Name
Abstract | ||
|---|---|---|
Question answering for Machine reading evaluation track is a aim to check machine understanding ability of a machine.As we analyzed most crusial part for efficient working of this system is to select text which needs to be considered for understanding since understanding text would involve a lot of NLP processing. This paper covers our submitted system for QA4MRE campaign, Which mostly focuses on two part first being selecting text from comprehension and background knowledge needed to be understand and second being eliminating or ranking options based on selected text from former step.Our main focus was on eliminating and ranking which boils down to tunning various parameter for selection whether to answer particular question if answered how to consider scores,Following methods like calculating cosine between question and passage sentences,cosine of named entities output of passage sentences and question were also considered for scoring .In addition to this basic frame work of our system negation of sentences were also considered to answers which received very close score.We also considered expansion of question and options respectively to collect relevant information from background collection.Entity Co-referencing and normalization were some of important preprocessing to consider on passage and background collection as we analyzed since it can increase score of sentence or option which do not directly mention entity. |
Year | Venue | Field |
|---|---|---|
2011 | CLEF (Notebook Papers/Labs/Workshop) | Question answering,Normalization (statistics),Negation,Ranking,Cosine similarity,Computer science,Preprocessor,Natural language processing,Artificial intelligence,Sentence,Comprehension |
DocType | Citations | PageRank |
Conference | 0 | 0.34 |
References | Authors | |
0 | 2 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Gaurav Arora | 1 | 18 | 2.43 |
| Prasenjit Majumder | 2 | 173 | 25.15 |