A Few Things I Know About Her: A Personally Machine Learning Inspired Approach to Understand Surrounding Nature (Intelligent Systems Reference Library, 219)

Description

This book reconsiders key issues, such as description and explanation, which affect data analytics.  For starters: the soul does not exist. Once released from this cumbersome roommate, we are left with complex biological systems: namely, ourselves, who must configure their environment in terms of worlds that are compatible with what they sense. Far from supplying yet another cosmogony, the book provides the cultivated reader with computational tools for describing and understanding data arising from his surroundings, such as climate parameters or stock market trends, even the win/defeat story of his son football team. Besides the superposition of the very many universes considered by quantum mechanics, we aim to manage families of worlds that may have generated those data through the key feature of their compatibility. Starting from a sharp engineering of ourselves in term of pairs consisting of genome plus a neuron ensemble, we toss this feature in different cognitive frameworks within a span of exploitations ranging from probability distributions to the latest implementations of machine learning. From the perspective of human society as an ensemble of the above pairs, the book also provides scientific tools for analyzing the benefits and drawbacks of the modern paradigm of the world as a service. Acknowledgements Executive Summary Figure Credits Contents Symbols 1 Genome + Synapses 1.1 Genome 1.2 Synapses 1.3 The Power of Artificial Neural Networks 1.4 Further Readings 1.4.1 Biological Aspects 1.4.2 Modeling Aspects References 2 What I Can Understand, What I Can't 2.1 A Compatible World 2.1.1 The Compatibility Cue 2.1.2 Compatible Futures: The Cat Remains in Superposition 2.1.3 Indeterminacy and Confidence Intervals 2.2 My Own Rudimentary Physics 2.2.1 Cause Effect Dependencies 2.2.2 The Symmetry Classes 2.2.3 Much More 2.3 Tackling Unawareness Pebbles 2.3.1 The Description Thread 2.3.2 The Explanation Thread 2.4 Further Reading 2.4.1 The Probability Toolbox 2.4.2 Random Number Generators 2.4.3 Entropy 2.4.4 Conditional Probability References 3 Data First 3.1 An Essential Cosmogony of Data 3.1.1 Sampling Mechanism 3.1.2 Summing Up 3.2 The Inferential Problem 3.3 Compatible Functions 3.3.1 The Discrete Version: Classification 3.3.2 Learning Linear Functions 3.3.3 Steering by Sight 3.4 From Much Complex to Much Simple, Maybe: tertium non datur 3.5 Which Data? 3.6 Further Readings 3.6.1 The Chicken Egg Dilemma 3.6.2 Delving into the Roots 3.6.3 A Few Basic Statistics 3.6.4 A Few Suitable Kernels 3.6.5 Computational Complexity References 4 How to Tackle Difficult to Understand Phenomena. Fuzziness, Cognitivity, Memory 4.1 The Fuzziness Recovery 4.1.1 Exploiting Fuzzy Sets 4.2 The Communication Recovery 4.3 Cognitive Algorithms 4.4 Unawareness Management 4.4.1 The Information Path 4.4.2 The Goal Path 4.4.3 The Place of Cognitive Algorithm 4.4.4 Continuation Paths After Cognition 4.4.5 And What About Quantum Computing? 4.5 From Data to Model and Back: The Eternal Golden Loop 4.6 Memory 4.7 Computation, Communication and Memory: The Platinum Triangle 4.7.1 Give a Meaning to Things 4.7.2 Sharing the Names 4.7.3 Optimizing 4.7.4 Feedback 4.8 Further Reading 4.8.1 Fuzzy Sets 4.8.2 Communication 4.8.3 Cognitive Algorithms 4.8.4 Quantum Computing 4.8.5 Intelligent Memory 4.8.6 Large Numbers in Nature 4.8.7 Who Needs Names? References 5 Out-of-the-Box Solutions Versus Answers for Free 5.1 The Universal Responder Paradigm 5.1.1 All That Glitters Is Not Gold 5.1.2 And the Wise Owl Said, Let the Hare Eat the Fox 5.2 The Out of Box Solutions Thread 5.2.1 Understanding Is Really a Social Revolution 5.3 Bordering Own Knowledge Realm 5.4 No Thanks, You Must Repeat, Please! References 6 Conclusions References