating the reasoning and cognition of apes in many fields. DL has achieved various state-of-the-art results in the fields of speech and image recognition and language translation. For example, DL algorithms used by social media sites are becoming more and more adept at identifying objects, people, and more detailed attributes of images. 71314NLP, natural language processing, and knowledge representation caused a machine called Watson to defeat Jeopardy Champion in 2010. The same technology can be used for better web searchers. 15 Another example is a Stanford University study "Deep neural networks are more accurate than humans in detecting the sexual orientation of facial images", written by Professor Kosinski of Organizational Behavior and Computer Scientist Wang. They showed how deep neural networks can surpass humans' ability to recognize human sexual orientation from facial images. The accuracy rates of men and women are 81% and 74%, respectively. In contrast, the accuracy rate of humans is 61% of that of men. % And 54% forwomen.1416DL algorithm is the source of the main latest results in the 2012 ImageNet Large-scale Visual Recognition Challenge (ILSVRC). The source of success is Alex Krizhevsky, a convolutional neural network (CNN) called AlexNet created by a computer scientist, and co-published with colleagues Ilya Sutskever and Krizhevsky's doctoral supervisor Geoffrey Hinton. The project ImageNet is a large-scale image data set for the research of visual object recognition software. Compete in the ILSVRC software program to correctly classify and detect objects and scenes. 1417 In the ILSVR competition AlexNet, by using a high-performance model (GPU) that is computationally expensive but feasible by merging graphics processing units, it achieved a top-5 error of 15.3% (that is, an accuracy of 84.7%), which is a whole lower than the previous record. 10.8 percentage points. The groundbreaking result is said to have contributed to the prosperity of artificial intelligence. Reinforcement learning (RL) is the foundation of many modern artificial intelligences. It is inspired by biology. The core idea is to have a basis function, which represents the problem, and is updated over time. The problem to be solved in RL is: How does an intelligent agent learn to make a good decision sequence? The agent does not know or understand what the good outcome of the decision is, but must learn from the entire experience. The basic challenge of artificial intelligence and machine learning is to learn to make correct decisions under uncertain circumstances. RL involves optimization, delayed results, exploration, and generalization. The goal of optimization is to find the best decision-making method that produces the best results, or at least a very good strategy. In real life, decisions made now may affect later things, namely delayed consequences. An example from "real life" is that we understand that studying now (deciding) helps to succeed in the exam three weeks later (positive result) (delayed result). Read Less