Scholar
Yaswanth Narsupalli
Google Scholar ID: 0xLN_4gAAAAJ
B.Tech + M.Tech in AI and ML, IIT Kharagpur
Citations & Impact
Citations
82
H-index
3
i10-index
1
Publications
8
Co-authors
10
list available
Publications
Resume (English only)
Academic Achievements
- First-author paper 'Review-Feedback-Reason (ReFeR)' under review at TMLR 2025; accepted to 2 NeurIPS 2024 workshops
- First-author paper 'DepNeCTI' accepted at EMNLP 2023 (Findings)
- Co-authored 'II-Bench' accepted as poster at NeurIPS Workshop 2024
- Co-authored 'VideoScore' accepted at EMNLP 2024
- Second-author paper 'YinYang-Align' under review
- Co-authored 'DPO Kernels' under review
- Served as Student Paper Reviewer for EMNLP 2023
- Microsoft Research grant doubled to $40,000 in Dec 2023
- Awarded Best Term Papers in 'NLP', 'Graph ML', and 'AI Design Lab' in Apr 2023
Research Experience
- FinAgent: Financial Trading Agent (Apr–Jun 2024, University of Waterloo TIGER Lab with Prof. Wenhu Chen) – Developed a multimodal agent for financial trading and portfolio optimization; built a dataset of financial news and stock prices for 100 companies
- Comparing GANs and diffusion models for image generation (Feb–Apr 2024, IIT Kharagpur)
- Event Extraction using LLMs (Oct 2023–Jan 2024, UT Dallas) – Led research on document- and sentence-level event extraction using GPT-3.5, GPT-4, and open-source LLMs
- Citation Integrity (Jun–Sep 2023, UIUC) – Built a new dataset to classify citations (e.g., 'irrelevant citation') and extract supporting evidence
- Machine Translation using LLMs (May–Sep 2023, UC Berkeley) – Enhanced 500K MT lines via synthetic LLM data; improved BLEU scores by 10% using NLLB model with advanced post-processing
- Nested Compound Parsing and Type Identification (Dec 2022 onward)
