Home

Online Learning

Degrees

FAQ

Schedule of Courses

Contact Us

Texts

Software

Other Resources

Catalog Description

Course Objectives

Prerequisites

Class Location

Grading

Assignments

Additional Information

Schedule of Assignments and Events

Contact Information/Office Hours

Texts

Hands-On Machine Learning with Scikit-Learn & TensorFlow (most recent ed.) by Aurélien Géron, Ingram/O'Reilly (2017) ISBN 9781491962299 (Required).

Software

• The Anaconda Python (version 3) distribution, including the following Python packages/frameworks:
• Warning: Do _NOT_ "mess with" any existing Python installation on your computer. If you do, expect your operating system to stop working, and for the "fix" to be a real pain. You can install Anaconda without messing up the Python installation that came with your OS, and on which your OS is probably dependent.
• General software requirements are described on the Online Learning page.
• TensorFlow version 1.8 (non-GPU!).
• IMPORTANT NOTE: Install Anaconda (with Python 3.6), but then follow the TensorFlow instructions to install TensorFlow in a virtual environment. Do not install TensorFlow via Anaconda's conda package manager, even though you will be using Anaconda's Python. Instead, install TensorFlow using the TensorFlow website's instructions for installing into a virtual environment.

Other Resources

• "Learn TensorFlow and deep learning without a Ph.D" by by Martin Görner. Use the "3-hour" "college" version.

You will need a suitable computer. The University's Computer Recommendations are appropriate for this course for Windows and OS X. If you use the Linux operating system this course, the recommendations for Windows computers will also be suitable for Linux.

Please see the Online Learning page for more information regarding expected student student access to the internet, and corresponding software.

Please see the Mercer's Distance Learning page for more information regarding expected student student access to the internet, and corresponding software.

Course Description

Special Topics.

• This course will focus on deep learning using TensorFlow.
• The programming language of instruction will be Python. C.f. "a minimal crash course..." and/or the "official" python tutorial.
• We recommend using the Anaconda distribution of Python (even though we will likely not use the version of TensorFlow included with Anaconda). Be sure to get the Python version 3.6 of Anaconda.

This Online Learning course begins on May 21, 2018 at 12:00 am Eastern time. All assignments are due by the end (11:59:59 pm) of the specified calendar day in the Eastern time zone. This course ends at 11:59:59 pm on June 22, 2018, in the Eastern time zone.

Back to Top

Course Objectives

Breadth and Depth in Defining, Analyzing, and Solving Complex Problems:

• After successfully completing this course, you should be able to demonstrate a basic capability to apply deep learning techniques to solve complex problems.

Communicate Technical Aspects of the Solution for Complex Software Engineering Problems to a Technical Audience:

• After successfully completing this course, you should be able to demonstrate a basic ability to communicate regarding the application of deep learning techniques to a technical audience.

Back to Top

Prerequisite(s)

Prerequisite: Permission of the instructor.

You will need to be at least somewhat proficient in Python (3) programming when you start this course. Please contact the professor if you are not already proficient in Python programming.

Please see the Online Learning page for more information regarding expected student abilities and aptitudes.

Back to Top

Class Location

• This course is entirely online.
• Please see the Canvas page for information regarding access to Canvas, Mercer's online learning system. Please note that the Canvas contents for this course are in addition to this syllabus, and do not substitute for this syllabus.

Back to Top

Grading

Item	Percent of Term Grade
Direct Activity Report 1	3
Direct Activity Report 2	3
Direct Activity Report 3	4
Project 1	30
Project 2	30
Project 3	30

Credit will be given only for original work.

The Mercer University Honor Code, including its provisions for academic honesty (including plagiarism), applies to all Mercer students.

Project reports are to be submitted complete, not as incremental partial submissions.

All course work time is categorized as either Direct or Non-Direct (but not both). More information about the Direct and Non-Direct categories is contained in the Direct and Non-Direct and Report Logs. Please read these pages very carefully.

Instruction Time

Federal and regional accreditation requirements stipulate that a 3-credit hour course must include 2250 minutes of direct instruction time, which is approximately 68 minutes per day. For this course, that direct instruction time includes"

• Collaborative discussions with other students regarding the content of readings assigned for the course.
• Collaborative discussions with other students regarding the content that goes beyond readings assigned for the course, but is not part of work on the projects assigned on the syllabus.
• Collaborative development of generic technology that may be useful for your project(s), but is not, in its generic form, part of your project.

In addition to the 2250 minutes of direct instruction time each week, students are expected to spend a minimum of 4500 additional minutes (approximately 136 minutes per day) completing reading and writing assignments:

• Doing the assigned reading.
• Working on the projects assigned on the syllabus.

More information about the Direct and Non-Direct categories is contained in the Direct and Non-Direct and Report Logs web pages. Please read these pages very carefully.

All deliverables (Activity Reports and Project Reports) are due on their assigned dates.

• Deliverables submitted no more than one day late may be graded, but with a 10 point penalty .
• Deliverables submitted more than one day (24 hours) late will receive a grade of zero.
• The Non-Direct Activity reports are part of the project reports. The Non-Direct Activity reports do not receive a separate grade, but are used to help in evaluating your project reports.

Use the Report Logs to help you to manage your time and keep on schedule. You should schedule your work evenly throughout the term.

More information about project grading is contained in the General Project Rubric .

If you have a question about any of this, please ask.

Back to Top

Assignments

• Purpose of projects:
• Use the projects to develop your capabilities, and
• Use the projects to demonstrate your capabilities.
• Deliverables for all projects:

  All course work time is categorized as either Direct or Non-Direct (but not both). Read Direct and Non-Direct and Report Logs before continuing. Please read these pages very carefully.

  • All Direct Activity Reports include a Direct Activity Log, and may include additional supporting material at your discretion.

  • All Project Reports include a Non-Direct Activity Log, in addition to your work on the project.

  • Projects shall be delivered via the course FTP site.
  • Project reports should each be a single, standalone .pdf document, with the Bookmark feature used to support easy document navigation. Note that both MS Word and LibreOffice (which is free open-source software) support bookmarks, and will carry the bookmarks over into the pdf document when they convert your .docx, .odf, or whatever file type to PDF.
  • Project reports should focus on the development and/or maintenance that you actually do. Reports should not focus on summarizing or explaining the text or other materials.
  • Inclusion of screenshots demonstrating your work, and code you actually wrote for the project, are appropriate. Long code listings can be placed in appendices if doing so does not interfere with you presenting your work on that code in the body of the text.
  • Your report should be organized in such a way as to make the topics that you want credit for covering easy to find, and demonstrate your capabilities clearly and convincingly. Everything you want considered for credit, including code and tests (including test results), should be included in the report.
  • Each project can be a single, integrated project that tries to actually do something, or a collection of exercises that demonstrate your capabilities but don't accomplish anything else, or any combination of these two alternatives that you find convenient, unless otherwise specified in the assignment.
  • You may include other material outside of the report (on the FTP site), if you wish, but this material may or _may not_ be considered in evaluating your work.
  • For each capability that you demonstrate within a report or major section of that report, you may present only the final result of your work; you need not demonstrate every step in the development of that result.
  • Use IEEE standards to cite references How to Cite References: IEEE Documentation Style. Careful attention to this will help you to avoid plagiarism.
• Some suggestions for doing a project are contained in the One Way to Do a Project page.
• You may arrange the teamwork pretty much as you please. The team members may develop a single deliverable (code, report, etc.), or multiple deliverables, one per member. In this latter case, team members should perform substantive peer reviews of each others's work as that work progresses, to help their team mates to do better work. Persuasive evidence of these reviews should be included in the reports.
• Project 1 (solo or team) TensorFlow and Deep Neural Networks:
• Preliminaries: make sure that your Python skills are ready to go Catalog Description.
• Scope: watch the first 59 minutes of the (longer version of) the presentation by Martin Görner. Other Resources.
• Slide Deck Part 1: Slides 1 - 33.
• Scope: Read the preface and chapters 1, 9, 10, and 11 of the book by Aurélien Géron
• Scope: Codelab slides 1 - 10.
• Get the codelab that accompanies the Görner presentation, through slide 10, working in Jupyter. Demonstrate and report this work in your Project 1 report.
• Project 2 (team) Convolutional Neural Networks:
• Scope: watch the (longer version of) the presentation by Martin Görner from minute 59 to 1 hour 47 minutes. Other Resources.
• Slide Deck Part 1: Slides 34 -51, and Part 2: Slides 1 - 23.
• Scope: Read chapter 13 of the book by Aurélien Géron. (Chapter 12 is beyond the scope of this course.)
• Scope: Codelab slides 11 - 15.
• Define and perform your own CNN project. Demonstrate and report this work in your Project 2 report. You are most likely better off to base your work on the textbook by Géron. Try to avoid getting caught up in the complexities of getting the data into your program.
• Project 3 (team) Recurrent Neural Networks:
• Scope: watch the (longer version of) the presentation by Martin Görner from 1 hour 47 minutes to the ends Other Resources.
• Slide Deck Part 2: Slides 24 - 61.
• Scope: Read chapter 14 of the book by Aurélien Géron, excluding the material on distributed networks. (Chapter 12 is beyond the scope of this course.)
• Define and perform your own RNN project. Demonstrate and report this work in your Project 3 report. You are most likely better off to base your work on the textbook by Géron. Try to avoid getting caught up in the complexities of getting the data into your program.
• Please raise questions and issues via the course listserv.

Back to Top

Additional Information

• Asynchronous Learning Nets (ALNs)
• • Self Study: You study the texts and any other course study material on your own(team).
  • Collaborative Projects: After your self-study, you collaborate (via the Internet [email]) with your partner(s) to produce a team product. Your collaborative work is substantial, asynchronous, and rapid.
  • This course is an online learning course. Please read this Online Learning web page regarding online learning in this course and this program.

Back to Top

• What Do I Do?

• Subscribe to the course email listserv before the second day of class. Your professor will tell you how to do this.
• Study the assigned material.
• Write and read email messages about ideas, problems and solutions to do with the assigned study material.
• Do the assigned work, deliver the assigned deliverables.
• Write and read email messages (to/from the listserv) about ideas, problems, projects, and solutions to do with the assigned homework.
• Review the deliverables produced by other people.

Changes to this Syllabus:

There will be changes to this syllabus, so check back frequently, and don't forget to hit "Reload" or "Refresh".

Schedule of Assignments and Events (tentative)

Project 2 due

Start Date	End Date	Activity/Event
May 21, 2018	June 22, 2018	Course in Progress
	May 21, 2018	Subscribe to the course listserv.
May 21, 2018	May 31, 2018	Work on Project 1
	May 31, 2018	Project 1 (DNN) due Direct Activity Report 1 due
	June 11, 2018	Project 2 (CNN) due Direct Activity Report 2 due
	June 22, 2018	Direct Activity Report 3 due Project 3 (RNN) due Course ends

Notes: The course begins at midnight (12:00 am) on the first day of class (see schedule above), and ends at 11:59:59 pm on the last day of class. Assignments are due at any time during the day specified on the schedule, that is, before 11:59:59 pm on that day. All times are Macon, Georgia (Eastern time zone) times.

Back to Top

---

Contact Information

• My Name: Dr. Paul E. MacNeil
• My E-mail: macneil_pe@mercer.edu .
• OfficeHours
• My Office Phone: 478 301-2185
• US Mail:
 
• Dr. Paul E. MacNeil
• School of Engineering
• Mercer University
• 1501 Mercer University Drive
• Macon, GA 31207

Back to Top

Important Additional Information:

"Firearms/weapons are not allowed in Mercer University campus or Regional Academic Center (RAC) facilities except when required for students who are in uniform (police, law enforcement officers, etc.) in an official duty status. All other weapons are prohibited including, but not limited to, hunting weapons, BB guns, stun guns, air rifles, air pistols, paintball guns, bowie knives, daggers, dirks, slingshots, leaded canes, metal knife sharpeners, switchblade knives, blackjacks, metallic knuckles, and any sharp, pointed or edged instrument (except instructional supplies, unaltered nail files and clips, and tools used solely for preparation of food, instruction, and maintenance). All fireworks or explosives (defined as any substance prepared for the purpose of producing visible or audible effects by combustion, explosion, or detonation) are forbidden. https://police.mercer.edu/

“Students requiring accommodations for a disability should inform the instructor at the close of the first class meeting or as soon as possible. The instructor will refer you to the ACCESS and Accommodation Office to document your disability, determine eligibility for accommodations under the ADAAA/Section 504 and to request a Faculty Accommodation Form. Disability accommodations or status will not be indicated on academic transcripts. In order to receive accommodations in a class, students with sensory, learning, psychological, physical or medical disabilities must provide their instructor with a Faculty Accommodation Form to sign. Students must return the signed form to the ACCESS Coordinator. A new form must be requested each semester. Students with a history of a disability, perceived as having a disability or with a current disability who do not wish to use academic accommodations are also strongly encouraged to register with the ACCESS and Accommodation Office and request a Faculty Accommodation Form each semester. For further information, please contact Katie Johnson, Director and ADA/504 Coordinator, at 301-2778 or visit the ACCESS and Accommodation Office website at http://www.mercer.edu/disabilityservices."

---

Accreditation: Mercer University is accredited by SACSCOC.