# PRIMO.ai

On Saturday September 26, 2020 PRIMO.ai has 611 pages

## Contents

- 1 Getting Started
- 2 Information Analysis
- 3 Algorithms
- 3.1 Predict values - Regression
- 3.2 Classification ...predict categories
- 3.3 Recommendation
- 3.4 Clustering - Continuous - Dimensional Reduction
- 3.5 Convolutional
- 3.6 Graph
- 3.7 Sequence / Time
- 3.8 Competitive
- 3.9 Semi-Supervised
- 3.10 Natural Language
- 3.11 Reinforcement Learning (RL)
- 3.12 Neuro-Symbolic
- 3.13 Other

- 4 Techniques
- 5 Development & Implementation

# Getting Started

### Overview

### Background

### AI Breakthroughs

- Capabilities
- Case Studies
- Artificial Intelligence | United States Patent and Trademark Office --> AI Patents after 2013

### AI Fun

- Google AI Experiments
- TensorFlow Playground ...learn more
- TensorFlow.js Demos
- Google AIY Projects Program - Do-it-yourself artificial intelligence
- NVIDIA Playground

- Competitions
- Try GPT
- AI Dungeon 2 AI generated text adventure

*.. more Natural Language Processing (NLP) fun...*

- CoreNLP - see NLP parsing techniques by pasting your text | Stanford
- Sentiment Treebank Analysis Demo

### How to...

- AI Solver for determining possible algorithms for your needs
- Strategy & Tactics for developing AI investments
- AI Governance to reduce unnecessary risks and assure success
- Evaluation of AI investments
- Checklists for ensuring consistency and completeness

### Forward Thinking

# Information Analysis

- Framing Context
- Data Science
- Master Data Management (MDM) / Feature Store / Data Lineage / Data Catalog
- Managed Vocabularies
- Datasets
- Benchmarks
- Batch Norm(alization) & Standardization
- Data Preprocessing
- Data Encoding
- Data Cleaning
- Feature Exploration/Learning
- Data Interoperability
- Data Augmentation, Data Labeling, and Auto-Tagging
- Imbalanced Data
- Privacy in Data Science
- Bias and Variances
- Hyperparameters
- Zero Padding
- Train, Validate, and Test
- Evaluation - Measures
- Excel - Data Analysis
- Explainable Artificial Intelligence (XAI)
- Visualization

# Algorithms

## Predict values - Regression

- Linear Regression
- Ridge Regression
- Lasso Regression
- Elastic Net Regression
- Bayesian Linear Regression
- Bayesian Deep Learning (BDL)
- Logistic Regression (LR)
- Support Vector Regression (SVR)
- Ordinal Regression
- Poisson Regression
- Tree-based...
- General Regression Neural Network (GRNN)
- One-class Support Vector Machine (SVM)
- Gradient Boosting Machine (GBM)

## Classification ...predict categories

- Supervised
- Naive Bayes
- K-Nearest Neighbors (KNN)
- Perceptron (P) ...and Multi-layer Perceptron (MLP)
- Feed Forward Neural Network (FF or FFNN)
- Artificial Neural Network (ANN)
- Deep Learning - Deep Neural Network (DNN)
- Kernel Approximation - Kernel Trick
- Logistic Regression (LR)
- Softmax Regression; Multinominal Logistic Regression

- Tree-based...
- Apriori, Frequent Pattern (FP) Growth, Association Rules/Analysis
- Markov Model (Chain, Discrete Time, Continuous Time, Hidden)

- Unsupervised

## Recommendation

## Clustering - Continuous - Dimensional Reduction

- Singular Value Decomposition (SVD)
- Principal Component Analysis (PCA)
- K-Means
- Fuzzy C-Means (FCM)
- K-Modes
- Association Rule Learning
- Mean-Shift Clustering
- Density-Based Spatial Clustering of Applications with Noise (DBSCAN)
- Expectation–Maximization (EM) Clustering using Gaussian Mixture Models (GMM)
- Restricted Boltzmann Machine (RBM)
- Variational Autoencoder (VAE)
- Biclustering
- Multidimensional Scaling (MDS)

### Hierarchical

- Hierarchical Cluster Analysis (HCA)
- Hierarchical Clustering; Agglomerative (HAC) & Divisive (HDC)
- Hierarchical Temporal Memory (HTM) Time
- Mixture Models; Gaussian

## Convolutional

### Deconvolutional

## Graph

- includes social networks, sensor networks, the entire Internet, 3D Objects (Point Cloud)

- Graph Convolutional Network (GCN), Graph Neural Networks (Graph Nets), Geometric Deep Learning
- Point Cloud
- A hierarchical RNN-based model to predict scene graphs for images
- A multi-granularity reasoning framework for social relation recognition
- Neural Structured Learning (NSL)

## Sequence / Time

- Transformer
- Generative Pre-trained Transformer (GPT)
- Attention Mechanism/Transformer Model
- Transformer-XL

- Sequence to Sequence (Seq2Seq)
- End-to-End Speech
- Neural Turing Machine
- Recurrent Neural Network (RNN)
- (Tree) Recursive Neural (Tensor) Network (RNTN)

### Time

- Temporal Difference (TD) Learning
- Predict values

### Spatialtemporal

Spatial-Temporal Dynamic Network (STDN)

## Competitive

- Generative Adversarial Network (GAN)
- Image-to-Image Translation
- Conditional Adversarial Architecture (CAA)
- Kohonen Network (KN)/Self Organizing Maps (SOM)
- Quantum Generative Adversarial Learning (QuGAN - QGAN)

## Semi-Supervised

In many practical situations, the cost to label is quite high, since it requires skilled human experts to do that. So, in the absence of labels in the majority of the observations but present in few, semi-supervised algorithms are the best candidates for the model building. These methods exploit the idea that even though the group memberships of the unlabeled data are unknown, this data carries important information about the group parameters. Reference: Learning Techniques

- Semi-Supervised Learning with Generative Adversarial Network (SSL-GAN)
- Context-Conditional Generative Adversarial Network (CC-GAN)

## Natural Language

- Natural Language Processing (NLP) involves speech recognition, (speech) translation, understanding (semantic parsing) complete sentences, understanding synonyms of matching words, and sentiment analysis

## Reinforcement Learning (RL)

an algorithm receives a delayed reward in the next time step to evaluate its previous action. Therefore based on those decisions, the algorithm will train itself based on the success/error of output. In combination with Neural Networks it is capable of solving more complex tasks. Policy Gradient (PG) methods are a type of reinforcement learning techniques that rely upon optimizing parametrized policies with respect to the expected return (long-term cumulative reward) by gradient descent.

- Monte Carlo (MC) Method - Model Free Reinforcement Learning
- Markov Decision Process (MDP)
- State-Action-Reward-State-Action (SARSA)
- Q Learning
- Deep Reinforcement Learning (DRL) DeepRL
- Distributed Deep Reinforcement Learning (DDRL)
- Evolutionary Computation / Genetic Algorithms
- Actor Critic
- Hierarchical Reinforcement Learning (HRL)

## Neuro-Symbolic

the “connectionists” seek to construct artificial neural networks, inspired by biology, to learn about the world, while the “symbolists” seek to build intelligent machines by coding in logical rules and representations of the world. Neuro-Symbolic combines the fruits of group.

## Other

# Techniques

- Math for Intelligence
- Arxiv Sanity Preserver to accelerate research

### Methods & Concepts

- Backpropagation
- Overfitting Challenge
- Manifold Hypothesis and Dimensional Reduction; identification - what influences an observed outcome
- Activation Functions
- Memory Networks
- Multiclassifiers; Ensembles and Hybrids; Bagging, Boosting, and Stacking
- Optimizers
- Neural Network Pruning
- Repositories & Other Algorithms
- DAWNBench An End-to-End Deep Learning Benchmark and Competition
- Knowledge Graphs
- Quantization
- Train, Validate, and Test
- Causation vs. Correlation
- Image Retrieval / Object Detection; Faster Region-based Convolutional Neural Networks (R-CNN), You only Look Once (YOLO), Single Shot Detector(SSD)
- Deep Features
- Local Features
- Human in the Loop

### Learning Techniques

- Text Transfer Learning
- Image/Video Transfer Learning
- Few Shot Learning
- Transfer Learning a model trained on one task is re-purposed on a second related task
- Ensemble Learning
- Multi-Task Learning (MTL)
- Apprenticeship Learning - Inverse Reinforcement Learning (IRL)
- Imitation Learning
- Simulated Environment Learning
- Lifelong Learning - Catastrophic Forgetting Challenge
- Neural Structured Learning (NSL)
- Meta-Learning
- Human-in-the-Loop Learning

### Opportunities & Challenges

- Generative Modeling
- Inside Out - Curious Optimistic Reasoning
- Nature
- Connecting Brains
- Architectures
- Integrity Forensics
- Metaverse
- Other Challenges in Artificial Intelligence
- Quantum

# Development & Implementation

- Building Your Environment
- AIOps / MLOps
- Service Capabilities
- AI Marketplace & Toolkit/Model Interoperability

## No Coding

- Automated Machine Learning (AML) - AutoML
- Neural Architecture Search (NAS) Algorithm
- Other codeless options, Code Generators, Drag n' Drop

## Coding

### Libraries & Frameworks

#### TensorFlow

- TensorBoard
- TensorFlow Playground
- TensorFlow.js Demos
- TensorFlow.js
- TensorFlow Lite
- TensorFlow Serving
- Related...

### Tooling

- Model Search
- Model Monitoring
- Notebooks; Jupyter and R Markdown

### Platforms: Machine Learning as a Service (MLaaS)

### ... and other leading organizations

- Allen Institute for Artificial Intelligence, or AI2
- OpenAI
- NIST
- Stanford University, MIT, UC Berkeley, Carnegie Mellon University, Princeton University, University of Oxford, University of Texas Austin, UCLA, Duke University, EPFL, Harvard University, Cornell University, ETH, Tsinghua University, National University of Singapore, University of Pennsylvania, Technion, University of Washington, UC San Diego, University of Maryland, Peking University, Georgia Institute of Technology, University of Illinois at Urbana-Champaign, University of Wisconsin Madison, University of Toronto, Université de Montréal - Mila, KAIST, Texas A&M University, RIKEN, University of Cambridge, Columbia University, UMass Amherst, National Institute for Research in Digital Science and Technology (INRIA), New York University, University College London, University of Southern California, Yale University, Yandex, Shanghai Jiao Tong University, University of Minnesota, University of Chicago, McGill University, Seoul National University, University of Tuebingen, University of Alberta, Rice University, Johns Hopkins University

If you get a 502 or 503 error please try the webpage again, as your message is visiting the island which the server is located, perhaps deciding to relax in the Sun before returning. Thank you.