Natural language processing

Publishing authors

Base scenario: 4.5k publishing authors. Paper count × 3.07 known authors per paper ÷ 3 papers per author per year. The range assumes 1–5 papers per author; the base and both ends of the range are at least the 2,912 distinct Author IDs already observed in the sample.

The publication rate is an assumption, not a measured rate for this field. The range is a scenario, not a confidence interval.

Random sample: 1,000 of 4,357 papers (23.0%). 12.3% of authorship records lack an Author ID; 10.1% of retrieved works lack an abstract. Retrieved 2026-09-26.

Mean known authors per paper: 3.07; 95% bootstrap interval 2.89–3.27. This describes sampling variability, not search accuracy or uncertainty in the number of researchers.

Applied coauthors are included. Missing or incorrectly linked author records affect these figures. They do not measure research jobs or everyone working in a field.

Private investment

An assumed allocation of broader investment segments. The range is 0.5–2× the base case. Shares are editorial assumptions, not measured deals or research spending.

Base case: $1.36B.

AI infrastructure/models/research/governance: $143.2B × 0.80%
Ed tech: $1.44B × 15.00%

Company investment, not revenue or research spending. Ranges are scenarios, not confidence intervals.

ACL 2026 — research topics

Scope and related research

Language-specific quality and translation remain distinct tasks alongside general pre-training.

Organizations

Selected examples

DeepL Product
Machine translation.
Google Cloud Tools & infrastructure
Cloud Translation offers neural and language-model translation.
Cohere Models
Rerank models rank relevant documents across languages.

More organizations (4)

Duolingo Product
AI language learning.
Meta Models
Llama 4 models support multilingual text understanding and generation.
Microsoft Tools & infrastructure
Azure Translator provides neural and language-model text translation.
Unbabel Product
Translation AI.

Information extraction and knowledge graphs

Entities, relations, events, knowledge graphs, ontologies, and neuro-symbolic knowledge representation.

Knowledge graphs Entity extraction

5,423 papers

1.9× vs. 2023

Authors · scenario: 4.2k–21k
Investment · scenario: $1.64B–$6.57B

Sources and methodology

Publication activity

2019: 868
2020: 1,278
2021: 1,526
2022: 1,978
2023: 2,793
2024: 3,915
2025: 5,423

Matches include papers applying these methods. Search precision and recall have not been systematically measured.

Title and abstract matches for articles, preprints and reviews; retracted work is excluded. Versions and overlapping areas may be counted more than once.

("information extraction" OR "named entity recognition" OR "relation extraction" OR "knowledge graph" OR "knowledge representation" OR "neurosymbolic") AND (("machine learning" OR "deep learning" OR "neural network" OR "artificial intelligence") OR ("language model" OR "large language model" OR "vision language model") OR "natural language processing")

Retrieved 2026-09-26.

Publishing authors

Base scenario: 6.9k publishing authors. Paper count × 3.84 known authors per paper ÷ 3 papers per author per year. The range assumes 1–5 papers per author; the base and both ends of the range are at least the 3,724 distinct Author IDs already observed in the sample.

The publication rate is an assumption, not a measured rate for this field. The range is a scenario, not a confidence interval.

Random sample: 1,000 of 5,423 papers (18.4%). 13.1% of authorship records lack an Author ID; 17.9% of retrieved works lack an abstract. Retrieved 2026-09-26.

Mean known authors per paper: 3.84; 95% bootstrap interval 3.60–4.09. This describes sampling variability, not search accuracy or uncertainty in the number of researchers.

Applied coauthors are included. Missing or incorrectly linked author records affect these figures. They do not measure research jobs or everyone working in a field.

Private investment

An assumed allocation of broader investment segments. The range is 0.5–2× the base case. Shares are editorial assumptions, not measured deals or research spending.

Base case: $3.28B.

Data management, processing: $31.6B × 2.00%
Pharmaceutical: $10.6B × 10.00%
Accounting/finance: $2.59B × 25.00%
Legal tech: $3.79B × 25.00%

Company investment, not revenue or research spending. Ranges are scenarios, not confidence intervals.

ACL 2026 — research topics AAAI 2026 — AI taxonomy KDD 2026 — research topics

Scope and related research

Focuses on extracting and organizing knowledge. Learning on graphs is covered under graph ML.

Organizations

Selected examples

ABBYY Product
Document AI.
Hyperscience Product
Document processing.
Instabase Product
Enterprise document AI.

More organizations (4)

Google Cloud Tools & infrastructure
Document AI extracts structured fields and entities from documents.
Kensho Product
Financial AI.
Neo4j Product
Graph data science and knowledge graphs.
Rossum Product
Transactional document AI.

Dialogue systems

Conversation management, grounded dialogue, question answering, and language interfaces.

Conversational AI Dialogue systems

2,586 papers

1.9× vs. 2023

Authors · scenario: 3.0k–8.2k
Investment · scenario: $1.08B–$4.32B

Sources and methodology

Publication activity

2019: 569
2020: 775
2021: 859
2022: 978
2023: 1,375
2024: 1,569
2025: 2,586

Matches include papers applying these methods. Search precision and recall have not been systematically measured.

Title and abstract matches for articles, preprints and reviews; retracted work is excluded. Versions and overlapping areas may be counted more than once.

("dialogue system" OR "dialog system" OR "conversational AI" OR "conversational agent" OR "task oriented dialogue" OR "conversational question answering")

Retrieved 2026-09-26.

Publishing authors

Base scenario: 3.0k publishing authors. Paper count × 3.19 known authors per paper ÷ 3 papers per author per year. The range assumes 1–5 papers per author; the base and both ends of the range are at least the 2,978 distinct Author IDs already observed in the sample.

The publication rate is an assumption, not a measured rate for this field. The range is a scenario, not a confidence interval.

Random sample: 1,000 of 2,586 papers (38.7%). 14.7% of authorship records lack an Author ID; 12.1% of retrieved works lack an abstract. Retrieved 2026-09-26.

Mean known authors per paper: 3.19; 95% bootstrap interval 2.97–3.40. This describes sampling variability, not search accuracy or uncertainty in the number of researchers.

Applied coauthors are included. Missing or incorrectly linked author records affect these figures. They do not measure research jobs or everyone working in a field.

Private investment

An assumed allocation of broader investment segments. The range is 0.5–2× the base case. Shares are editorial assumptions, not measured deals or research spending.

Base case: $2.16B.

AI agents: $8.02B × 10.00%
Retail: $4.10B × 10.00%
Medical and healthcare: $11.8B × 5.00%
Ed tech: $1.44B × 25.00%

Company investment, not revenue or research spending. Ranges are scenarios, not confidence intervals.

ACL 2026 — research topics

Scope and related research

Agent research covers action execution; human–AI interaction covers how people use these systems.

Organizations

Selected examples

Sierra Product
Customer-service agents.
Intercom Product
AI customer support.
Convai Product
Embodied conversational characters.

More organizations (7)

Anthropic Product
Claude provides conversational assistance for analysis and complex tasks.
Decagon Product
Customer-service agents.
ElevenLabs Product
Speech and voice agents.
Google DeepMind Models
Gemini models support multimodal dialogue and question answering.
OpenAI Product
ChatGPT supports contextual conversations and question answering.
Replika Product
Conversational companions.
Speak Product
AI language tutor.

Text understanding and generation

Semantics, syntax, discourse, classification, sentiment, summarization, and task-specific text generation.

NLP Text generation Summarization

11,797 papers

1.6× vs. 2023

Authors · scenario: 6.6k–33k
Investment · scenario: $1.02B–$4.10B

Sources and methodology

Publication activity

2019: 3,369
2020: 4,518
2021: 5,209
2022: 5,650
2023: 7,383
2024: 9,293
2025: 11,797

Matches include papers applying these methods. Search precision and recall have not been systematically measured.

Title and abstract matches for articles, preprints and reviews; retracted work is excluded. Versions and overlapping areas may be counted more than once.

("natural language understanding" OR "text classification" OR "sentiment analysis" OR "text summarization" OR "semantic parsing" OR "discourse parsing" OR "natural language generation")

Retrieved 2026-09-26.

Publishing authors

Base scenario: 11k publishing authors. Paper count × 2.80 known authors per paper ÷ 3 papers per author per year. The range assumes 1–5 papers per author; the base and both ends of the range are at least the 2,757 distinct Author IDs already observed in the sample.

The publication rate is an assumption, not a measured rate for this field. The range is a scenario, not a confidence interval.

Random sample: 1,000 of 11,797 papers (8.5%). 12.0% of authorship records lack an Author ID; 15.7% of retrieved works lack an abstract. Retrieved 2026-09-26.

Mean known authors per paper: 2.80; 95% bootstrap interval 2.68–2.92. This describes sampling variability, not search accuracy or uncertainty in the number of researchers.

Applied coauthors are included. Missing or incorrectly linked author records affect these figures. They do not measure research jobs or everyone working in a field.

Private investment

An assumed allocation of broader investment segments. The range is 0.5–2× the base case. Shares are editorial assumptions, not measured deals or research spending.

Base case: $2.05B.

AI infrastructure/models/research/governance: $143.2B × 0.80%
Legal tech: $3.79B × 20.00%
Ed tech: $1.44B × 10.00%

Company investment, not revenue or research spending. Ranges are scenarios, not confidence intervals.