01Understanding Academic Research Papers
Sections 1–12 are general academic writing guidance that applies to most computer science and engineering venues. Section 13 explains, separately, what is known about the historical CSE 2022 submission process. CSE 2022 closed in 2022; this website has no submission system.
A research paper is an argument supported by evidence: it states a problem, claims a contribution, and gives readers enough information to check whether the claim is true.
Papers differ by type. An original research paper presents new methods or results; a survey organizes existing work; a position paper argues for a direction; a reproducibility study re-examines a published result. Know which you are writing, because reviewers judge each type differently.
02Research Question Development
A good research question is specific, answerable with the resources available, and significant to someone beyond the author. Compare:
| Weak | Stronger |
|---|---|
| Can machine learning help engineering? | Can a learned surrogate reproduce the peak stress predicted by a finite element model of a given bracket within 2%, at a fraction of the run time? |
| Is our algorithm fast? | How does the algorithm's running time scale with graph size on sparse road networks, compared with Dijkstra's algorithm with a binary heap? |
Write the question down before collecting results. Questions formed after seeing data invite unconscious bias.
03Literature Review
A literature review establishes what is already known, identifies the gap your work addresses, and shows that comparisons are fair. Search systematically across library databases and key venues, follow citations forward and backward from central papers, and record search terms so the review can be repeated. Organize related work by theme or approach rather than paper by paper, and describe others' work accurately and generously.
04Research Methodology
Computational research typically uses one or more of these methods:
- Theoretical: proofs of correctness, complexity bounds or convergence.
- Experimental: controlled computational experiments comparing methods.
- Simulation-based: studying a system through a validated model (see computational science).
- Empirical/observational: measuring real systems or datasets.
- Design science: building and evaluating an artefact that solves a practical problem.
Justify why the chosen method can answer the research question, and state its assumptions.
05Experimental Design
- Baselines: compare against strong, current methods, tuned with the same effort as your own.
- Ablations: remove one component at a time to show which parts matter.
- Controls: change one factor while holding others fixed.
- Repetition: run multiple trials and report variation, such as standard deviation or confidence intervals.
- Pre-specified metrics: choose evaluation measures before seeing results.
06Data and Reproducibility
Terminology varies, but a common distinction is that reproducibility means obtaining the same results using the original authors' data and code, while replicability means obtaining consistent results in a new study with new data. Both depend on clear reporting.
When data cannot be shared for privacy or licensing reasons, explain why and describe them in enough detail for others to obtain comparable data.
07Results Presentation
Present results so a reader can verify the claims quickly. Each figure and table should be understandable on its own, with labelled axes, units and a caption stating what it shows. Report uncertainty. Avoid selecting only the best runs or datasets. Distinguish statistically meaningful differences from noise, and resist describing small improvements as breakthroughs.
08Discussion and Limitations
The discussion interprets results: what they mean, why they may have occurred, and how they relate to earlier work. A candid limitations section strengthens a paper rather than weakening it. State where the method was not tested, which assumptions might not hold, and what could invalidate the conclusions.
09Citations and References
Cite sources for every claim that is not your own contribution or common knowledge. Use the citation style required by the venue consistently, cite the original source rather than a secondary mention, and read what you cite. Reference managers help keep bibliographies accurate. Never cite a work you have not been able to confirm exists.
10Plagiarism and Research Integrity
Research misconduct is commonly defined as fabrication (making up data or results), falsification (manipulating data, methods or images) and plagiarism (using others' words or ideas without credit). Related problems include duplicate submission of the same work to several venues, undisclosed self-plagiarism, honorary authorship and selective reporting.
When using automated writing or coding tools, follow the venue's policy, disclose their use where required, and remain fully responsible for every statement and result. Authorship should reflect substantial contribution, and every author should approve the final manuscript.
11Peer Review Principles
Peer review is the evaluation of a manuscript by independent experts before publication. Reviewers assess novelty, correctness, significance, clarity and whether the evidence supports the claims. Good reviewers are constructive and specific, declare conflicts of interest, keep manuscripts confidential, and do not use unpublished ideas they encounter. Authors should respond to every point respectfully, revising where reviewers are right and explaining clearly where they disagree.
12Preparing Conference Manuscripts
Conference papers have strict page limits and fixed deadlines. Before submitting to any conference:
- Use the official template and respect page and formatting limits exactly.
- Follow anonymization rules if the review is double-blind.
- Write the abstract and introduction last, once contributions are clear.
- Check that every figure is legible when printed in greyscale.
- Verify the venue's publisher and indexing claims independently, in the databases themselves.
- Confirm registration and presentation requirements before acceptance.
13Understanding Historical CSE 2022 Submission Rules Closed
The original CSE 2022 website invited authors to submit papers in the conference's research areas through a paper submission page. As with most conferences, accepted authors were expected to prepare a final version and register to present.
We do not reproduce the historical CSE 2022 formatting rules, page limits, review criteria, deadlines or publication arrangements, because we cannot verify them precisely, and reproducing them inaccurately could mislead authors who are tracing past papers. The key facts are:
- The conference took place on December 17–18, 2022; all deadlines were before that date.
- Submissions are permanently closed.
- This website does not accept, review or publish papers, and has no connection with any publication that may have resulted from the conference.
For current venues, consult the official call for papers of the conference you intend to submit to. Browse the research learning resources for more background.