This is a 10-minute Final Year Project proposal defense for a university committee that has not seen this project before. The audience is faculty evaluators judging whether the research direction is well-justified, technically sound, and worth approving — not a general audience and not investors. What it needs to accomplish: Convince the committee that (1) this is a genuine, well-documented research gap — Urdu has no benchmark or evaluated framework for omission-based misleading news, distinct from fake news/clickbait/sensationalism; (2) the research questions are concrete and each one maps to a specific, already-designed methodology step, not vague aspiration; (3) the project is a legitimate research contribution even though it adapts an existing methodology (OM-GUARD) rather than inventing a new one — the Urdu corpus and the empirical test of whether the method transfers to a right-to-left, SOV-structured language are the novel parts; and (4) the project is a research-first thesis (dataset + fine-tuned model), not a product build — the prototype is a lightweight validation tool, not the deliverable. What should happen after they see it: The committee approves the proposal to proceed, without requiring a change in direction — or, if they push back, their questions are answerable directly from what's already documented in the full proposal (methodology, evaluation plan, feasibility, cost) rather than exposing a gap the team hasn't thought throug
This is a 10-minute Final Year Project proposal defense for a university committee that has not seen this project before. The audience is faculty evaluators judging whether the research direction is well-justified, technically sound, and worth approving — not a general audience and not investors.
What it needs to accomplish:
Convince the committee that (1) this is a genuine, well-documented research gap — Urdu has no benchmark or evaluated framework for omission-based misleading news, distinct from fake news/clickbait/sensationalism; (2) the research questions are concrete and each one maps to a specific, already-designed methodology step, not vague aspiration; (3) the project is a legitimate research contribution even though it adapts an existing methodology (OM-GUARD) rather than inventing a new one — the Urdu corpus and the empirical test of whether the method transfers to a right-to-left, SOV-structured language are the novel parts; and (4) the project is a research-first thesis (dataset + fine-tuned model), not a product build — the prototype is a lightweight validation tool, not the deliverable.
What should happen after they see it:
The committee approves the proposal to proceed, without requiring a change in direction — or, if they push back, their questions are answerable directly from what's already documented in the full proposal (methodology, evaluation plan, feasibility, cost) rather than exposing a gap the team hasn't thought throug
Created using ChatSlide
This project aims to address the misleading nature of Urdu news previews by adapting the OM-GUARD framework. It will explore how headline-image pairs influence interpretation and differentiate between omission, fake news, and clickbait. The study will benchmark omission in Urdu by collecting previews from major news outlets and pairing them with full articles. Key contributions include establishing the first Urdu omission benchmark, creating an annotated multimodal research corpus, and...