search

Call-for-Code-for-Racial-Justice / TakeTwo-DataScience

Call for Code Diverse Representation Problem 3 media bias data science
Apache License 2.0
8 stars 8 forks source link

Develop Data Science MVP Version 2 #17

Open naokiabe opened 3 years ago

naokiabe commented 3 years ago

Background

Solution

DS-MVP-2 An extended machine learning module that can learn to detect racially biased expressions in sequential context based on input labeled data. It would be natural to implement this version utilizing recurrent deep neural networks such as the LSTM architecture. Here, the labeled data consists of <context, expression, classification> triples.

Tasks

Acceptance Criteria

Technical Requirements

github-actions[bot] commented 2 years ago

:wave: Hi! This issue has been marked stale due to inactivity. If no further activity occurs, it will automatically be closed in 14 days.

Preetikasri commented 2 years ago

Tried LSTMs with Glove embeddings, but training done till only 10 epochs due to computational resource constraint. Average accuracy about ~40% on test sample. More epochs needed to be trained for in-order to fully exploit the algorithm. Awaiting GPU allotment.

demilolu commented 2 years ago

not stale or done.

github-actions[bot] commented 2 years ago

:wave: Hi! This issue has been marked stale due to inactivity. If no further activity occurs, it will automatically be closed in 14 days.