Papers by Aditya Shrivastava
X-RiSAWOZ: High-Quality End-to-End Multilingual Dialogue Datasets and Few-shot Agents (2023.findings-acl)
Copied to clipboard
Mehrad Moradshahi, Tianhao Shen, Kalika Bali, Monojit Choudhury, Gael de Chalendar, Anmol Goel, Sungkyun Kim, Prashant Kodali, Ponnurangam Kumaraguru, Nasredine Semmar, Sina Semnani, Jiwon Seo, Vivek Seshadri, Manish Shrivastava, Michael Sun, Aditya Yadavalli, Chaobin You, Deyi Xiong, Monica Lam
| Challenge: | X-RiSAWOZ dataset has more than 18,000 human-verified dialogue utterances for each language . Xiaoping and Xinhui are the main challenges for task-oriented dialogue research . |
| Approach: | They develop a toolkit to accelerate the post-editing of a new language dataset after translation . their dataset, code, and toolkit are released open-source . |
| Outcome: | The proposed toolkit accelerates the post-editing of a new language dataset after translation. |
Corpus Creation and Emotion Prediction for Hindi-English Code-Mixed Social Media Text (N18-4)
Copied to clipboard
| Challenge: | Emotion Prediction is a natural language processing task dealing with detection and classification of emotions in monolingual and bilingual texts. |
| Approach: | They propose a machine learning system which uses various machine learning techniques to detect emotion associated with tweets. |
| Outcome: | The proposed system uses various machine learning techniques to detect emotion associated with the text. |
“Knowledge is Power”: Constructing Knowledge Graph of Abdominal Organs and Using Them for Automatic Radiology Report Generation (2023.acl-industry)
Copied to clipboard
Kaveri Kale, Pushpak Bhattacharyya, Aditya Shetty, Milind Gune, Kush Shrivastava, Rustom Lawyer, Spriha Biswas
| Challenge: | conventional radiology workflows involve dictating diagnosis to transcriptionists, which is prone to delay and error. |
| Approach: | They propose to generate a set of knowledge graphs from a large collection of free-text radiology reports and use them to generate automatic radiology report generation. |
| Outcome: | The proposed model improves the reported BLEU-3, ROUGE-L, METEOR, and CIDEr scores by 2%, 4%, 2% and 2% respectively. |
An Automatic Method to Estimate Correctness of RAG (2025.coling-industry)
Copied to clipboard
| Challenge: | Existing methods to assess the correctness of RAG models fail to capture the model’s internal state during answer generation. |
| Approach: | They propose a method to predict the correctness of RAG models by modeling the model’s uncertainty on quantified perturbations of input. |
| Outcome: | Extensive experiments across multiple large language models show that the proposed approach quantifies RAG robustness by aligning predictions with ground truth with a MSE 0.002 while offering flexibility for diverse qualitative metrics. |