VIRTUAL ARENA AI

Academic Production (Tech/AI Subset)

Technical subset of indexed academic production: concentration areas and science→market gap.

Tech Only2026-06-13 00:25 UTC (1m ago)
62,486papers in subset·35,067IA·2,832Brazil
📄
Total Papers
62,486
arXiv + other sources
7 sources
🔬
AI Papers
35,067
Vision + ML + AI + NLP
56%
🇧🇷
Brazil
2,832
BR authors · tech only
4.5%
📚
Last 90 Days
43,184
recent papers
+246.7%

Publication Trend

Comparison between last 90 days and the previous 90 days

246.7%
Last 90 days
43,184
Previous 90 days
12,455

Data Sources

74%
21%
SourceVolume%Pipeline RoleCoverage
arXiv
46,35974.2%Global preprint base●●●●○
OpenAlex
13,00920.8%Multi-disciplinary enrichment●●●○○
OpenAlex University
1,7092.7%Institutional repository●●○○○
OpenAlex Brazil
8021.3%Brazil coverage●●○○○
semantic-scholar
4220.7%●○○○○
openalex_dissertations_br
1750.3%●○○○○
BDTD
100.0%BR theses & dissertations●○○○○

Brazil Focus

Coverage by Source
Total Brazil · tech2,832
BR authors · tech/AI filter applied
OpenAlex Brazil802
indexed via OpenAlex
BDTD (teses/dissertações): 10 items — insufficient base for chart; partial integration underway.
Coverage Asymmetry

Of 2,832 Brazilian papers, only 802 are indexed via OpenAlex and 10 via BDTD. Real Brazilian coverage is partial and biased toward CS/Physics via arXiv.

⚠ Absence of indexation ≠ absence of production. Institutional repositories from 15 universities are not yet fully integrated.

Research Areas

Ranked by publication volume. Computer Vision leads, followed by Machine Learning. (count by primary arXiv category — differs from "Research vs Market" below which uses taxonomy_bridge and yields smaller volumes)

Computer Visioncs.CV
11,789
Machine Learningcs.LG
10,825
NLP & Computational Ling.cs.CL
6,370
Artificial Intelligencecs.AI
6,083
Roboticscs.RO
3,963
Computer Science (General)Computer science
2,959
Cryptography & Securitycs.CR
1,337
Statistical MLstat.ML
927
Other areas (12)
4,239

Papers by Institution

⚠ Institutional data is ~48 days old (OpenAlex). Paper counts are historical accumulations.

#InstitutionCWUR worldCWUR in countryCWUR researchOpenAlex citationsCitations per workPapers in sample%
1National University of Singapore#79#1#32
37910.8%
2Universidade de São Paulo#119#1#8222.3M48.2
35710.2%
3KU Leuven#98#1#52
2537.2%
4Centre National de la Recherche Scientifique#1083#27#1036
2527.2%
5Universidade Estadual de Campinas (UNICAMP)
2487.1%
6Delft University of Technology#268#10#221
2396.8%
7ETH Zurich#31#1#45
2376.8%
8Universidade Federal de Pernambuco#891#14#8502.2M24.8
2216.3%
9Universidade Federal de Minas Gerais#508#6#4845.8M41.2
1965.6%
10Universidade Federal de Santa Catarina#732#9#6973.1M29.2
1905.4%
11Nagoya University#127#5#233
1905.4%
12Technical University of Munich#77#5#78
1905.4%
13Universidade Federal Fluminense#1006#18#9641.6M20.7
1895.4%
14University of Oslo#88#1#101
1815.2%
15United Arab Emirates University#958#2#918
1805.1%

Research vs Market

Each row compares paper volume (research) with job volume (market) via taxonomy_bridge v0.1. Paper volumes here are lower than "Research Areas" totals above because taxonomy_bridge uses semantic topic matching — not the full arXiv category count.

Research (papers)
Market (jobs)
Full Stack
1.2K/42.0K
Machine Learning
34.6K/2.8K
GenAI / LLMs
24.5K/77
MLOps
18.2K/81
Frontend / HCI
1.1K/12.2K
Distributed Computing
987/6.3K
Cybersecurity
1.9K/4.0K
Data Engineering
990/2.7K
QA/Testing
1.2K/208
UX/UI Design
1.1K/140
Research→Market Gap

35,067 AI papers represent the largest concentration of academic production. The research→jobs conversion rate remains structurally low.

E1Robotics (cs.RO) is 5th in papers but barely appears in job listings.
E2Software Engineering has high job volume compared to papers.
CSGenAI/LLMs has high academic output (24.5K papers via taxonomy_bridge) but few explicit jobs (77), suggesting title matching underestimates actual demand.
Methodology & Limitations
Coverage

arXiv dominates with ~79%. Bias toward CS/Physics. Medicine, engineering and humanities underrepresented.

Taxonomy

taxonomy_bridge v0.1 with matching via arxiv_category and job_keyword (regex word-boundary). Initial coverage — narrow topics may have weak matching.

Main limitation

No citation/impact data for arXiv papers. Brazilian papers frequently without identified institution (NULL field).

Dominant source

arXiv (74%)

Most Frequent Terms

LLM (11,635)Computer science (11,285)Multimodal (4,573)Reinforcement Learning (4,359)Artificial intelligence (3,960)Diffusion Models (3,595)Engineering (1,431)Process (computing) (1,246)Work (physics) (1,087)Context (archaeology) (1,060)Medicine (1,005)Key (lock) (976)Machine learning (915)Physics (887)Mathematics (836)

Global academic competitiveness

Loading rankings…