The No Code AI and Machine Learning: Building Data Science Solutions Program lasts 12 weeks. The program will begin with blended learning elements, including recorded lectures by MIT Faculty, case studies, projects, quizzes, mentor learning sessions, and webinars.
These 12 weeks will be distributed in the following manner:
Module 1: Introduction to AI Landscape
Module 2: Data Exploration - Structured Data, Networks, and Graphical Models
Module 3: Prediction Methods - Regression
Module 4: Decision Systems
Module 5: Data Exploration - Unstructured Data
Module 6: Recommendation Systems
Module 7: Data Exploration - Temporal Data
Module 8: Prediction Methods - Deep Learning and Neural Networks
Module 9: Computer Vision Methods
Module 10: Workflows and Deployment
Week 1 - Module 1
Introduction to the AI Landscape
To offer a general overview of the four blocks upon which this No Code AI and Machine Learning Program is focused.
- Understanding the data: What is it telling us?
- Prediction: What is going to happen?
- Decision Making: What should we do?
- Causal Inference: Did it work?
Week 2 - Module 2
Data Exploration - Structured Data
To learn the basic principles of applying data exploration techniques, such as dimensionality projection and clustering on structured data.
- Asking the right questions to understand the data
- Understanding how data visualization makes data clearer
- Performing Exploratory Data Analysis using PCA
- Clustering the data through K-means & DBSCAN clustering
- Evaluating the quality of clusters obtained
Week 3 - Module 3
Prediction Methods - Regression
To understand the concept of linear regression and how it can be used with historical data to build models that can predict future outcomes.
- The idea of regression and predicting a continuous output
- How do you build a model that best fits your data?
- How do you quantify the degree of uncertainty?
- What do you do when you don't have enough data?
- What lies beyond linear regression?
Week 4 - Module 4
To understand the concept of classification and understand how tree-based models achieve prediction of outcomes that fall into two or more categories.
- Understand the Decision Tree model and the mechanics behind its predictions
- Learn to evaluate the performance of classification models
- Understand the concepts of Ensemble Learning and Bagging
- Learn how Random Forests aggregate the predictions of multiple Decision Trees
Week 5 - Learning Break
Week 6 - Module 5
Data Exploration - Unstructured Data
To understand the concept of Natural Language Processing and how natural language represents an example of unstructured data, the business applications for this kind of data analysis, and how data exploration and prediction are performed on natural language data.
- Understand the concept of unstructured data and how natural language is an example
- Understand the business applications of Natural Language Processing
- Learn the techniques and methods to analyze text data
- Apply the knowledge gained towards the business use case of sentiment analysis
Week 7 - Module 6
To understand the idea behind recommendation systems and potential business applications.
- Learn the concept of recommendation systems and potential business applications
- Understand the sparse data problem that necessitates recommendation systems
- Learn about potentially simple solutions to the recommendation system
- Understand the ideas behind Collaborative Filtering Recommendation Systems
Week 8 - Module 7
Data Exploration - Temporal Data
To understand the critical concept of temporal data, and its differences from structured and unstructured data, the idea behind Time Series Forecasting and the preprocessing required to obtain stationarity in Time Series.
- Understand temporal data and how it represents a different data modality
- Understand the idea behind Time Series Forecasting
- Learn about the concept of Stationary Time Series, testing for stationarity and conversion techniques to transform non-stationary time series into stationary
Week 9 - Learning Break
Week 10 - Module 8
Prediction Methods - Neural Networks
To understand the ideas behind Neural Networks, their introduction of non-linearities into the encoding and predictive process through a hierarchical structure, and the various steps involved in their forward propagation and back propagation cycle to minimize prediction error.
- Understand the key concepts involving Neural Networks
- Learn about the encoding process taking place in the neural network layers and how non-linearities are introduced
- Understand how forward propagation happens through the layered architecture of neural networks and how the first prediction is achieved
- Learn about the cost function used to evaluate the neural network's performance and how gradient descent is used in a back propagation cycle to minimize error
- Understand the critical optimization techniques used in gradient descent
Week 11 - Module 9
Computer Vision Methods
To understand how images represent a spatial form of unstructured data and hence, a different data modality, how the Convolutional Neural Network (CNN) structure achieves generalized encoding abilities from image data and acquire an understanding of what CNNs learn.
- Learn about spatial concepts of images such as locality and translation invariance
- Understand the working of filters and convolutions, and how they achieve feature extraction to generate encodings
- Learn about how these concepts are used in the structure of Convolutional Neural Networks (CNNs) and understand what CNNs actually learn from image data
Week 12 - Module 10
Workflows and Deployment
To obtain additional perspective on how the same takeaways from the conceptual modules discussed prior have been applied in various business scenarios and problem statements by industry leaders who have achieved success in practical applications of Data Science and AI.