Skip to content

An AI Generated Text detect model using KearsNLP, improved model performance by several methods, including trying different pre-trained models (BERT, RoBERTa, Deberta), data augmentation, trying different model architectures and layers, learning rate scheduling, hyperparameter tuning, adding Dropout layers, trying different word embeddings, etc.

Notifications You must be signed in to change notification settings

Anna-Andrea/An-LSTM-Based-Model-for-LLM-Generated-Text-Detection

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Kaggle Cometitions: LLM - Detect AI Generated Text

Summary:

● Implemented an LTSM-based model for LLM text identification, including embedding layer, LSTM layer, fully connected layer, and other layers.

● The process included data preprocessing, feature extraction, clustering, dimension reduction, model optimization and other steps.

● Model optimization approaches including using pre-trained language models (BERT, RoBERTa, Deberta, etc.), data enhancement, optimizing network architecture and number of layers, learning rate scheduling, hyperparameter optimization, adding Dropout layers, trying different word embeddings, etc.

● This model has a large improvement compared to the baseline (based on LR, SGD and NB Bernoulli's voting).

About

An AI Generated Text detect model using KearsNLP, improved model performance by several methods, including trying different pre-trained models (BERT, RoBERTa, Deberta), data augmentation, trying different model architectures and layers, learning rate scheduling, hyperparameter tuning, adding Dropout layers, trying different word embeddings, etc.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published