Vyas Raina

prepend_acoustic_attack. Prepend universal audio attack segment to mute Whisper

attack-comparative-assessment. Adversaial attack comparative assessment Large Language Model

AGNewsClassification. Classification of AG News dataset

IMDB_Sentiment_Classification. Sentiment Analysis IMDB dataset using different model architectures: ELECTRA, BERT, RoBERTa, XLNet

GrammaticalErrorCorrection. Train a Transformer-based seq2seq GEC model to correct grammatical errors

gec_llm_combination. Using LLMs for blackbox GEC ensembling

adv_train_nlp. Different forms of adversarial training for nlp

adv_attack_whisper. Adversarial attack Whisper Model

AudioSafety. Perturb the audio to selectively mute a target ASR sysem

mbr_gec. Minimum Bayes' Risk for Grammatical Error Correction

tabular_neural_vs_gbt. Comparing Neural and Gradient Boosted Tree Models on Domain shifted tabular data

deep_pron. End-End deep learning approach to spoken language assessment

NAACL-2022-Residue-Detector. Code for paper submission to NAACL 2022

img_attackability. Identify attackable samples

ML_FPL. Machine Learning Fantasy Premier League

Xvector_grader. Uses x-vector representation to predict SLA grades

AttackDetectGEC. Detection based defence of adversarial attacks on Grammatical Error Correction Systems