Project Checkpoint 1

Due by 10pm Friday 10/21

Baseline Model

The goal of this checkpoint is to make sure that you have been able to obtain and prepare your dataset, and build and evaluate a baseline model for it. A baseline model is a simple model that performs the same end-to-end calculation as your final project model, but is simpler. e.g. for classification tasks, you could build a logistic classifier, or 3-layer convolutional network. For translation, you could build a basic sequence-to-sequence model without attention. What you build for the baseline depends on the nature of your project. If you're not sure what a reasonable baseline model would be, please ask the staff.

By the end of this checkpoint, you should have a complete pipeline for testing your model and can spend the rest of the project developing, innovating and documenting it. Your baseline model will help you quantify the performance of your final model.

Writeup

For this checkpoint, you should produce a 4-page write-up. You should submit a pdf file. You can re-use parts of your project proposal write-up. This write-up should include these sections:

The Team

Give a list of team members and their responsibilities.

Problem Statement and Background (should include state-of-the-art)

A high-level statement of the problem you are addressing, i.e. recognition of objects in images. Be specific - i.e. "recognition" in this case means outputting a label for the object and also a bounding box for its location.

Give some background on the problem: why it is important - and why it is challenging.

NEW: describe what you think is the state-of-the-art model for this problem, and how it relates to the final model you plan to build.

The Data Source and Preprocessing

Describe the nature and source of the dataset you are using. Describe any pre-processing you did.

Baseline and Final Model

Describe the baseline model you built this time, and the final model you plan to build later. How do you expect the final model to improve performance?

Description of the Tools Used

Describe the tools used. Discuss any issues or challenges you faced so far related to those tools.

Evaluation

How did you evaluate your baseline model? Try to give informative measures such as ROC plots, in addition to basic accuracy or AUC measures. Use graphs.

Submission

Use this link to submit your writeup:

https://bcourses.berkeley.edu/courses/1453965/assignments/7746703

This writeup will be a group submission.