Welcome to my blog!
This blog is a documentation for the interesting aspects of my work, my thoughts and more. Navigate to the GSoC tag in the tags section for following my work on the DBpedia Hindi Chapter.
Featured
-
GSoC Weeks 11-12: Predicate Linking Integration and Final Week Touches
Published: at 07:00 AMWeeks 11-12 of GSoC'25: Integrating Predicate Linking and Closing the Project
-
GSoC Week 10: Predicate Linking
Published: at 07:00 AMWeek 10 of GSoC'25: implementing predicate linking for Hindi DBpedia
-
GSoC Week 8-9: Paper Review, SPARQL Endpoint Demonstration, and Synthetic Data Generation
Published: at 07:00 AMWeeks 8-9 of GSoC'25: Reviewing multilingual SPARQL paper, demonstrating Hindi DBpedia endpoint capabilities, and generating synthetic data for finetuning
-
GSoC Week 7: Synthetic Data Generation and Performance Testing of Hindi Chapter
Published: at 07:46 AMWeek 7 of GSoC'25: Rewriting the synthetic data generation script and performance testing of SPARQL endpoint
-
GSoC Week 6: IndIE Enhancement + Finetuning
Published: at 12:14 AMWeek 6 of GSoC'25: Pushing the performance of our pipeline for Hindi information extraction.
-
GSoC Week 5: Link Prediction and indIE Enhancement
Published: at 11:16 PMWeek 5 of GSoC'25: Implementing and benchmarking link prediction models for our existing Hindi knowledge graph
-
GSoC Week 4: Enhancing OIE using ReAct prompting + Link Prediction
Published: at 06:14 PMWeek 4 of GSoC'25: Trying new LLMs, using the ReAct prompting strategy and analysing our graph for link prediction models.
-
GSoC Week 3: Pushing the boundary of LLMs for Hindi Relation Extraction
Published: at 11:42 AMWeek 3 of GSoC'25: Building, evaluating, and benchmarking LLM-based information extraction for Hindi using the BenchIE dataset. Models, prompts, metrics, and what we learned.
-
GSoC Weeks 1-2: Bootstrapping DBpedia Hindi Extraction
Published: at 05:37 PMWeeks 1 & 2 of my GSoC'25 journey: setting up the DBpedia Information Extraction Framework and Neural Extraction Framework for the Hindi chapter, overcoming roadblocks, and first learnings.