Papers by Kumar Saunack
How low is too low? A monolingual take on lemmatisation in Indian languages (2021.naacl-main)
Copied to clipboard
| Challenge: | Prior work on ML based lemmatization focused on high resource languages, where data sets (word forms) are readily available. |
| Approach: | They propose to use neural methods to relate inflected forms of words to their dictionary form to reduce the sparse data problem. |
| Outcome: | The proposed methods can give competitive accuracy even in low resource setting. |
A Graph-Based Method for Unsupervised Knowledge Discovery from Financial Texts (2022.lrec-1)
Copied to clipboard
| Challenge: | A financial analyst's work involves manually reviewing lengthy filings and financial news articles in order to extract relevant pieces of information. |
| Approach: | They propose an end-to-end, fully unsupervised method for knowledge discovery from financial texts that integrates existing resources to construct a knowledge graph of companies and related entities. |
| Outcome: | The proposed method calculates the environmental rating for companies in the S&P 500 based on company filings with the SEC and provides an independent assessment of its outputs with an independent MSCI source. |
Analysing cross-lingual transfer in lemmatisation for Indian languages (2020.coling-main)
Copied to clipboard
| Challenge: | Inference-based scripts such as Abjad are difficult for cross-lingual models to learn in extremely low resource scenarios. |
| Approach: | They evaluate cross-lingual approaches for low resource languages and compare their performance against other models using different linguistic factors. |
| Outcome: | The proposed model on six low resource languages from two different families is compared with monolingual models on morphologically rich Indian languages. |