name	literature-review-methodology
description	Systematic search strategy, inclusion/exclusion criteria, evidence synthesis, gap identification. Use when searching for papers, summarizing related work, or structuring a literature review.

Literature Review Methodology

When conducting literature reviews, summarizing related work, or structuring research gaps, use a systematic rather than ad-hoc approach.

1. Defining the Search Strategy

Formulate specific search strings (e.g., "federated learning" AND ("privacy" OR "differential privacy")).
Use high-quality academic indexes (e.g., ACM Digital Library, IEEE Xplore, Google Scholar, DBLP).
Track the venues (e.g., SIGMOD, VLDB, OSDI, NeurIPS) to prioritize high-impact peer-reviewed work.

2. Inclusion & Exclusion Criteria

Document precisely why papers are included or excluded stringently.

Inclusion: Papers published in the last 10 years, focusing on automated tuning, published in Tier-1 databases or ML venues.
Exclusion: Papers focused solely on tangential sub-problems, non-peer-reviewed preprints (unless highly cited).

3. The Extraction Process

Do not read papers blindly from start to finish. Scan deliberately:

Title & Abstract: Determine relevance.
Introduction & Conclusion: Understand the core claim and takeaway.
Figures & Evaluation: How do they prove their claim? What baselines did they use?
Methodology: Read thoroughly only if the paper is highly relevant.

Create a synthesis matrix (e.g., in a spreadsheet or markdown table) tracking: [Paper Citation] | [Target Domain] | [Core Technique] | [Key Limitation / Gap].

4. Writing the Synthesis

DO NOT write a chronological or paper-by-paper list ("Alice did X. Bob did Y. Charlie did Z.").
DO write a thematic synthesis. Group literature by approach or problem angle.
- Example: "While centralized approaches to model training [Alice, Bob] achieve strong accuracy, they require aggregating sensitive data. In contrast, federated methods [Charlie, Dave] offer better privacy guarantees but struggle with heterogeneous data distributions..."
Identify the Gap: The entire purpose of the related work section is to highlight the specific gap that your work fills. Conclude the section by explicitly contrasting your approach with the limitations of the reviewed literature.

Literature Review Methodology

When conducting literature reviews, summarizing related work, or structuring research gaps, use a systematic rather than ad-hoc approach.

1. Defining the Search Strategy

Formulate specific search strings (e.g., "federated learning" AND ("privacy" OR "differential privacy")).

Use high-quality academic indexes (e.g., ACM Digital Library, IEEE Xplore, Google Scholar, DBLP).

Track the venues (e.g., SIGMOD, VLDB, OSDI, NeurIPS) to prioritize high-impact peer-reviewed work.

2. Inclusion & Exclusion Criteria

Document precisely why papers are included or excluded stringently.

Inclusion: Papers published in the last 10 years, focusing on automated tuning, published in Tier-1 databases or ML venues.

Exclusion: Papers focused solely on tangential sub-problems, non-peer-reviewed preprints (unless highly cited).

3. The Extraction Process

Do not read papers blindly from start to finish. Scan deliberately:

Title & Abstract: Determine relevance.

Introduction & Conclusion: Understand the core claim and takeaway.

Figures & Evaluation: How do they prove their claim? What baselines did they use?

Methodology: Read thoroughly only if the paper is highly relevant.

Create a synthesis matrix (e.g., in a spreadsheet or markdown table) tracking: [Paper Citation] | [Target Domain] | [Core Technique] | [Key Limitation / Gap].

4. Writing the Synthesis

DO NOT write a chronological or paper-by-paper list ("Alice did X. Bob did Y. Charlie did Z.").

DO write a thematic synthesis. Group literature by approach or problem angle.

Example: "While centralized approaches to model training [Alice, Bob] achieve strong accuracy, they require aggregating sensitive data. In contrast, federated methods [Charlie, Dave] offer better privacy guarantees but struggle with heterogeneous data distributions..."

Identify the Gap: The entire purpose of the related work section is to highlight the specific gap that your work fills. Conclude the section by explicitly contrasting your approach with the limitations of the reviewed literature.

literature-review-methodology