705.603.81 - Creating AI-Enabled Systems

Course Structure

The course materials are divided into five Topics. Each Topic will be dedicated to study a specific system, culminating in a case study submission for the student to demonstrate an in-depth understanding of a specific ML system in its real-life context. This course heavily relies on a case study structure to explore complex issues, generate insights, and illustrate theories or concepts through detailed analysis. 

Each Topic's materials will be accessible at the beginning of their corresponding Module. You are encouraged to preview all sections of the topic as early as possible:

Module	Topic	Submissions
1	Introduction to ML Systems (Part 1): Requirements Engineering; Data Engineering; Dataset Design	Discussion 1
2	Introduction to ML Systems (Part 2): Requirements Engineering; Data Engineering; Dataset Design	Tool Assessment Assignment Discussion 2
3	Object Detection (Multi-Objective) Systems (Part 1): Problem Statement; Requirements Engineering; Data	Discussion 3
4	Object Detection (Multi-Objective) Systems (Part 2):  Model; Metrics; Deployment	Discussion 4
5	Object Detection (Multi-Objective) Systems (Part 3): Case Study Analysis	TechTrack Implementation TechTrack Design and Analysis Discussion 5
6	Vector Search Systems (Part 1): Problem Statement; Requirements Engineering; Data	Discussion 6
7	Vector Search Systems (Part 2): Model; Metrics; Deployment	Discussion 7
8	Vector Search Systems (Part 3): Case Study Analysis	IronClad Implementation IronClad Design and Analysis Discussion 8
9	Retrieval Augmented Generation Systems (Part 1): Problem Statement; Requirements Engineering; Data	Discussion 9
10	Retrieval Augmented Generation Systems (Part 2): Model; Metrics; Deployment	Discussion 10
11	Retrieval Augmented Generation Systems (Part 3): Case Study Analysis	TextWave Implementation TextWave Design and Analysis Discussion 11
12	Recommender Systems (Part 1): Problem Statement; Requirements Engineering; Data	Discussion 12
13	Recommender Systems (Part 2): Model; Metrics; Deployment	Discussion 13
14	Recommender Systems (Part 3): Case Study Analysis	MovieMate Design and Analysis SecureBank Final Case Study

Course Topics

• Fundamental Components of ML/AI Systems
  • System Requirements
  • Data
  • Model
  • Metrics
  • Deployment
• Perception (Computer Vision) Systems
  • Object Detection Models
  • Mean Average Precision
• Vector Search Systems
  • Embedding Models
  • Indexing and Search Strategies
  • Binary (Relevance) Ranking Metrics
• Retrieval Augmented Generation Systems
  • Text Generation
  • Reranking
  • Text Matching Metrics
• Recommender Systems
  • Recommender Models
  • Relevance Metrics
    
  • Cold-start scenarios

Course Goals

This course prepares students for creating AI-enabled systems by considering the full-lifecycle for building and designing practical AI solutions. We will study issues involving requirements, data, metrics, model, and deployment. This course demonstrates how this fulfills a deployable pipeline to construct working AI systems.

Course Learning Outcomes (CLOs)

• Utilize the end-to-end lifecycle of creating, deploying, and operating AI-enabled systems
• Employ strategies to decompose problem sets for applying AI capabilities to improve operations
• Perform data engineering tasks and design a data system to support an AI-enabled system
• Select AI algorithms to address the requirements to design an AI-enabled system
• Apply evaluation metrics to include addressing bias to test AI-enabled systems
• Explain the trade-space of different deployment strategies to include continuous monitoring techniques
• Create a portfolio of machine learning solutions available in github and dockerhub

Textbooks

No textbook is required. Each module will have a reading list that will include papers and textbooks chapters from O’Reilly online. Please use your JHED ID for logging into O’Reilly (https://www.oreilly.com). This is an excellent source for books and interactive tutorials. I encourage your exploration of such a valuable tool. If students are interested in learning more about the materials related to the basic foundations of this course, these books are recommended:

• Designing Machine Learning Systems by Chip Huyen
• Machine Learning Design Patterns by Valliappa Lakshmanan, Sara Robinson, Michael Munn

Required Software

Assignments require a python development environment to generate Jupyter notebooks and python scripts. A tutorial on Jupyter notebooks can be found here.

Docker (docker.com) is required to create the necessary development environment and more. The class provides the necessary instructions in Module 1. This allows the creation of a docker container that contains jupyter lab, an environment to create and run jupyter notebooks and python scripts. This container also includes most of the python libraries required for the assignments and git, a version control for your created software. A video tutorial can be found here. You will also use docker to create deployable images to run your work. There is an interactive docker tutorial available here.

Docker containers avoid library and operating system conflicts to one to also run the containers regardless of their environment (as long as it supports docker). With git and docker you will build an accessible and managed machine learning portfolio of your classwork. Docker also allows straightforward and isolated installation of other software discussed in the course such as NoSQL databases (e.g. MongoDB) and messaging software (e.g. Kafka). Docker images/containers are easy to remove when no longer needed. Docker containers are far superior than installing and deinstalling software and all the various python libraries directly in your computer operating system.

You are required to have access to an operating system that supports Docker. Current Windows, Mac, and Linux all support Docker. If this is not available, you can access cloud virtual machines and install docker there.

Student Coursework Requirements

It is expected that each module will take approximately 10-15 hours per week to complete. Here is an approximate breakdown:

• Reading the assigned sections of the texts, some outside reading (approximately 1–2 hours per week).
• Listening and watching to the audio annotated slide presentations (approximately 1 hour per week).
• Participating in discussions (approximately 1–2 hour per week).
• Completing assignments (approximately 7–10 hours per week). Note: The estimate depends on your prior Python coding ability. Those students with strong coding background should expect less time. 

All submissions are due according to the dates in the Calendar. Unless specified, all submissions should be submitted to GitHub and/or Canvas. You will be instructed as to the structure of your provided repository. This course will consist of the following basic student requirements:

Assignments (65% of Final Grade)

Assignments will be comprised of data analysis, software delivery, and a written explanation of the system design focused on the implementation your system. The assignments are meant to act as a way for students to receive early feedback for your case studies, and are encouraged to edit your system design and analysis based on the feedback your receive. Late submissions will be reduced by 10 points (out of 100) letter grade for each week late (no exceptions without prior coordination with the instructors). All deductions will not be prorated.

Final Project (25% of Final Grade)

The purpose of the final case study is to provide an in-depth understanding of a specific ML system. Each case study will explore systems to uncover insights about a problem and a proposed solution. These case studies begins with an open-ended description of a hypothetical problem, and students must propose an ML-system solution to address the presented problem. The general process to tackle case studies will include:

1. Deeply learn about the problem
2. Gather additional information
3. Brainstorm possible solutions
4. Evaluate alternatives
5. Develop the proposed solution
6. Analyze solution
7. Prepare the report

Case study submissions will be in the form of a Markdown file containing problem and system design descriptions. A good case study submission should include: 1) a description of the problem, 2) a description of your proposed design, 3) an argument for how your design addresses each requirement, and 4) test/evaluate hypotheses of your proposed design. This file should contain diagrams, tables, charts, etc. to argue for your design decisions. 

Discussion Participation (10% of Final Grade)

The discussions supplements the Final Project by allowing students to delve deeper into their chosen project through question prompts. The weekly discourse is composed of two parts (10 points total):

• Initial Posts (6 points): Students will post their initial response to the module's discussions prompt. Be detailed in your postings addressing these prompts. They may include figures, graphs, images . They are expected to post by the following Sunday of that module week. 
• Peer Responses (4 points):  Students will then choose at least two posts to submit a response by the following Friday (five days after the due date of the Initial Post). For responses, we want you to interact with your classmates, provide constructive feedback, and possibly influence the way your peers view problems and/or system. You are encouraged to agree or disagree with your classmates, and provide alternative solutions/approaches/insights. However, please ensure that your postings are civil and constructive. You will receive a severe penalty for inflammatory and disrespectful responses. An excellent thread will have some degree of closure for all concepts and ideas generated within the conversation.

I will monitor module discussions and will respond to some of the posts to provide feedback. You are encouraged to edit your responses based of the feedback you receive.  

Grading Policy: 

Assignments are due according to the dates posted in your Canvas course site. You may check these due dates in the Course Calendar or the Assignments in the corresponding modules.

Grades are typically posted one week after assignment due dates.

The course will use the following grading system:
100-90 = A,
89-80 = B
79-70 = C
69-63 = D
<63 = F

Final grades will be determined by the following weighting:

Submission	Weight
Assignments	65%
Final Project	25%
Discussions	10%

Grading Policy

Assignments are due according to the dates posted in your Canvas course site. You may check these due dates in the Course Calendar or the Assignments in the corresponding modules.
Grades are typically posted one week after assignment due dates.

The course will use the following grading system:
100-90 = A,
89-80 = B
79-70 = C
69-63 = D
<63 = F

Academic Policies

Deadlines for Adding, Dropping, and Withdrawing from Courses

Students may add a course up to one week after the start of the term for that particular course. Students may drop courses according to the drop deadlines outlined in the EP academic calendar. Between the 6th week of the class and prior to the final withdrawal deadline, a student may withdraw from a course with a W on their academic record. A record of the course will remain on the academic record with a W appearing in the grade column to indicate that the student registered and withdrew from the course. 

Academic Misconduct Policy

All students are required to read, know, and comply with the Johns Hopkins University Krieger School of Arts and Sciences (KSAS) / Whiting School of Engineering (WSE) Procedures for Handling Allegations of Misconduct by Full-Time and Part-Time Graduate Students. This policy prohibits academic misconduct, including but not limited to the following: cheating or facilitating cheating; plagiarism; reuse of assignments; unauthorized collaboration; alteration of graded assignments; and unfair competition. Course materials (old assignments, texts, or examinations, etc.) should not be shared unless authorized by the course instructor. Any questions related to this policy should be directed to EP’s academic integrity officer at ep-academic-integrity@jhu.edu.

Students with Disabilities - Accommodations and Accessibility

Johns Hopkins University values diversity and inclusion. We are committed to providing welcoming, equitable, and accessible educational experiences for all students. Our courses are designed with a proactive approach to accessibility to minimize the need for disability disclosure and accommodation requests, but we recognize that you may need additional support. Students with disabilities (including those with psychological conditions, medical conditions, and temporary disabilities) can request accommodations for this course by providing an Accommodation Letter issued by Student Disability Services (SDS). Please request accommodations for this course as early as possible to provide time for effective communication and arrangements.  For further information or to start the process of requesting accommodations, please contact EP Student Disability Services at ep-disability-svcs@jhu.edu. 

Student Conduct Code

The fundamental purpose of the JHU regulation of student conduct is to promote and to protect the health, safety, welfare, property, and rights of all members of the University community as well as to promote the orderly operation of the University and to safeguard its property and facilities. As members of the University community, students accept certain responsibilities which support the educational mission and create an environment in which all students are afforded the same opportunity to succeed academically. For a full description of the code please visit the Student Conduct Code website.

Classroom Climate

JHU is committed to creating a classroom environment that values the diversity of experiences and perspectives that all students bring. Everyone has the right to be treated with dignity and respect. Fostering an inclusive climate is important. Research and experience show that students who interact with peers who are different from themselves learn new things and experience tangible educational outcomes. At no time in this learning process should someone be singled out or treated unequally on the basis of any seen or unseen part of their identity. If you have concerns in this course about harassment, discrimination, or any unequal treatment, or if you seek accommodations or resources, please reach out to the course instructor directly. Reporting will never impact your course grade. You may also share concerns with your program chair, the Assistant Dean for Diversity and Inclusion, or the Office of Institutional Equity. In handling reports, people will protect your privacy as much as possible, but faculty and staff are required to officially report information for some cases (e.g. sexual harassment).

Course Auditing

When a student enrolls in an EP course with “audit” status, the student must reach an understanding with the instructor as to what is required to earn the “audit.” If the student does not meet those expectations, the instructor must notify the EP Registration Team (EP-Registration@exchange.johnshopkins.edu) in order for the student to be retroactively dropped or withdrawn from the course (depending on when the "audit" was requested and in accordance with EP registration deadlines). All lecture content will remain accessible to auditing students, but access to all other course material is left to the discretion of the instructor.