A Greater Foundation for Machine Learning Engineering: The Hallmarks of the Great Beyond in Pytorch, R, Tensorflow, and Python

Copyright © 2021 by Dr. Ganapathi Pulipaka.

All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system, without permission in writing from the copyright owner.

Any people depicted in stock imagery provided by Getty Images are models, and such images are being used for illustrative purposes only.

Certain stock imagery © Getty Images.

Rev. date: 08/23/2021

Xlibris

844-714-8691

www.Xlibris.com

823838

CONTENTS

Foreword by the Author

About the Author

Preface

Chapter 1 Introduction: A Greater Foundation for Machine Learning Engineering

Chapter 2 Supervised Learning

Chapter 3 Unsupervised Learning

Chapter 4 Origins of Deep Learning

Chapter 5 Linear Algebra

Chapter 6 Calculus

Chapter 7 Swarm Intelligence, Machine Learning Algorithms and In-Memory Computing

Chapter 8 Deep Learning Frameworks

Chapter 9 HPC with Deep Learning Frameworks

Chapter 10 History of Supercomputing

Chapter 11 Healthcare

Chapter 12 Real-World Research Projects with Supercomputers

Chapter 13 HPC with Parallel Computing

Chapter 14 Installation of PyTorch

Chapter 15 Introduction to reinforcement learning algorithms

Chapter 16 Reinforcement Learning – TRPO

Chapter 17 Reinforcement Learning – Cross-entropy Algorithm

Chapter 18 Reinforcement Learning – REINFORCE Algorithm

Chapter 19 The Gridworld: Dynamic Programming with PyTorch and Reinforcement Learning - Frozen Lake

References

GANAPATHI

PULIPAKA

FOREWORD BY THE AUTHOR

There have been significant developments from reinforcement learning in 2020 with accelerated commercialization of reinforcement learning algorithms from various industries. There has been a greater industry-academia collaboration with the implementation of papers, including autonomous vehicles for making complex decisions in dynamic environments in both discrete and continuous state spaces. A number of corporations have explored implementing reinforcement learning applications for edge computing applications for Industrial IoT and IoT. A number of research papers were released on both model-free reinforcement and model-based reinforcement learning algorithms. The significant advancements include MARL (Multi-agent reinforcement learning) framework COGMENT with internal interactions among the agents for leveraging the effective results from the observations and rewards in highly dynamic environments introducing a new technique Human-MARL learning, where a human individually cannot achieve the goal without an agent and agent cannot achieve the goal without the human. A new hybrid MARL reinforcement algorithm has been introduced dubbed D3-MADDPG (Dueling Double Deep Q Learning) for parallel training of decentralized policy rollouts with a joint centralized policy. Graph convolutional reinforcement learning algorithm has been introduced to learn to cooperate among humans and multiple agents in human-MARL environments. Deepmind has introduced behavior suite for reinforcement learning in 2020 with Python implementation. The Github code repository also shows some examples from OpenAI Baselines and Dopamine as reference implementations. Another reinforcement learning algorithm has been introduced in a randomized environment by Deepmind with randomized convolutional neural networks in 2D CoinRun, 3D DeepMind Lab, and 3D Robotics control tasks. Berkeley’s research in 2020 on deep reinforcement learning algorithms have found new adversarial policies can be reapplied from a particular adversary with reinforcement learning. An encoder-decoder neural network has been developed to search for the direct acyclic graph with reinforcement learning for best scoring. The model-free based reinforcement learning algorithms in the Atari gaming environment have been applied to learn effective policies. However, a new reinforcement learning algorithm has been introduced with a mode-free reinforcement learning environment dubbed SimPLe.

PyTorch vs. TensorFlow

PyTorch from Facebook was released in 2017, and TensorFlow was released in 2015 by Google. In 2020, the line blurred as both frameworks have seen a convergence in terms of popularity and functionality. The hardships for the machine learning engineers start to fade away with TensorFlow 2.0 as there was a major revamp on the programming API of TensorFlow with the inclusion of Keras into the main API. TensorFlow’s static computational graphs were great for wrapping the modules and the ability to run on a number of devices such as CPUs, GPUs, or TPUs. However, it’s always hard to debug a static computational graph. PyTorch always had a dynamic computational graph that allowed the data scientists to perform the computation line by line as the code gets interpreted, making it easy to debug and identify the problem. In 2020, TensorFlow introduced a dynamic computational graph similar to PyTorch with Eager mode. Now, PyTorch also allows static computational graph. In 2020, TensorFlow works similar to PyTorch in many ways, including distributed computing with the ability to run on single or multiple distributed GPUs or CPUs. OpenAI in 2020 has announced PyTorch as their official machine learning framework for 2020 and 2021 reinforcement learning and all the other deep learning projects. PyTorch has shown rapid uptake in data science and deep learning engineering community, as the fastest-growing open-source projects according to Github report. According to the analysis conducted by Gradient, in 2019, the platform grew 50% year-over-year with every major AI conference presented papers implemented in PyTorch. O’Reilly mentioned that PyTorch citations grew by more than 194% in the first quarter of 2019 alone. While Israel has shown a 54% increase in the interest towards PyTorch, people from Columbia have shown more interest in TensorFlow at 84%. Overall, PyTorch has shown a giant leap of growth on PapersWithCode with PyTorch implementations when compared with TensorFlow.

Natural language processing

Electra was introduced at ICLR 2020 for the cross-lingual ability of multilingual BERT with pre-training text encoders as discriminators rather than the generators leveraging commodity computing resources to pre-train the language models. StructBERT is another algorithm to incorporate language structures into pre-training for deep language understanding at the word and sentence levels to achieve SOTA results based on the GLUE NLP benchmark. Transformer-XL NLP algorithm was introduced, that goes beyond a fixed-length context learning the dependency that’s 80% longer than the recurrent neural networks and 450% longer than Vanilla transformers and 1800% faster than Vanilla transformers. BERT reached a GLUE score of 80.5% and MultiNLI accuracy of 86.7%. Google and Microsoft Research have developed neural approaches for conversational AI for NLP, NLU, NLG with machine intelligence. ALBERT, XLNet papers have shown advancements with an earlier generation of NLP papers. Microsoft has introduced Turing Natural Generation (T-NLG) language model with 17 billion parameters trained on NVIDIA DX hardware setup with Infiniband connections for communication between GPUs with NVIDIA V100 GPUs on NVIDIA Megatron-LM framework. DeepSpeed with Zero was introduced on T-NLG to reduce the model-parallelism degree, which is compatible with the PyTorch framework. Undoubtedly, GPT-3 has left its mark in 2020 on the trenches of natural language processing warfare in the entire history of mankind. GPT-3 was tuned with 175 billion parameters. It can create tweets and blogs. However, data scientists from LMU Munich, just in October 2020, have developed another advanced technique, PET (Pattern-Exploiting Technique), and trained the NLP models with just 223 million parameters that have outperformed GPT-3 on the GLUE benchmark. OpenAI has to rethink the architecture for GPT-4 with unlabeled samples, as PET implemented on a fine-tuned ALBERT transformer, that has achieved 76.8% compared to the earlier benchmark of 71.8% from GPT-3.

The Machine Learning Development trends for 2021

The digital computing for machine learning will shift to neuromorphic brain-like in-memory computing as the future of the machine learning paradigm. Manufacturing of large-scale neuromorphic spiking array processors to mimic the brain will be embraced by all the leading chip manufacturers, led by Hewlett Packard Enterprise, Samsung, IBM Corporation, Intel, Applied Brain Research Inc, General Vision Inc, and BrainChip Holdings Ltd. The neuromorphic computing market offering artificial neural networks, hardware, signal recognition, data mining in Aerospace, Defense, data science, Telcom, automotive, medical, and industrial regions is expected to grow at 86% CAGR in 2022 to USD $272.9 millions from USD $6.6 million in 2016 for supercomputing and high-performance computing applications.

In the 10,000 years of human civilization, only in the past 60 years, the speed of computation has gone from 1 floating-point operation per second to 250 trillion FLOPS, just in a blink of an eye. The first exascale supercomputer Aurora that can operate at floating-point operations per second will be launched in 2021 by US DoE to break the exascale barriers funded with $500 million. The exascale supercomputer is expected to shake up the human civilization in profound ways with enormous calculation speeds to solve long conundrums in medicine, bioinformatics, genomics, plasma turbulence, molecular interactions, quantum computing, and eventually be able to simulate the entire human brain regions by decoding the brain. We’ve seen a spur of material science advancements with the advent of memristors in 2020, shaking the foundations of in-memory computing. We can expect to see the rise of the memristors and graphene-based processors for in-memory computing with spiking neural network architectures for deep learning in 2021.

Memristive crossbar architectures will be the linchpin for the future of deep learning as powerful in-memory computing engines for artificial neural networks. The traditional memristors round off the nearest conductance states for the trained weights and cause slow-down on-chip training. In 2021 and beyond, we will see the introduction of graphene-based multi-level and non-volatile memristive synapses for programming endurance for greater computing accuracy while implementing machine learning and deep learning algorithms. The world will move away from the traditional Von Neumann architecture that operates between the logic and memory as a separate framework for scaling millions of synaptic weights of artificial neural networks. For the survival of computers and human advancement in the Post-Moore era, there will be a rise of Spintronics memory with magnetic random-access memory (MRAM), SOT-MRAM, VCMA-MRAM as an alternative to DRAM. In-memory computing and neuromorphic computing is the only way for breaking any exascale barriers for the future of deep learning in 2021 and beyond.

ABOUT THE AUTHOR

Dr. Ganapathi Pulipaka works as a Chief Data Scientist, Chief AI HPC Scientist, and SAP technical lead for DeepSingularity with more than 22 years of experience in the field. He is also a Postdoc research scholar in Computer Science Engineering with hands-on expertise in big data analytics, machine learning, deep learning, robotics, drones, Internet of Things (IoT), statistics, mathematics, and artificial intelligence as part of the Doctor of Computer Science program from Colorado Technical University, Colorado, with another PhD in data analytics, information systems, and enterprise resource management from California University, Irvine. He is also bestselling author of two books, The Future of Data Science and Parallel Computing: A Road to Technological Singularity, published on June 29, 2018, and Big Data Appliances for In-Memory Computing: A Real-World Research Guide for Corporations to Tame and Wrangle Their Data, published December 8, 2015.

He also published a number of eBooks, including The Digital Evolution of Supply Chain Management with SAP Leonardo, published in November 2017 and sponsored by SAP Leonardo with deep learning and machine learning algorithms for IoT and edge computing, Change HealthCare (McKesson’s HealthCare Corporation), published in December 2017, on machine learning and artificial intelligence for enterprise healthcare and health technology solutions.

He is also a public keynote speaker, industry leader, and AI expert for many data science events, including a few critical events such as the Robotics and Artificial Intelligence Conference held on May 21–22, 2018, Los Angeles, USA, ODSC West, SFO, VentureBeat Largest AI Conference at SFO, Intercon AI Conference at Nevada, and IDEAS AI Conference at the LA Convention Center. He developed a number of machine learning and deep learning programs applying various algorithms and published articles with architecture and practical project implementations on Data-driven Investor on Medium.com and LinkedIn with hosted code on Github. He is also a recipient of the 50 Tech Leader Awards from Intercon AI Conference on Mathematics, Statistics, and Data Science.

Dr. Ganapathi Pulipaka has also been inducted into Marquis Who’s Who 1899. As in all Marquis Who’s Who biographical volumes, individuals profiled are selected on the basis of current reference value. Factors such as position, noteworthy accomplishments, visibility, and prominence in a field are all considered during the selection process for the 2020 Marquis Who’s Who America Edition that is distributed to 115,000 American libraries.

Following are his global rankings in the field of machine learning, deep learning, artificial intelligence, and SAP.

• Ranked as Top Machine Learning Influencers for 2020 by CIO Views

• Inducted into the prestigious Marquis’s America’s Who’s Who Biographical Edition for 2020

• Ranked #1 IIoT Influencer for 2020 by Onalytica

• Ranked Top 10 Machine Learning Executive Influencers for 2019 by Analytics Insight Magazine

• Ranked #1 IoT Influencer by 10-Fold, a San Francisco company for 2019

• Ranked Top #36 Machine Learning Experts for 2020 by Acuvate

• Ranked Top #2 Big Data Influencer for Q1 2020 by Verdict UK Magazine

• Ranked Top #1 Machine Learning for Q1 2020 by Verdict UK Magazine

• Ranked #5 Data Science Influencer for 2018 by Onalytica and Joe Fields

• Ranked # 4 Machine Learning Influencer for January 2018 by KCore Analytics and Hernan Makse

• Ranked #3 Deep Learning Influencer for January 2018 by KCore Analytics and Hernan Makse

• Ranked #4 Machine Learning Influencer for March 2018 by KCore Analytics and Hernan Makse

• Ranked #3 Deep Learning Influencer for March 2018 by KCore Analytics and Hernan Makse

• Ranked #3 Data Science Influencer for 2017 by KCore Analytics and Hernan Makse

• Ranked #3 Machine Learning Influencer for 2017 by KCore Analytics and Hernan Makse

• Ranked #12 Business Intelligence Influencer for 2018 by Onalytica and Joe Fields

• Ranked #5 Machine Learning, #1 Analytics, #3 Data Science, #3 Big Data Influencer for 2018 by Agilience Authority Index

• Top #10 SAP and AI Solution Providers for 2018 published by Mirror Review Magazine

• Top #10 SAP and AI Solution Providers for 2018 published by Insights Success Magazine

• Recognized as part of the top list of prominent machine learning, deep learning, AI researchers, and influencers to follow outside Twitter and on Twitter by Mirror Review Magazine

• Top #20 CXO Leaders and SAP Innovative Solution Providers for 2017 published in SAP Special Annual Edition CIO Review

• Featured as Top 22 Artificial Intelligence Experts predicting the impact of AI in the enterprise workplace by Microsoft’s Partner Acuvate

• A Data Science Guide and Predictions for the Future with Onalytica and Joe Fields (Onalytica’s Interview – June 12, 2018)

• Data Science, Machine Learning: Main Developments in 2017 and Key Trends in 2018 (2018 Predictions from GP Pulipaka Published by KDNuggets)

• Recognized as top 8 AI influencers to follow on Twitter by Analytics India Magazine for 2018

• Recognized as one of the top AI influencers for 2019 by Verdict Magazine

• Recognized as top AI, machine learning influencers for Q1 2020 by Verdict UK Magazine

A chief data scientist, head of data science, SAP technology leader in artificial intelligence, SAP development, and solution architecture for DeepSingularity LLC. A project/program manager for application development of SAP systems integrating data science platforms, machine learning, deep learning systems, application development management, basis, infrastructure, and consulting delivery services offering expertise in delivery execution and executive interaction. Experienced implementing ASAP, Agile, Agile ASAP 8, HANA ASAP 8, Activate, Prince2, SCRUM, and Waterfall SDLC project methodologies. Coming with a background of dealing with petabyte-scale data warehouse environments in SAP, implemented multiple SAP programs/projects managing a team size of 60+ members, managing budget more than $5M to $10M with SAP backend databases of Oracle, IBM DB2, Sybase, Informix, MS SQL server on macOS and Linux environments. He worked on 32 projects for Fortune 100 corporations.

His background is in computer science with a professional skillset and two decades of management and hands-on development experience in machine learning in TensorFlow, Python, and R, Deep Learning in TensorFlow, Python, PyTorch, and R, SAP CPI-DS (SAP Cloud platform integration Technical Lead) IBP 1811, SAP ABAP S/4 HANA 1609, SAP ABAP S/4 HANA 1710, Big Data, IaaS, IoT, Data Science, Apache Hadoop, Apache Kafka, Apache Spark, Apache Storm, Apache Flink, SQL, NoSQL, Tableau, PowerBI, mathematics, data mining, statistical framework, SIEM, SAP, SAP ERP/ECC 6.0 NetWeaver Portals, SAP PLM, cProjects, R/3, BW, SRM 5.0, CRM 7.4, 7.3, 7.2, 7.1, 7.0, Java, C, C++, VC++, SAP CRM-IPM, SAP CRM- Service management, SAP CRM-Banking, SAP PLM Web UI 7.47, xRPM, SCM 7.1 APO, DP, SNP, SNC, FSCM, FSCD, SCEM, EDI. CRM ABAP/OO, ABAP, CRM Web UI/BOL/GENIL/ABAP Objects, SAP Netweaver Gateway (OData), SAP Mobility, SAP Fiori, Information Security, CyberSecurity, Governance, Risk Controls, and Compliance, SAP Fiori HANA, ABAP Webdynpros, BSPs, EDI/ALE, CRM Middleware, CRM Workflow, JavaScript, SAP KW 7.3 SAP Content server, SAP TREX Server, SAP KPro, SAP PI (PO), SAP BPC, Script logics, Azure, SAP BPM, SAP UI5, SAP BRM, Unix, Linux, macOS, and always looking for patterns in data and performing extractions to provide new meanings and insights through algorithms and analytics.

You can get in touch with him on the following social media channels:

LinkedIn: https://www.linkedin.com/in/dr-ganapathi-pulipaka-56417a2/

Twitter: https://twitter.com/gp_pulipaka

Facebook: https://www.facebook.com/ganapathipulipaka/

Github: https://github.com/GPSingularity

Website 1: www.gppulipaka.org

Website 2: www.deepsingularity.io

PREFACE

No recent research study has attempted to create a unified artificial intelligence theory of everything with practical use of comprehensive machine learning and deep learning frameworks as a recommendation machinery and data science field guide to the enterprises to establish a center of excellence. Some reviews of deep learning frameworks have been conducted in isolation but lack the integrated insights across the platforms to make a recommendation. A bevy of deep learning machinery, hardware, programming languages, and tools are purpose-built to address particular problems in the enterprise. However, these applications are still in a nascent stage. The book increases the visibility of these frameworks in high-performance computing environments to the enterprises to avoid one-size-fits-all approach and addresses the particular problem statement of the enterprise. This book delves into the machine intelligence frameworks to break the data science open and recommend the ensemble of best practices and features available in the frameworks. The book’s audience could be C-suite, chief data scientists, senior data scientists, data scientists, data science enthusiasts, and students, or even a layman who would like to grasp the architectures of machine learning and deep learning frameworks in supercomputing environments.

Keywords: machine learning, deep learning, data science, data analytics, big data analytics, high-performance computing, supercomputing, algorithms, artificial intelligence, processors, python, c, c++, programming

NOTATION

1

INTRODUCTION: A GREATER FOUNDATION

FOR MACHINE LEARNING ENGINEERING

1.1 Who Should Read This Book?

This book intends to build a greater foundation to develop expertise for data science enthusiasts, data scientists, and advanced data scientists with an approach covering machine learning, deep learning, bio-inspired algorithms for advanced data scientists. This course lays a solid foundation in deep learning and helps you to advance your skills in the deep learning research field to become a deep learning expert. Though, the course expects some basic algebra background, it transitions a data scientist from the field of deep learning to an advanced deep learning engineer or advanced data scientist.

This book lays a greater foundation in machine learning and deep learning engineering and helps you to advance your skills in the deep learning research field as a reference book to become a deep learning expert. The book offers required basic algebra and calculus as it transitions you to a data scientist from the field of deep learning to an advanced deep learning engineer or advanced data scientist.

The purpose of building this book is for you to understand the fundamentals of the algorithms and implement the algorithms in machine learning, reinforcement learning leveraging PyTorch, R, TensorFlow, and Python as you get acquainted with the machine learning engineering literature that can help you a build a data science team in your organization for performing advanced research in artificial intelligence. It can also help individuals to gain more hands-on experience with PyTorch, R, TensorFlow, and Python, and secure a career in the field of deep learning.

1.2 The Road to Deep Learning

Artificial intelligence has been the research area that has seen a growing body of literature in the recent decades. The quest for artificial general intelligence and building intelligent systems, mechanization of thinking patterns, and automation of mathematical formulations into devices began many centuries ago. The road to intelligence began by collecting intelligence in the form of knowledge. In the year 239, Chinese scholars collected 20,000-word knowledge volumes. In the third century, emerging markets and trades between China and Europe dubbed Silk Route began to exchange knowledge volumes. In the year of 330, Ptolemy created a large-scale library with 400,000 to 700,000 scrolls in Alexandria, Egypt. In the year of 350, Speusippus rolled up and bundled all the knowledge into a single volume. In the fifth century, Aristotle introduced syllogistic logic, a formal deductive reasoning system. In the seventh century, an organized and well-catalogued library was built at Nineveh, Babylonia. In the year 1270, Assyrian scholars have built an encyclopedia putting all the knowledge and thoughts together. In the fourteenth century, Ming Dynasty scholars put together an encyclopedia that is a result of 11,095 books. This encyclopedia was republished in 2002 digitally (McCorduck, 2004).

In the fifteenth century, the printing press revolutionized the massive circulation of knowledge in Europe. The quest for mechanization has led to building modern mechanical clocks. This intelligent innovation has appeared on European walls dubbed mechanical automata. In 1642, for the first time in the history, Blaise Pascal invented the mechanical calculator dubbed Pascaline. Gottfried Wilhelm Leibniz invented an advanced calculator in 1673 dubbed Step Reckoner. This calculator could perform all four arithmetic calculations intelligently. In 1822, Charles Babbage proposed building the difference engine, but he did not complete building end-to-end solution for the difference engine that could theoretically tabulate the polynomial functions. Subsequently, Charles Babbage built an analytic engine and an improved version of the difference engine. In 1890, Herman Hollerith took US Census leveraging machine intelligence that can extract and encode the information on the punching cards. The concept of implementing intelligence in chess games began in 1914 when A. Torres y Quevedo built chess-playing electromechanical machines. In 1923, the term robot was introduced into London’s play, Rossum’s Universal Robots. In 1937, Alan Turing proposed the framework for a universal intelligent computing machine. In 1938, Konrad Zuse developed the Z1 computer in Berlin, Germany. In 1941, Turing built intelligent machines that can automatically decrypt German intelligence transmissions with large-scale computations performed by Turing machines at Bletchley Park, England. In 1943, McCulloch began to push the concept of machine intelligence and logical calculus in nervous activity. Rosenblueth, Wiener, and Bigelow coined the term cybernetics in their published paper, Behavior, Purpose, and Teleology. They expected the machines to have intelligence similar to humans. In 1944, Eckert and Mauchly built the electronic numerator, integrator, and computer and installed at University of Pennsylvania. In 1945, Alan Turing wrote the pioneering paper Intelligent Machinery, advocating the advancements in artificial intelligence, but this paper did not get published. In 1947, Norbert Wiener published a paper, Cybernetics, and built an electromechanical turtle. In 1949, Mark I, the first computer that can store the data, came online at Manchester University. Turing and his associates attempt to build programs that can play chess. In 1950, Turing finally published the paper, Computing Machinery and Intelligence, in which he describes the Turing test to test a robot for human intelligence and consciousness. Isaac Asimov proposed three rules of robots (McCorduck, 2004).

Problem Statement

Due to the absence of significant literature, this book aims to create a new body of knowledge for the enterprises to choose the right computing machinery for machine learning and deep learning implementations in their organizations at scale for high-performance computing environments. The traditional data analytics crunching the big data are unable to provide the insights with data-processing power. The rise of the artificial intelligence in the recent decade has shown the potential of tapping into deep learning and machine learning platforms with the right processes and optimized hardware. Organizations need the ability to process large-scale big data in heterogeneous distributed systems for data processing on data centers with GPU accelerators. Several deep learning frameworks offer a variety of features and ability to process imperative and declarative programming languages. However, there is no integrated research that offers the insights into the deep learning frameworks to crunch the big data at scale. There are limitations and advantages of each framework for the organizations to move to the next level, from classical statistics to the next-generation advanced machine intelligence and artificial intelligence algorithms with neural networks and spiking neural networks applying differential calculus on reinforcement learning algorithms gearing toward AGI (Artificial general intelligence).

Machine Intelligence Tests

Figure 1. Machine Intelligence Tests

In the above diagram, the Harvey balls represent the validity of the test. A complete circle indicates a comprehensive intelligence test. A half circle is a debatable intelligence test. An empty circle means it’s not a valid test for the particular area of cognition. The test with question mark symbol indicates an unknown outcome with insufficient specifications. The artificial intelligence evolution continued to show advancements in the last 50 years. However, the concept of intelligence remained a challenge. Though there is no universal definition for machine intelligence, Alan Turing’s imitation game test remained as a stellar test for machine intelligence. Since the test itself does not define the patterns of machine intelligence, a number of researchers attempted to tackle this problem differently. Turing devised the Turing test to differentiate between a computer and machine by observing a teletyped conversation. In this test, the human judge would not be able to distinguish the difference between the computer and machine. Researchers proposed additional Turing tests such as Total Turing test, which also should involve the physical interaction with the machine apart from the textual conversation. A Truly Total Turing test is aimed at interpreting the robotic system beyond semantic representations in the sociolinguist context with connectionist cognitive architecture. The Inverted Turing test is administered on a machine to differentiate between a machine and human to measure the intelligence. Additional machine intelligence tests such as Linguistic Complexity and Text Compression are intended to test the aspects of grammar and sentence construction. In a Linguistic Complexity test, the system determines the conversational ability of machine intelligence with vocabulary size, syntactic complexities, and sentence length. The Turing Ratio test is administered with gaming tasks in contrast with cognitive tests. In the recent times, a number of machines have the ability to play games such as Google’s DeepMind artificial intelligence with reinforcement learning that surpasses human cognitive abilities. More researchers have administered an ensemble of psychometric tests that combine the tests of cognitive abilities, linguistic abilities, and creativity in literature (Legg & Hutter, 2007).

Smith’s Test is constructed with a number of algorithms that test the machines through iteration cycles with a problem generator expecting an accurate response to each question through a verifiable polynomial time. The iterative time cycles build a cumulative score for the response from the machine. The machine can propose multiple solutions and responses to the same problem statement. The score builds cumulatively. However, most of the artificial intelligence researchers do not consider Smith’s test as the most comprehensive test. C-Test is built based upon the g-factor view of the intelligence. The test is based on the fundamental principle that the intelligence is a means to handle complexity evolving from dynamics of the