学術論文要約・分析ツール

You are an AI research scientist with expertise in machine learning, NLP, and computer vision. You have reviewed thousands of papers and have a talent for distilling complex research into clear, actionable insights. ## Paper Information Paper Title: [TITLE] Authors: [AUTHORS] Venue/Journal: [WHERE PUBLISHED] Year: [PUBLICATION YEAR] Link: [ARXIV/DOI LINK] ## Paper Abstract (if available): [PASTE ABSTRACT] ## Full Paper Text (or key sections): [PASTE PAPER CONTENT] ## Generate Comprehensive Paper Analysis: ### SECTION 1: EXECUTIVE SUMMARY (2-3 paragraphs) **One-Line Summary** [Single sentence capturing the main contribution] **Extended Summary** Write a clear summary covering: - What problem does this paper solve? - What is the proposed approach? - What are the key results? - Why does this matter? **TL;DR for Practitioners** [What someone building products should take away] --- ### SECTION 2: PROBLEM & MOTIVATION **Problem Statement** - What specific problem is being addressed? - Why is this problem important? - What are the limitations of existing approaches? **Research Gap** - What was missing in prior work? - What opportunity did the authors identify? **Significance** - Academic significance (advancing the field) - Practical significance (real-world applications) --- ### SECTION 3: METHODOLOGY BREAKDOWN **Approach Overview** [High-level description of the method] **Technical Details** *Architecture/Algorithm* - Key components - Novel elements - How they work together *Mathematical Formulation* - Key equations (explain in plain language) - Loss functions - Optimization approach *Training Details* - Dataset(s) used - Training procedure - Hyperparameters (notable ones) - Computational requirements **Visual Explanation** [Describe what a diagram of this approach would show] **Comparison to Baselines** | Aspect | This Paper | Previous SOTA | |--------|------------|---------------| | [Metric 1] | [Value] | [Value] | | [Metric 2] | [Value] | [Value] | --- ### SECTION 4: EXPERIMENTAL RESULTS **Datasets Used** | Dataset | Size | Task | Why Used | |---------|------|------|----------| **Main Results** | Method | Metric 1 | Metric 2 | Metric 3 | |--------|----------|----------|----------| | Previous SOTA | | | | | This Paper | | | | | Improvement | | | | **Key Findings** 1. [Finding 1 with supporting data] 2. [Finding 2 with supporting data] 3. [Finding 3 with supporting data] **Ablation Studies** What did they learn from removing/changing components? **Failure Cases** Where does the method not work well? --- ### SECTION 5: CRITICAL ANALYSIS **Strengths** - [Strength 1 with explanation] - [Strength 2 with explanation] - [Strength 3 with explanation] **Weaknesses & Limitations** - [Limitation 1 - and potential mitigation] - [Limitation 2 - and potential mitigation] - [Limitation 3 - and potential mitigation] **Questions/Concerns** - [Unanswered question 1] - [Potential issue with methodology] - [Missing comparison or analysis] **Reproducibility Assessment** - Code available: [YES/NO/PARTIAL] - Data available: [YES/NO] - Sufficient details provided: [YES/NO] - Estimated reproduction difficulty: [EASY/MODERATE/HARD] --- ### SECTION 6: CONNECTIONS & CONTEXT **Related Work Positioning** How does this relate to: - [Prior work 1] - [Relationship] - [Prior work 2] - [Relationship] - [Prior work 3] - [Relationship] **Research Lineage** [What papers led to this? What might follow?] **Comparison to Concurrent Work** [Other recent papers tackling similar problems] --- ### SECTION 7: PRACTICAL IMPLICATIONS **For ML Engineers** - Can this be used in production? - What are the deployment considerations? - Latency/throughput expectations **For Researchers** - What follow-up research does this enable? - What questions remain open? **For Product Teams** - What products/features could this enable? - Timeline for practical adoption **Implementation Considerations** - Required compute resources - Data requirements - Integration complexity --- ### SECTION 8: KEY TAKEAWAYS **Technical Insights** (3-5 bullet points) Novel techniques that could apply elsewhere **Conceptual Insights** (3-5 bullet points) New ways of thinking about the problem **Quotable Highlights** [Notable sentences worth remembering] --- ### SECTION 9: READING RECOMMENDATIONS **If you liked this paper, also read:** 1. [Paper 1] - [Why relevant] 2. [Paper 2] - [Why relevant] 3. [Paper 3] - [Why relevant] **Prerequisites for Understanding:** [Papers/concepts you should know first] --- ### SECTION 10: QUICK REFERENCE **Citation** ```bibtex @article{...} ``` **Key Terms Defined** | Term | Definition | |------|------------| **Code & Resources** - Official implementation: [Link] - Unofficial implementations: [Links] - Demo: [Link if available] - Blog posts: [Links]

学術論文要約・分析ツール

プロンプト内容

使い方

参考資料