phd-research-companion

PhD Research Companion v1.5.0

Professional full-stack research management skill for Computer Science PhD students, providing complete automated support from project initialization to journal submission with scientific traceability compliance.

🎯 Overview & Purpose

This comprehensive skill transforms the fragmented, manual process of academic research into a streamlined, trackable workflow that ensures every step meets publication standards and maintains an audit trail for reproducibility and scientific integrity.

What It Solves

🏗️ Architecture Overview

phd-research-companion/
├── init_research_project.py         # Entry point - creates full research environment
├── run                              # Quick CLI wrapper for all commands  
├── scripts/                         # Core analysis & generation tools
│   ├── multi_source_search.py      # Literature collection (arXiv, SemanticScholar, DBLP)
│   ├── paper_analyzer.py           # Deep extraction of contributions/methodology  
│   ├── create_experiment_design.py # Comparison/ablation/robustness YAML configs
│   ├── generate_latex_template.py  # IEEE/ACM/NeurIPS templates with proper structure
│   ├── revision_tracker.py         # Track improvement rounds systematically
│   ├── verify_math_notation.py     # Mathematical proof consistency validator  
│   └── check_compliance.py         # Final submission readiness checker
├── references/                      # Best practices & documentation
└── SKILL.md                        # This comprehensive guide

🚀 Quick Start Guide

Installation & Setup

# Clone or copy skill to workspace
cd /home/user/workspace/skills/phd-research-companion
Make run script executable (one-time setup)  
chmod +x run
Verify installation
./run --version

Initialize Your Research Project

# Method 1: Interactive wrapper (recommended)  
./run init -d "machine unlearning with certified forgetting guarantees" \
            -j "IEEE TIFS" \
            -o ./my-research-project-2024
Method 2: Direct Python execution
python3 init_research_project.py --domain "Your Research Topic"

Output created:

research-project-2024/
├── 00-dashboard/                    # Project overview & tracking
├── 01-literature-survey/           # BibTeX, PDFs, analysis outputs
├── 02-methodology-dev/             # Theorems, formal proofs  
├── 03-paper-drafting/              # LaTeX templates, drafts  
├── 04-experiments/                 # Designs (YAML), results archive  
├── 05-revision-rounds/             # Systematic improvement tracking  
├── 06-collaboration/               # Advisor feedback, peer reviews  
└── 07-audit-trail/                 # Scientific traceability evidence

📚 Complete Module Reference

1️⃣ Multi-Source Literature Search (scripts/multi_source_search.py)

Automatically collect papers from arXiv, Semantic Scholar, DBLP with deduplication and citation export.

#### Basic Usage

# Quick search (foreground - instant feedback)  
./run search -q "machine unlearning differential privacy" -l 30
Background execution for large searches 
./run search -q "federated learning security" \
             --sources arxiv semanticscholar \
             -l 50 \
             -o ./my-project/01-literature-survey \
             --background

#### Advanced Filters

# Temporal filtering with keyword constraints  
./run search \
    -q "adversarial robustness certified defenses" \
    --from-year 2020 --to-year 2024 \
    --sources arxiv,ieee\dblp \
    -l 75
Output: search-results-20240310.bibtex + search-summary.md

#### Background Monitoring

# Terminal 1: Start background task  
./run search -q "topic" --background -o ./results &
Terminal 2: Monitor progress (real-time)
watch -n 5 'cat results/search-progress-search.json'
Check completed status after finish
cat results/search-progress-search.json | grep '"summary"'

Files Generated:

2️⃣ Paper Analysis & Deep Extraction (scripts/paper_analyzer.py)

Extract key contributions, methodology components, and mathematical formalisms from downloaded papers.

#### Modes Available

Deep Mode (10-30 minutes for batch of 50+ papers)

./run analyze -i ./my-project/01-literature-survey/*.pdf --mode deep --background
Output per paper: analysis-{filename}.md with sections:
  - Key innovations extracted  
  - Methodology components mapped
  - Mathematical definitions identified
  - Limitations noted

Quick Mode (2-5 minutes for fast overview)

./run analyze -i ./papers/*.pdf --mode quick
Fast metadata extraction: title, authors, venue, year only

#### Batch Analysis with Progress Tracking

# Start background analysis of 100 papers
./run analyze \
    -i "./literature/*.pdf" \
    --mode deep \
    -o ./analysis-output \
    --background &
Monitor in another terminal:
while [ -f ./analysis-output/analysis-progress-analysis.json ]; do 
    sleep 10
    cat ./analysis-output/analysis-progress-analysis.json | jq '.{progress_percent,total_papers,stage}'
done

Comparison Report Generated:

3️⃣ Experimental Design Generation (scripts/create_experiment_design.py)

Create reproducible experiment specifications in YAML format covering three key categories required for top-tier publications.

#### A) Baseline Comparison Experiments

./run experiment \
    --type comparison \
    --datasets "CIFAR-10,Fashion-MNIST,CelebA" \
    --baselines "Retraining,SISA,NAU,Certificate-based,MF-GAN" \
    --metrics "test_accuracy,fps,latency_ms,gdpa_certificates" \
    -o ./my-project/04-experiments/design-baseline
Output: experiment-comparison-design.yaml + report.md with:
  - Dataset specifications (split sizes, class distributions)  
  - Baseline paper citations and implementation references
  - Evaluation metrics with formulas
  - Expected compute time & GPU requirements

#### B) Ablation Studies ``bash ./run experiment \ --type ablation \ --components "privacy_layer,adversarial_training,noising_mechanism" \ --base_model "ResNet-18" \ -o ./experiments/ablation-studies

Documents: What happens when each component is removed?

#### C) Robustness Stress Tests  

4️⃣ LaTeX Template Generation (scripts/generate_latex_template.py)
Generate conference/journal-ready templates with proper formatting for IEEE, ACM, NeurIPS, ICLR.
#### IEEE TIFS (Transactions)

#### ACM Transactions on Information Systems (TISSEC)  

#### NeurIPS Conference Format  

Key Features:
✅ Proper bibliography support (biblatex with IEEEtran/ACM styles)
✅ Figure placement guidelines ([htbp] with positioning notes)

✅ Abstract, introduction, conclusion structure scaffolding
✅ Theorem/enumerate environments configured for proofs
✅ References section with placeholder citations ready
5️⃣ Revision Round Tracking (scripts/revision_tracker.py)
Systematically document every improvement round (6-8 cycles typical) before final submission.
#### Add Revision Round Entry

#### Generate Revision Timeline Report

#### Track Specific Issue Resolution

6️⃣ Mathematical Notation Verification (scripts/verify_math_notation.py)
Automate detection of undefined symbols, inconsistent notation, or missing proofs.
#### Full Scan  

#### Specific Checks

7️⃣ Final Compliance Check (scripts/check_compliance.py)
Run systematic verification before advisor review or journal submission.
#### Full Audit  

#### Generate Submission Readiness Report

🛠️ Automation & Integration
Daily Literature Watch Updates
Set up cron job for continuous domain monitoring:

Bash Automation Wrapper Example

🔍 Troubleshooting
Literature Search Issues

LaTeX Compilation Errors

📊 Status & Maintenance Information
Version: 1.5.0 (March 2026)  
Tested With: Python 3.8+, IEEE LaTeX template v2.4, arXiv API v2  
Supported Venues: IEEE TIFS/TIP/TKDE, ACM TISSEC/CSUR, NeurIPS/ICLR/AAAI  
Update Check