Class participation, Google Folder.

(Back to Course Index Page)

For the midterm assignment, work between weeks 3 and 7 to write an 800-word abstract

Read: Sklearn User Guide (Section 1: Supervised Learning). Focus on the Iris dataset examples.

Locate and review the Iris examples in the following topics:

• Logistic Regression (Regularization path of L1-logistic regression)

• Support Vector Machines (SVM-Anova: SVM with univariate feature selection)

• Nearest Neighbors Classification

• Decision Trees (decision surface + tree structure)

• Ensemble Methods (ensemble decision surfaces)

• Voting Classifier

Compare the algorithms on a common setup: using the same train/test split (or 5-fold cross-validation), compare these methods on Iris.

• Accuracy (required)

• Macro F1-score (recommended)

• Confusion matrix (required for at least 2 models)

• Decision boundary visualization (required for at least 2 models)

Write a short summary (AAAI format): Submit a 1–2 page report in AAAI two-column format. Your report should include:

• A comparison table of results across models

• 1–2 figures (decision boundaries or confusion matrices)

• A short discussion: Which model worked best? Which was easiest to interpret? Why?

AAAI is one of the top four AI conferences

Read the guide from AAAI 2026

AAAI 2026 Guidelines for papers

Use the LaTeX format

Overleaf Tutorials

AAAI 2026 Template on Overleaf

Link sharing, each team work on the same overleaf project

Review the example dataset (EEGEyeNet, NeurIPS Datasets & Benchmarks 2021): read the paper and skim the GitHub repository. Pay attention to the task definition, dataset structure, labels, and the baseline results table.

Link: https://openreview.net/forum?id=Nc2uduhU9qa

Choose datasets for your Midterm + Final Project: identify two candidate datasets from the NeurIPS Datasets & Benchmarks track (2021–2024).

Start here: https://neurips.cc/Conferences/2025/CallForDatasetsBenchmarks

Present your dataset analysis: for each dataset, briefly summarize (1) what the ML task is, (2) what the inputs/labels are, and (3) what models are most appropriate. Confirm the dataset can support multiple ML methods we will cover in this course (e.g., traditional ML baselines, deep learning, sequence models, transformers).

Submit a short written summary (1 page max): explain why you selected these datasets, what makes them suitable for the midterm/final project, and which metrics you plan to use for evaluation.

Each studentonly submits one file, lab_3_lastname.zip, including

Start a shared Github project for midterm and final project, add the instructor’s account 'xqu'

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.