CS357: Foundations of Artificial Intelligence - Literature Review (100 Points)

Purpose, Task, and Criteria

Purpose: To test the plan your team has already committed to against the published evidence rather than against your intuitions, and to make your understanding of the problem multi-disciplinary by reading what the stakeholder's own field says about it, while there is still time for what you find to change the project.

Task: Individually annotate 2-3 credible sources, then as a team write a one-page synthesis that connects them to your problem statement, tests the gap your proposal claimed, and states plainly what the evidence confirms, complicates, or changes about the plan you proposed.

Criteria: I grade this on faithful summaries in your own words, specific credibility assessments, relevance tied to the problem statement, a required span of at least two disciplinary perspectives, a gap that follows from the sources, and an honest account of what the evidence changes about your proposed plan. The rubric below has the details.

Assignment Goals

The goals of this assignment are:

To research the team's stakeholder problem through published sources, assessing each source's credibility and relevance to the problem statement (Goal 11)
To build a multi-disciplinary understanding of the problem by drawing sources from at least two disciplinary perspectives, including the stakeholder's own discipline (Goal 12)
To synthesize individually gathered sources into a team-level account of what is known and the specific gap the project addresses, and to carry that account back into the project plan as a stated confirmation, revision, or risk
To practice information literacy, finding, reading, and critically evaluating scholarly and technical sources

Background Reading and References

Please refer to the following readings and examples offering templates to help get you started:

The Assignment

In this Project Thread milestone, your team finds out what the world already knows about your stakeholder’s problem, and whether that knowledge agrees with the plan you are proposing. The review runs in two phases. In Phase 1, each team member works alone and builds an annotated bibliography of 2-3 sources. In Phase 2, the team writes one page that connects everyone’s sources to the problem statement, names the gap your project will address, and says what the evidence changes about the plan. The assignment goes out before your Final Project Proposal is due, and the course schedule has the due date for each phase.

A literature review is an argument: here is what is known, here is who knows it, here is the hole in it, and here is why that hole matters to our stakeholder.

Phase	Who submits	What you submit	How it is graded
Phase 1: Individual Annotated Bibliography	Each member, individually, via the LMS	2-3 annotated sources, with how you found each one and a complete citation	A small completion rubric, so that you get feedback before the synthesis
Phase 2: Team Synthesis	One PDF per team	One-page synthesis, plan-impact section, combined bibliography, each member’s reflection, signatures, AI-use disclosure	The full rubric below, including your individual annotations

Phase 1 is due first, and it is the longer half. Phase 2 falls across the Thanksgiving break, so schedule the team meeting now.

On using AI here. You may use AI tools to find candidate sources or to explain a difficult passage. They are also the most reliable way to end up citing a paper that does not exist, so every source you cite must be one you opened, with a link or DOI (digital object identifier) that resolves. Write every annotation yourself from the source, check every quote and claim against the source, and say in your disclosure what you used and how you checked. A bibliography with a made-up source is an integrity violation, not a formatting error.

Phase 1: Individual Annotated Bibliography

An annotated bibliography is a list of sources where each entry carries a paragraph of your own: what the source says, whether to trust it, and why it matters here. Keep your Stakeholder Brief and your proposal open, because each annotation ties back to them. You are the primary author of your own annotations.

Step 1.1: Divide the territory with your team

If everyone searches the same keywords, your team ends up with three copies of the same paper. So in a standup, agree on who searches which angle, so that the team’s combined sources span at least two disciplines and at least one comes from the stakeholder’s discipline. Log the division.

You should have. A logged division, with each member’s name next to the angle they will search.

Step 1.2: Search and keep a log

Start at the library databases (through the Ursinus library site) and Google Scholar, because they surface work the open web does not. Then:

Search the stakeholder’s vocabulary and yours. The education literature says “formative feedback” where computer science says “evaluation loop.”
When you find one good source, chain it. Citation chaining means following its references backward to older work, and using Google Scholar’s “Cited by” to go forward to newer work. It is the fastest way to map a literature.
To reach the stakeholder’s discipline, look up which journals their field reads (your interview’s “what should we read?” question was for this), or ask a librarian.
Log your search terms as you go, because each annotation reports how you found its source.

You should have. A short list of candidates, each with a link or DOI and the search terms (or the source you chained from) that found it.

Step 1.3: Choose your 2-3 sources

At least one must be a scholarly source: a peer-reviewed article, conference paper, or academic book chapter, whose claims other experts checked before publication. Technical documentation (official docs, standards, agency or nongovernmental organization reports) is allowed. It is authoritative about how a thing works, but it is not evidence about whether it helps anyone, so it supplements the scholarly requirement rather than replacing it.

If it fails. If none of your picks is scholarly, chain from your best technical source into the research it cites or that cites it. If a source will not open, drop it.

Step 1.4: Read each source critically

You do not have to read a paper in order. Read the abstract, then the introduction’s last paragraph (the claimed contribution), then the figures and tables, then the conclusion’s limitations. Then decide whether the middle is worth your hour. As you read, answer three questions:

What is the actual evidence (sample, method, effect size)?
Who is not represented in it?
What do the authors themselves say it cannot show?

For technical documentation, ask instead: what version is this, is it current, and does it describe designed behavior or measured behavior?

Step 1.5: Write one annotation per source

Write 150-250 words per source, from the source itself rather than the abstract, with three labeled parts. Here is a worked example. The paper is constructed for this page, so do not cite it.

Citation: Rivera, A. and Chen, L. (2024). "Automated feedback in an
  introductory writing course." Journal of Example Studies, 12(3), 45-60.
How I found it: Google Scholar, "automated feedback" "first-year writing";
  chained from the survey our partner recommended.

Summary: Rivera and Chen compared two sections of one writing course over a
semester.  Students who received automated feedback on drafts revised more
often, but their final essays scored about the same as the other section's.
They conclude that the tool changed how students work and not what they
produce.

Credibility assessment: This is a peer-reviewed journal, and the authors
report no funding from the tool's vendor.  The study has a small sample, 62
students at one college, and the sections were not randomly assigned.  It can
support the claim that feedback changes revision habits in a setting like
ours.  It cannot support a claim about long-term learning, and it did not
measure any.

Relevance to our problem: Our problem statement says the writing center cannot
respond to drafts quickly enough.  This source supports the assumption that
faster feedback increases revision, and it complicates our success criterion,
because we planned to measure final essay scores, and those may not move.

Notice what the credibility assessment does. It goes past “peer-reviewed, so credible” to say what this source can and cannot support, based on the venue, method, incentives, and recency. The relevance part names an element of the problem statement, and it says what changes.

Step 1.6: Fill in the skeleton and submit

Use this skeleton for each source, fill every line, and submit individually via the LMS. Run the Individual annotated bibliography items in the Self-Check first.

Citation (complete, consistently formatted):
How I found it (database and search terms, or chained from):
Summary:
Credibility assessment:
Relevance to our problem statement:

Phase 2: Team Synthesis

A synthesis connects sources to each other and to your problem, instead of summarizing them one at a time. The synthesis is one page, with a named primary author who holds drafting responsibility. The whole team reviews and signs.

Step 2.1: Outline by idea, not by source

Meet after every Phase 1 bibliography is in, and name the primary author. Then outline the page by idea, with each heading tied to your problem statement. Here is the difference, in two constructed sentences:

Weak:    Rivera and Chen studied automated feedback.  Okafor studied writing centers.
Strong:  Two of our sources agree that fast feedback increases revision, but neither
         finds better final essays, and that is the outcome our partner cares about.

The strong sentence makes a claim that the sources support together. At least once, show where your disciplines converge, conflict, or measure success differently, and what that means for the project.

Watch out. The team’s combined review must span at least two disciplinary perspectives (for example, computer science and public health), including the stakeholder’s discipline. A review whose every source is a computer science paper fails this requirement, regardless of quality.

You should have. An outline whose headings are ideas, each listing the sources that support or complicate it.

Step 2.2: Name the gap, and your weakest evidence

The gap is the specific missing thing your project will address: a question no source answers, a population no study covers, or a tool no one has built for this context. Show that it follows from the sources: what exists, what is missing, and why the missing piece matters to your stakeholder. “More work is needed” is not a gap. Compare it with the gap your proposal claimed, which may turn out to be the same, narrower, or different.

Then say honestly where your sources are weakest, such as a question only one source addresses, or a perspective missing from the set.

Step 2.3: State what this changes about the plan

This is where the reading reaches the project. Under its own heading, write one entry per proposal element. If the sources leave the plan standing, say so and say why. That is a finding, not a formality.

Proposal element:
Confirmed, complicated, or changed:
Source(s) responsible:
What we are doing about it (or why the plan stands):

Record any resulting scope change in your decision log, and raise it at your next intra-team check-in.

Step 2.4: Review, sign, and assemble the team PDF

Every member reviews and signs. The PDF holds the one-page synthesis with the primary author named, the what this changes about the plan section, the combined bibliography, each member’s reflection below, all members’ typed signatures, and the team AI-use disclosure. Run the Team synthesis items in the Self-Check first.

Deliverables

File or artifact	What it shows	Rubric row
Phase 1: your own annotated bibliography, submitted individually via the LMS	You read, judged, and connected real sources	Individual Annotations; Disciplinary Breadth; Information Literacy
Phase 2: one team PDF containing the one-page synthesis, primary author named	The sources talk to each other and the gap follows from them	Team Synthesis; Disciplinary Breadth
In the team PDF: the what this changes about the plan section	The evidence was carried back into the proposal	Team Synthesis
In the team PDF: the combined bibliography, all members’ typed signatures, and the team AI-use disclosure	Complete citations and the Project Thread process elements	Information Literacy
In the team PDF: each member’s answers to the Reflection Prompts below	What the reading changed for you	Part of a complete submission

Reflection Prompts

Each member answers both, under their own name, in the Phase 2 PDF, keyed to the Open Questions (Goal 15):

How can we understand the world? Pick the source from another discipline that most changed how you see the problem. What does that discipline accept as evidence that surprised you, and has your idea of what “knowing something” means shifted at all?
What should matter to me? The gap your team named is a claim about what deserves work. Whose needs does that gap center, and whose did your sources let you overlook?

Self-Check Before You Submit

Individual annotated bibliography

Every source is one I actually opened, and every link or DOI resolves.
Each annotation says what the source claims, how it knows, and what it means for our problem, rather than summarizing an abstract.
Each annotation includes a specific credibility assessment (venue, method, funding or incentive, recency, and what the source can and cannot support) and ties the source to a named element of our problem statement.
At least one of my sources is scholarly, and each annotation notes how I found the source (database, search terms, or citation chain) and gives a complete citation.
Across the team’s combined set, the sources span at least two disciplinary perspectives, and at least one comes from the stakeholder’s discipline.

Team synthesis

The synthesis is organized by theme or argument, not source by source.
Where sources disagree, the disagreement is stated and not smoothed over, and at least one observation says where the disciplines converge, conflict, or define success differently, and what that means for the project.
The gap our project addresses is named, and it follows from the sources rather than being asserted before them.
We have said what the evidence confirms, complicates, or changes about our proposal, naming the element affected and the source responsible.
If nothing about the plan changed, we have said why the sources leave it standing rather than leaving the question unanswered.
Every claim traces to a citation, and every citation traces to a source someone on the team read.
We noted honestly where our sources are weakest.
AI disclosure names what was AI-assisted and how citations were verified.
The synthesis names its primary author, the combined bibliography is consistently formatted, and the PDF carries all members’ typed signatures.

Submission

In your submission, please include answers to any questions asked on the assignment page, as well as the questions listed below, in your README file.

If you wrote code as part of this assignment, please describe your design, approach, and implementation in a separate document prepared using a word processor or typesetting program such as LaTeX. This document should include specific instructions on how to build and run your code, and a description of each code module or function that you created suitable for re-use by a colleague.

In your README, please include answers to the following questions:

Describe what you did, how you did it, what challenges you encountered, and how you solved them.
Please answer any questions found throughout the narrative of this assignment.
If collaboration with a buddy was permitted, did you work with a buddy on this assignment? If so, who? If not, do you certify that this submission represents your own original work?
Please identify any and all portions of your submission that were not originally written by you (for example, code originally written by your buddy, or anything taken or adapted from a non-classroom resource). It is always OK to use your textbook and instructor notes; however, you are certifying that any portions not designated as coming from an outside person or source are your own original work.
Approximately how many hours it took you to finish this assignment (I will not judge you for this at all...I am simply using it to gauge if the assignments are too easy or hard)?
Your overall impression of the assignment. Did you love it, hate it, or were you neutral? One word answers are fine, but if you have any suggestions for the future let me know.
Using the grading specifications on this page, discuss briefly the grade you would give yourself and why. Discuss each item in the grading specification.
Any other concerns that you have. For instance, if you have a bug that you were unable to solve but you made progress, write that here. The more you articulate the problem the more partial credit you will receive (it is fine to leave this blank).
Please describe any use of outside resources you may have engaged in the completion of this assignment, including the use of generative Artificial Intelligence.

Assignment Rubric

Description	Pre-Emerging (< 50%)	Beginning (50%)	Progressing (85%)	Proficient (100%)
Individual Annotations, Summary, Credibility, and Relevance (Goal 11) (40%)	Annotations are missing, or are copied abstracts rather than the student's own account	Two to three sources are listed but annotations summarize only, with no credibility assessment or no connection to the team's problem statement, or the required source types (at least one scholarly; at least one from the stakeholder's discipline across the set) are not met	Each annotation contains a summary, a credibility assessment, and a relevance statement, but at least one element is thin, a credibility assessment that says "peer-reviewed, so credible" or a relevance claim that does not mention the stakeholder's problem	Each of the student's 2-3 annotations contains a faithful summary in the student's own words, a specific credibility assessment (venue, method, funding or incentive, recency, and what the source can and cannot support), and a relevance statement that ties the source to a named element of the team's problem statement; at least one source is scholarly and technical documentation is used appropriately where included (Goal 11)
Disciplinary Breadth (Goal 12) (20%)	All sources are from a single perspective, or the stakeholder's discipline is absent from the team's combined review	A second discipline appears in the source list but the annotations do not engage with how that discipline frames the problem	The combined review spans two disciplinary perspectives including the stakeholder's discipline, and annotations engage each field's framing, but the perspectives never meet, no annotation or synthesis paragraph notes where the fields agree, disagree, or measure differently	The team's combined review spans at least two disciplinary perspectives, at least one source comes from the stakeholder's discipline, and the writing shows the perspectives interacting, at least one explicit observation of where the disciplines converge, conflict, or define success differently, and what that means for the project (Goal 12)
Team Synthesis and Gap Identification (Goals 11, 12) (25%)	No synthesis is submitted, or it is a list of the annotations restated	The synthesis connects some sources to each other but not to the problem statement, or names no gap	The one-page synthesis connects the sources to the problem statement and proposes a gap, but the gap is generic ("more work is needed") or is not actually supported by the reviewed sources, or the account of what the evidence changes about the proposed plan is present but perfunctory	The one-page synthesis, with a named primary author, weaves the team's sources into a coherent account of what is known about the stakeholder's problem, states the specific gap the project addresses, shows that the gap follows from the reviewed sources (what exists, what is missing, why it matters to the stakeholder), states explicitly what the evidence confirms, complicates, or changes about the plan the team proposed (naming the affected element of the proposal, and where nothing changed, saying why the sources leave the plan standing), and honestly notes where the team's sources are weakest (Goals 11, 12)
Information Literacy, Process, and Submission (15%)	An incomplete submission is provided	Citations are incomplete or unformatted, search strategies are undocumented, or the Project Thread signatures are missing	Citations are complete and consistent and the submission includes all process elements, with a minor gap such as an undocumented search strategy for one source	Every source has a complete, consistently formatted citation; each annotation notes how the source was found (database, search terms, or citation chain); the submission carries all members' signatures and names the synthesis's primary author; every citation was verified against the actual source

Please refer to the Style Guide for code quality examples and guidelines.