CS 4364 Mid-term Project

  • Lab 6 Review, Due Wednesday, 02/25/2026, by midnight (23:59, EST)

  • Midterm: Due Wednesday, 03/04/2026, by midnight (23:59, EST)

Announcements

1. Your mid-term project

  • Select a dataset from the NeurIPS datasets 2021 to 2024

  • You can choose one of the following topics for your project:

  • Classification (Supervised Learning)

    1. Objective: Train a classification model similar to the Iris dataset example.

    2. Suggested Approach: Use a labeled dataset, such as Iris or EEGEyeNet, to classify data points into predefined categories.

    3. Deliverable: A classification model and accuracy analysis, including insights from model evaluation (e.g., precision, recall, F1-score).

  • Regression (Supervised Learning)

    1. Objective: Build a regression model similar to the Diabetes dataset example.

    2. Suggested Approach: Use a dataset where the goal is to predict a continuous value, such as housing prices or stock values.

    3. Deliverable: A regression model with a report on performance metrics (e.g., R-squared, mean squared error).

  • Clustering (Unsupervised Learning)

    1. Objective: Group data points without predefined labels, as in the Iris dataset clustering example.

    2. Suggested Approach: Apply clustering algorithms (e.g., K-Means, DBSCAN) to an unlabeled dataset and visualize the clusters.

    3. Deliverable: A clustering model with a discussion on the number of clusters and how well the data points are grouped.

  • Outlier Detection

    1. Objective: Develop a model to detect anomalies or outliers in a dataset.

    2. Suggested Approach: Use unsupervised or semi-supervised methods (e.g., Isolation Forest, One-Class SVM) on a dataset with a known structure but unexpected outliers.

    3. Deliverable: An outlier detection system and an analysis of the performance in identifying true anomalies.

  • Alternative: Research Review Paper

    1. If you prefer theoretical research, you may write a review paper on one of the selected topics.

    2. Your review paper should summarize recent advancements and propose improvements or future research directions. It should cover at least thirty academic papers published within the last five years.

  • Write the midterm paper in AAAI format (two columns, using the overleaf template for AAAI). Your drafts should include:

    1. Title, anonymous authors

    2. Introduction

    3. Related work

    4. Methods

    5. Results (optional)

    6. Discussion (optional)

    7. Conclusion (optional)

  • The Section of Introduction should includes

    1. Research Questions

    2. Motivation

  • The Section of Related work should includes

    1. Literature Review

    2. State-of-the-art

  • (Optional) Compare the performance of machine learning algorithms:

    1. Select at least two algorithms for comparison.

    2. This step is optional for the midterm but required for the final project.

  • (Optional) Submit your Python code:

    1. Ensure the README is clear for recreating your results.

    2. Test the code on both computers, and include screenshots of the output.

    3. Cite references, especially for existing code, and note if results were replicable.

    4. Do not email the dataset. Instead, provide instructions for downloading it, similar to the README example of EEGEyeNet.

  • Submit the midterm draft in a double-blind way for future internal peer review.

  • Hide the author’s information for double-blind peer review.

  • Hide author information everywhere in your submission, including PDF, TXT, and Python files.

  • Double-blind example, NeurIPS

  • 'Authors are responsible for anonymizing their submissions. In particular, they should not include author names, author affiliations, or acknowledgements in their submissions and they should avoid providing any other identifying information (even in the supplementary material).'

  • Submit the midterm in a double-blind way as required by the AAAI conference.

  • Your midterm paper should be a starting point to consider the rest of the final paper.

  • Your final paper will have more details, especially for the following sections:

    1. Methods

    2. Results

    3. Discussion

    4. Conclusion

2. Submission Guide

  • Each team should submit a single file on behalf of the entire group. Ensure that all team members are Mentioned in a separated author.txt file

  • The files should be named as midterm_lastname.zip, the zip file size should be smaller than 5M, including

    1. midterm_lastname1_lastname2.PDF for your midterm paper in AAAI format.

    2. (Optional) A code folder for ALL your Python Code and support files for your midterm paper.

5. Notes

  • Submit your zip file for the lab to BlackBoard.

  • Lab assignments will typically be released on Thursday and will be due by midnight on the following Wednesday.

  • Lab 6 Review, Due Wednesday, 02/25/2026, by midnight (23:59, EST)

  • Midterm: Due Wednesday, 03/04/2026, by midnight (23:59, EST)