Aparture

Appearance

arXiv Categories ​

Understanding arXiv's taxonomy and how to select relevant categories.

What are arXiv Categories? ​

arXiv organizes papers into a hierarchical taxonomy with:

  • Major categories - Top-level domains (cs, math, physics, etc.)
  • Subcategories - Specialized topics within domains
  • Cross-listings - Papers appearing in multiple categories

Papers can belong to multiple subcategories, with one designated as primary.

Category Structure ​

Computer Science (cs) ​

Artificial Intelligence

  • cs.AI - Artificial Intelligence
  • cs.LG - Machine Learning
  • cs.NE - Neural and Evolutionary Computing
  • cs.CV - Computer Vision and Pattern Recognition
  • cs.CL - Computation and Language (NLP)

Theory

  • cs.CC - Computational Complexity
  • cs.DS - Data Structures and Algorithms
  • cs.LO - Logic in Computer Science

Systems

  • cs.DC - Distributed, Parallel, and Cluster Computing
  • cs.OS - Operating Systems
  • cs.PL - Programming Languages
  • cs.SE - Software Engineering

Other

  • cs.CR - Cryptography and Security
  • cs.DB - Databases
  • cs.HC - Human-Computer Interaction
  • cs.IR - Information Retrieval
  • cs.RO - Robotics

Full list includes 40+ cs subcategories.

Mathematics (math) ​

Pure Mathematics

  • math.AG - Algebraic Geometry
  • math.NT - Number Theory
  • math.GT - Geometric Topology
  • math.RT - Representation Theory

Applied Mathematics

  • math.NA - Numerical Analysis
  • math.OC - Optimization and Control
  • math.PR - Probability
  • math.ST - Statistics Theory

Interdisciplinary

  • math.CO - Combinatorics
  • math.DS - Dynamical Systems
  • math.MP - Mathematical Physics

Statistics (stat) ​

  • stat.ML - Machine Learning
  • stat.ME - Methodology
  • stat.AP - Applications
  • stat.CO - Computation
  • stat.TH - Theory

Physics (physics, astro-ph, etc.) ​

Astrophysics (astro-ph)

  • astro-ph.CO - Cosmology and Nongalactic Astrophysics
  • astro-ph.GA - Astrophysics of Galaxies
  • astro-ph.SR - Solar and Stellar Astrophysics
  • astro-ph.EP - Earth and Planetary Astrophysics
  • astro-ph.HE - High Energy Astrophysical Phenomena
  • astro-ph.IM - Instrumentation and Methods

General Physics (physics)

  • physics.comp-ph - Computational Physics
  • physics.data-an - Data Analysis, Statistics and Probability
  • physics.optics - Optics
  • physics.plasm-ph - Plasma Physics

High Energy Physics (hep-)

  • hep-th - High Energy Physics - Theory
  • hep-ph - High Energy Physics - Phenomenology
  • hep-ex - High Energy Physics - Experiment

Quantitative Biology (q-bio) ​

  • q-bio.BM - Biomolecules
  • q-bio.GN - Genomics
  • q-bio.NC - Neurons and Cognition
  • q-bio.QM - Quantitative Methods

Economics (econ) ​

  • econ.EM - Econometrics
  • econ.TH - Theoretical Economics

Electrical Engineering (eess) ​

  • eess.IV - Image and Video Processing
  • eess.SP - Signal Processing
  • eess.SY - Systems and Control

Full taxonomy

See arxiv.org/category_taxonomy for the complete list of 150+ categories.

Choosing Categories ​

Strategy 1: Narrow & Focused ​

When to use:

  • Clear, specific research interests
  • Limited time for daily reading
  • Want high signal-to-noise ratio

Example: Deep learning in astronomy

Selected categories:
- cs.LG (Machine Learning)
- astro-ph.GA (Astrophysics of Galaxies)
- astro-ph.CO (Cosmology)

Typical volume: 5-15 papers/day Pros: Highly relevant results Cons: May miss related work

Strategy 2: Broad & Exploratory ​

When to use:

  • Interdisciplinary research
  • Want to discover new connections
  • Have time for exploration

Example: Bayesian methods across fields

Selected categories:
- stat.ME (Statistics Methodology)
- stat.ML (Machine Learning)
- astro-ph.IM (Instrumentation)
- physics.data-an (Data Analysis)
- econ.EM (Econometrics)

Typical volume: 20-50 papers/day Pros: Discover unexpected connections Cons: More noise, higher costs

Strategy 3: Domain-Focused ​

When to use:

  • Deep expertise in one field
  • Want comprehensive coverage
  • Daily reading is core activity

Example: Complete astrophysics coverage

Selected categories:
- All astro-ph.* subcategories (6 total)
- physics.comp-ph
- stat.AP

Typical volume: 30-80 papers/day Pros: Miss nothing in your domain Cons: High volume, expensive

Cross-Listing Considerations ​

Papers often appear in multiple categories:

Example paper:

  • Primary: cs.LG
  • Cross-listed: stat.ML, cs.AI

Aparture handling:

  • Fetches unique papers (no duplicates)
  • Considers all selected categories
  • Shows primary category in results

Implications:

  • Selecting both cs.LG and stat.ML won't duplicate papers
  • But you'll get papers where either is primary
  • More categories = better coverage

Category Selection Tips ​

Start Small ​

Begin with 2-3 highly relevant categories:

  1. Test with dry run - See paper volume
  2. Run minimal analysis - Check relevance
  3. Add categories gradually - Expand as needed

Monitor Volume ​

Track papers per category over time:

  • Low volume (<5/day) - Consider adding related categories
  • High volume (>50/day) - Consider enabling Quick Filter
  • Overwhelming (>100/day) - Narrow selection or increase thresholds

Check Cross-Listings ​

Some papers appear in multiple categories you care about:

Example workflow:

  1. Run analysis with cs.LG only
  2. Check where top papers are cross-listed
  3. Add those categories for better coverage

Use Research Criteria ​

Category selection + research criteria = powerful combination:

Broad categories:

Categories: cs.AI, cs.LG, cs.CV, cs.CL
Research criteria: "Bayesian deep learning"

Narrow categories:

Categories: cs.LG, stat.ML
Research criteria: "Any machine learning advances"

Common Combinations ​

Machine Learning Researcher ​

Core:
- cs.LG (Machine Learning)
- cs.AI (Artificial Intelligence)
- stat.ML (Statistics - Machine Learning)

Optional:
- cs.CV (Computer Vision)
- cs.CL (NLP)

Volume: 20-40 papers/day

Computational Astrophysicist ​

Core:
- astro-ph.CO (Cosmology)
- astro-ph.GA (Galaxies)
- astro-ph.IM (Instrumentation)

Optional:
- physics.comp-ph (Computational Physics)
- physics.data-an (Data Analysis)
- stat.AP (Statistics Applications)

Volume: 15-35 papers/day

Applied Statistician ​

Core:
- stat.ME (Methodology)
- stat.AP (Applications)
- stat.ML (Machine Learning)

Optional:
- cs.LG (Machine Learning)
- econ.EM (Econometrics)

Volume: 10-25 papers/day

Theory-Focused CS ​

Core:
- cs.CC (Computational Complexity)
- cs.DS (Data Structures & Algorithms)
- cs.LO (Logic)

Optional:
- math.CO (Combinatorics)
- cs.CR (Cryptography)

Volume: 5-15 papers/day

Category Updates ​

arXiv occasionally updates its taxonomy:

Recent changes:

  • 2020: Added eess.* (Electrical Engineering)
  • 2017: Reorganized cs.* subcategories
  • 2016: Added econ.* (Economics)

Staying current:

  • Check arXiv taxonomy annually
  • Watch for new subcategories in your field
  • Update Aparture configuration as needed

New subcategory?

If you notice a new relevant subcategory, add it to your configuration and re-run historical analyses to catch papers you missed.

API Considerations ​

Fetching Limits ​

arXiv API has rate limits:

  • 1 request per 3 seconds
  • Maximum 1000 results per request

Aparture handling:

  • Automatically respects rate limits
  • Batches requests for multiple categories
  • Paginates large result sets

Performance ​

More categories = longer fetching:

  • 1-3 categories: ~1 minute
  • 5-10 categories: ~2-3 minutes
  • 20+ categories: ~5-10 minutes

Next Steps ​

Released under the MIT License.

Copyright © 2025 Josh Speagle. Created in collaboration with Claude.