Pinned Repositories
100LinesOfCode
Let's build something productive in less than 100 Lines of Code.
601.771
BLI-for-Indic-languages
This is the code for our paper <put link here>
character-bert
Main repository for "CharacterBERT: Reconciling ELMo and BERT for Word-Level Open-Vocabulary Representations From Characters"
character-bert-pretraining
Code for pre-training CharacterBERT models (as well as BERT models).
CSCBLI
Code for the ACL2021 paper "Combining Static Word Embedding and Contextual Representations for Bilingual Lexicon Induction"
DeepLearningForShortStories
This is for the course Deep Learning for the Processing and Interpretation of Literary Texts
Handling-English-VPE-for-English-Hindi-MT
English-Hindi machine translation systems have difficulty interpreting verb phrase ellipsis (VPE) in English, and commit errors in translating sentences with VPE. We present a solution and theoretical backing for the treatment of English VPE, with the specific scope of enabling English-Hindi MT, based on an understanding of the syntactical phenomenon of verb-stranding verb phrase ellipsis in Hindi (VVPE). We implement a rule-based system to perform the following sub-tasks: 1) Verb ellipsis identification in the English source sentence, 2) Elided verb phrase head identification 3) Identification of verb segment which needs to be induced at the site of ellipsis 4) Modify input sentence; i.e. resolving VPE and inducing the required verb segment. This system obtains 94.83 percent precision and 83.04 percent recall on subtask (1), tested on 3900 sentences from the BNC corpus [Leech, 1992]. This is competitive with state-of-the-art results. We measure accuracy of subtasks (2) and (3) together, and obtain a 91 percent accuracy on 200 sentences taken from the WSJ cor- pus[Paul and Baker, 1992]. We carried out a manual analysis of the MT outputs of 100 sentences after passing it through our system. We set up a basic metric (1-5) for this evaluation, where 5 indicates drastic improvement, and obtained an average of 3.55.
north-indian-dialect-modelling
Collecting data for "dialects" in the North Indian "Hindi belt". Modelling the dialect system to gain insight and to develop NLP research for low-resource languages.
XORQA
This is the official repository for NAACL 2021, "XOR QA: Cross-lingual Open-Retrieval Question Answering".
niyatibafna's Repositories
niyatibafna/north-indian-dialect-modelling
Collecting data for "dialects" in the North Indian "Hindi belt". Modelling the dialect system to gain insight and to develop NLP research for low-resource languages.
niyatibafna/BLI-for-Indic-languages
This is the code for our paper <put link here>
niyatibafna/Handling-English-VPE-for-English-Hindi-MT
English-Hindi machine translation systems have difficulty interpreting verb phrase ellipsis (VPE) in English, and commit errors in translating sentences with VPE. We present a solution and theoretical backing for the treatment of English VPE, with the specific scope of enabling English-Hindi MT, based on an understanding of the syntactical phenomenon of verb-stranding verb phrase ellipsis in Hindi (VVPE). We implement a rule-based system to perform the following sub-tasks: 1) Verb ellipsis identification in the English source sentence, 2) Elided verb phrase head identification 3) Identification of verb segment which needs to be induced at the site of ellipsis 4) Modify input sentence; i.e. resolving VPE and inducing the required verb segment. This system obtains 94.83 percent precision and 83.04 percent recall on subtask (1), tested on 3900 sentences from the BNC corpus [Leech, 1992]. This is competitive with state-of-the-art results. We measure accuracy of subtasks (2) and (3) together, and obtain a 91 percent accuracy on 200 sentences taken from the WSJ cor- pus[Paul and Baker, 1992]. We carried out a manual analysis of the MT outputs of 100 sentences after passing it through our system. We set up a basic metric (1-5) for this evaluation, where 5 indicates drastic improvement, and obtained an average of 3.55.
niyatibafna/XORQA
This is the official repository for NAACL 2021, "XOR QA: Cross-lingual Open-Retrieval Question Answering".
niyatibafna/100LinesOfCode
Let's build something productive in less than 100 Lines of Code.
niyatibafna/601.771
niyatibafna/character-bert
Main repository for "CharacterBERT: Reconciling ELMo and BERT for Word-Level Open-Vocabulary Representations From Characters"
niyatibafna/character-bert-pretraining
Code for pre-training CharacterBERT models (as well as BERT models).
niyatibafna/CSCBLI
Code for the ACL2021 paper "Combining Static Word Embedding and Contextual Representations for Bilingual Lexicon Induction"
niyatibafna/DeepLearningForShortStories
This is for the course Deep Learning for the Processing and Interpretation of Literary Texts
niyatibafna/Email-Clustering-on-the-Enron-Dataset
Project for CS-303. Working with public available dataset Enron (https://www.cs.cmu.edu/~./enron/), that contains approximately 0.5 million messages collected from 150 users, to model a classification.
niyatibafna/embeddings-transfer-indian-languages
Transferring embeddings to low resource Indian languages using close relationships to other higher resource languages such as Hindi, Bangla, Marathi, etc.
niyatibafna/gina
Learning a Hindi lexicon from parallel corpora. Monsoon 2018. Google Cloud NLP API.
niyatibafna/HateSpeech-Hindi-English-Code-Mixed-Social-Media-Text
niyatibafna/Hindi-Sentence-Completion
Cleaned final code from Hindi-Verb-Prediction
niyatibafna/llm-eval-crosslingual-generalization
niyatibafna/lm_hf_skeleton
Skeleton scripts in HF.
niyatibafna/misc
Useful things
niyatibafna/mlmm-evaluation
Multilingual Large Language Models Evaluation Benchmark
niyatibafna/mt_hf_skeleton
Setting up MT in HF
niyatibafna/OBPE
niyatibafna/pgns-for-lrmt
niyatibafna/political_health
This is for measuring hate on Twitter against certain groups, and comparing these metrics over time
niyatibafna/retaining-source-terms-nmt
When we are translating technical material from English to Hindi, we may often want to retain certain terminology for consistency and coherence in Hindi. This task deals with constrained decoding of English-Hindi NMT to accomplish this goal i.e. given source English text, and a list of English terms that we want to retain, we want the output in target language Hindi that uses the required English terminology.