Conference Papers

Permanent URI for this collectionhttps://idr.nitk.ac.in/handle/123456789/28506

Browse

Author: search.filters.author.Ravi, S.S.Has files: No

Search Results

Now showing 1 - 3 of 3
  • No Thumbnail Available
    Item
    ARS NITK at MEDIQA 2019: Analysing various methods for natural language inference, recognising question entailment and medical question answering system
    (Association for Computational Linguistics (ACL), 2019) Agrawal, A.; George, R.A.; Ravi, S.S.; Kamath S․, S.S.; Anand Kumar, M.A.
    This paper includes approaches we have taken for Natural Language Inference, Question Entailment Recognition and Question-Answering tasks to improve domain-specific Information Retrieval. Natural Language Inference (NLI) is a task that aims to determine if a given hypothesis is an entailment, contradiction or is neutral to the given premise. Recognizing Question Entailment (RQE) focuses on identifying entailment between two questions while the objective of Question-Answering (QA) is to filter and improve the ranking of automatically retrieved answers. For addressing the NLI task, the UMLS Metathesaurus was used to find the synonyms of medical terms in given sentences, on which the InferSent model was trained to predict if the given sentence is an entailment, contradictory or neutral. We also introduce a new Extreme gradient boosting model built on PubMed embeddings to perform RQE. Further, a closed-domain Question Answering technique that uses Bi-directional LSTMs trained on the SquAD dataset to determine relevant ranks of answers for a given question is also discussed. Experimental validation showed that the proposed models achieved promising results. © 2019 Association for Computational Linguistics
  • No Thumbnail Available
    Item
    Leveraging multimodal behavioral analytics for automated job interview performance assessment and feedback
    (Association for Computational Linguistics (ACL), 2020) Agrawal, A.; George, R.A.; Ravi, S.S.; Kamath S․, S.; Anand Kumar, M.
    Behavioral cues play a significant part in human communication and cognitive perception. In most professional domains, employee recruitment policies are framed such that both professional skills and personality traits are adequately assessed. Hiring interviews are structured to evaluate expansively a potential employee’s suitability for the position - their professional qualifications, interpersonal skills, ability to perform in critical and stressful situations, in the presence of time and resource constraints, etc. Therefore, candidates need to be aware of their positive and negative attributes and be mindful of behavioral cues that might have adverse effects on their success. We propose a multimodal analytical framework that analyzes the candidate in an interview scenario and provides feedback for predefined labels such as engagement, speaking rate, eye contact, etc. We perform a comprehensive analysis that includes the interviewee’s facial expressions, speech, and prosodic information, using the video, audio, and text transcripts obtained from the recorded interview. We use these multimodal data sources to construct a composite representation, which is used for training machine learning classifiers to predict the class labels. Such analysis is then used to provide constructive feedback to the interviewee for their behavioral cues and body language. Experimental validation showed that the proposed methodology achieved promising results. © 2017 Association for Computational Linguistics
  • No Thumbnail Available
    Item
    Benchmarking semantic, centroid, and graph-based approaches for multi-document summarization
    (Springer Science and Business Media Deutschland GmbH info@springer-sbm.com, 2021) Agrawal, A.; George, R.A.; Ravi, S.S.; Kamath S․, S.
    Multi-document summarization (MDS) is a pre-programmed process to excerpt data from various documents regarding similar topics. We aim to employ three techniques for generating summaries from various document collections on the same topic. The first approach is to calculate the importance score for each sentence using features including TF-IDF matrix as well as semantic and syntax similarity. We build our algorithm to sort the sentences by importance and add it to the summary. In the second approach, we use the k-means clustering algorithm for generating the summary. The third approach makes use of the Page Ranking algorithm wherein edges of the graph are formed between sentences that are syntactically similar but are not semantically similar. All these techniques have been used to generate 100–200 word summaries for the DUC 2004 dataset. We use ROUGE scores to evaluate the system-generated summaries with respect to the manually generated summaries. © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd 2021.