Set B - Question 3

Question Details

• A. A postgraduate student is selected to present a research paper at an academic conference. The student has only 10 minutes to communicate the research effectively to a mixed audience of researchers and industry professionals. As the presenter, design an appropriate presentation plan by selecting the key elements of the research paper and explain how you would organize and communicate them to maintain clarity, audience engagement, and effective delivery. (5 Marks)
• B. A postgraduate student has completed a research project and is invited to present the work at a national conference, where only an extended abstract is required. Due to the limited space, the student must carefully select and organize the key aspects of the research. As the student, design a suitable structure for the extended abstract and justify the selection and sequence of the research problem, objectives, methodology, major findings, and significance to communicate the research contribution effectively to the conference audience. (5 Marks)
• C. Identify the type of thesis evaluation in which a researcher presents and defends the research findings before an expert panel. Explain its purpose in evaluating the researcher's work. (5 Marks)
• D. A research scholar has developed an AI-based system for identifying crop diseases using plant leaf images. The scholar has collected and preprocessed image data, developed a machine-learning model, evaluated its performance, and obtained results such as accuracy, sensitivity, specificity, F1-score, and confusion matrix. The research scholar is ready to document the outcomes of the proposed study. Suggest an appropriate research report structure and justify the contents included in its major sections. (5 Marks)

---

Model Answer

A. Design a 10-minute conference presentation plan for a mixed audience of researchers and industry professionals. (5 Marks)

1. 10-Minute Slide Budget & Organization (3 Marks)

Time Slot	Slide Topic	Content Communicated	Delivery Focus
0:00 - 1:30	Context & Problem (1 slide)	High-level industry motivation and specific research problem.	Hook both audiences with practical real-world impact.
1:30 - 3:30	Proposed Solution / Architecture (2 slides)	Core methodological novelty and workflow schematic diagram.	Emphasize engineering design and theoretical soundness.
3:30 - 7:00	Key Results & Benchmarks (2 slides)	Quantitative evaluation, comparison against SOTA, and ablation tables.	Show statistical rigor (researchers) and accuracy/speed (industry).
7:00 - 8:30	Practical Implications & ROI (1 slide)	Deployment latency, compute cost, and practical field usability.	Highlight business value, scalability, and operational adoption.
8:30 - 10:00	Conclusion & Q&A (1 slide)	3 bullet takeaways, open limitations, and invite discussion.	Keep strictly within time to respect session chairs and delegates.

2. Strategies for Mixed Audience Engagement (2 Marks)

• Dual-Layered Content: Present intuitive visual architectures on slides while verbally referencing formal algorithmic parameters.
• Clarity Over Clutter: Limit text to max 5 bullet points per slide; use bold callout metrics rather than raw unformatted data tables.

---

B. Design a suitable structure for an extended abstract and justify the selection and sequence of its components. (5 Marks)

1. Extended Abstract Structure (2-Page / ~1000 Words) (2.5 Marks)

1. Title & Abstract Summary: Descriptive title, author affiliations, and a 100-word teaser summary.
2. Introduction & Problem Statement: Specific research gap, practical motivation, and scope.
3. Research Objectives: 2-3 precise bulleted objectives.
4. Methodology / Technical Blueprint: High-level system architecture, dataset, baseline models, and experimental setup.
5. Key Results & Findings: 1 essential comparative table or graph showing accuracy/efficiency gains.
6. Significance, Impact & References: Practical contribution, limitations, and top 5 citations.

2. Sequence Justification (2.5 Marks)

The sequence mirrors the funnel model of scientific deduction (Why $\rightarrow$ What $\rightarrow$ How $\rightarrow$ Results $\rightarrow$ Value):

• Why: Problem formulation justifies why conference reviewers should care.
• What & How: Objectives and methodology demonstrate technical execution and reproducibility.
• Results & Impact: Concludes by proving empirical success, enabling conference committees to judge quality under strict page constraints.

---

C. Identify the type of thesis evaluation where a researcher presents and defends findings before an expert panel, and explain its purpose. (5 Marks)

1. Evaluation Type (1 Mark)

• Oral Defense / Viva-Voce Examination.

2. Purpose in Thesis Evaluation (4 Marks)

1. Authentication of Authorship: Verifies that the candidate personally conducted the experimental/analytical research and understands the foundational mechanics.
2. Defense of Critical Decisions: Tests the candidate's rationale for selecting specific datasets, algorithm architectures, loss functions, and statistical tests over alternatives.
3. Depth of Spontaneous Domain Mastery: Evaluates the researcher's ability to think critically in real time without written scripts when challenged with edge cases.
4. Formative Peer Review & Revisions: Provides an expert academic forum where external examiners identify technical omissions or required manuscript corrections prior to final archiving.

---

D. Suggest an appropriate research report structure for an AI crop disease detection study and justify each major section. (5 Marks)

1. Recommended Report Structure (2.5 Marks)

• Front Matter: Title, Abstract, Keywords, Table of Contents, List of Figures/Tables.
• Section 1: Introduction: Agricultural disease context, economic impact, motivation for UAV/edge imaging, and formal objectives.
• Section 2: Literature Review: Prior work in computer vision for agronomy, CNN vs. ViT architectures, and identified research gaps.
• Section 3: Dataset & Methodology: Image collection, preprocessing/data augmentation, model architecture, loss functions, and training hyperparameters.
• Section 4: Experimental Results & Analysis: Quantitative performance metrics: Accuracy, sensitivity, specificity, F1-score, and confusion matrix.
• Section 5: Discussion & Field Constraints: Model performance under variable field lighting, background occlusion, edge inference latency, and limitations.
• Section 6: Conclusion & Future Scope: Final technical summary, hardware deployment path, and references.

2. Justification of Key Sections (2.5 Marks)

• Methodology Section: Essential for reproducibility-ensures other agronomists and AI researchers can replicate the image preprocessing and network architecture.
• Results (Confusion Matrix / F1-Score): Class imbalance is common in crop disease imagery; reporting sensitivity/specificity proves the system avoids catastrophic false negatives in field diagnoses.
• Discussion & Limitations: Bridges the gap between idealized laboratory test benches and real-world farm deployment conditions.