http://repository.iiitd.edu.in/xmlui/handle/123456789/1 CSE Wed, 02 Sep 2026 23:36:13 GMT 2026-09-02T23:36:13Z http://repository.iiitd.edu.in/xmlui/handle/123456789/2078 Target-augmented shared fusion based multimodal sarcasm explanation generation Goel, Palaash; Akhtar, Md. Shad (Advisor) Multimodal Sarcasm Explanation (MuSE) is a challenging natural language understanding task that deals with training machines to understand the semantic incongruence present in sarcastic social media posts comprising of an image and a corresponding text caption and generating a natural language explanation to reveal the implicit (hidden) meaning behind them. We use the MORE dataset for the same. The current state of the art for this task (14) makes use of a ‘multi-source semantic graph’ which incor porates external world knowledge along with concepts extracted from both the images and their corre sponding captions to facilitate the reasoning process and lead to a better explanation model. After careful analysing of some of their limitations, we proposed the novel Target-aUgmented shaRed fusion-Based sarcasm explanatiOn model, aka. TURBO, for the task of MuSE that: 1. Utilizes a knowledge graph to incorporate external knowledge, similar to TEAM, while overcom ing the aforementioned limitations 2. Incorporates a novel shared fusion mechanism for learning important information from both the visual and textual modalities 3. Utilizes manually annotated information about the target of sarcasm in our model 4. Beats the current state-of-the-art by roughly 2-3% on average It is important to note that TURBOfixes a problem in the implementation of the previous version of this model (presented as part of last semester’s work). Additionally, we replace the term “cause of sarcasm” with “target of sarcasm” since the latter more appropriately represents the purpose/meaning of the annotation and is thus, easier to understand as well. Thu, 28 Nov 2024 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/2078 2024-11-28T00:00:00Z http://repository.iiitd.edu.in/xmlui/handle/123456789/2076 Generative genomics Verma, Shriya; Sethi, Tavpritesh (Advisor) Generative genomics refers to the field of study focused on developing computational models and algorithms capable of generating synthetic genomic data. This report explores the application of transformer-based architectures for generating synthetic genomic data. The literature review gives an overview of different transformer-based architectures explored through the term. Some challenges faced during implementation and the innovative solutions for them are also included. The potential applications of generative genomics and the scope for future research in the evolving field are also covered in this report. Wed, 27 Nov 2024 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/2076 2024-11-27T00:00:00Z http://repository.iiitd.edu.in/xmlui/handle/123456789/2074 Generative genomics Kamra, Samanyu; Sethi, Tavpritesh (Advisor) Generative genomics generally refers to a study area that undertakes computational model devel opment and algorithms capable of developing synthetic genomic data. This report explores the application of transformer-based architectures for generating synthetic genomic data. The litera ture review provides an overall view of different architectures based on the transformer explored through the term. It also discusses some challenges during implementation and their solutions with innovation presented for the first time. It also covers the possible applications of generative genomics and the area where the scope for future research in the evolving field lies. Retrieval-Augmented Generation (RAG) is a powerful technique in the field of Natural Language Processing (NLP) that combines the benefits of both retrieval-based and generative models. RAG models integrate a re trieval step into the generative process, allowing the model to access external knowledge stored in large-scale databases or corpora during the generation of text. Wed, 27 Nov 2024 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/2074 2024-11-27T00:00:00Z http://repository.iiitd.edu.in/xmlui/handle/123456789/2070 Application of DL in NER in recipes Singh, Ritisha; Bagler, Ganesh (Advisor) Named Entity Recognition (NER) is a pivotal technique for extracting structured information from unstructured or semi-structured data. This project focuses on applying deep learning (DL) tech niques to NER for ingredient phrases, leveraging a manually annotated dataset of approximately 10,000 rows. Multiple approaches were explored, including statistical methods, fine-tuning deep learning-based language models, and few-shot prompting with large language models (LLMs). Mod els such as spaCy-transformer, DistilBERT, BERT, and DistilRoBERTa were initially implemented as described in reference literature, achieving F1 scores in the 90s. Through custom optimizations, the performance of these models improved to 93–94%, yet still fell short of benchmark results. To address these limitations, a second dataset comprising 10,000 rows was integrated after resolving inconsistencies, creating a unified dataset of 18,000 rows. However, combining the datasets led to a performance drop, revealing challenges related to dataset quality and model generalization. Further evaluations, including cross-validation and bucket analysis, were conducted with BERT, the best performing model, to gain deeper insights into its strengths and limitations. This report presents a comprehensive overview of the methods, challenges, and results, emphasizing the need for continuous refinement and highlighting opportunities for enhancing model performance. Sun, 01 Dec 2024 00:00:00 GMT http://repository.iiitd.edu.in/xmlui/handle/123456789/2070 2024-12-01T00:00:00Z