- Massaging the task hierarchy categories is somewhat encouraged.
- If a subcategory (e.g.
Question Answering -> Contextual Question Answering -> Extractive
) is mentioned, don't mention its parent category (Question Answering -> Contextual Question Answering
) src/auto_add_domain.py
is helpful for adding domains for all tasks in any particular dataset
Abuse Detection
Question Answering
Question Answering -> Commonsense Question Answering
Question Answering -> Contextual Question Answering
Question Answering -> Contextual Question Answering -> Extractive
Question Answering -> Contextual Question Answering -> Abstractive
Question Answering -> Fill in the Blank
Question Answering -> Multiple Choice Question Answering
Question Answering -> Open Question Answering
Question Answering -> Incorrect Answer Generation
Question Generation
Question Generation -> Contextual Question Generation
: Generate questions based on given context e.g. a task to create a question based on a paragraph.Question Generation -> Question Composition
: Compose questions by concating questions in the inputQuestion Generation -> Fill in the Blank
Author Identification
Classification
Classification -> Verification
: Verify whether a given descriptive attribute applies to a given text or not (binary output) e.g. if the paragraph contains offensive content or notClassification -> Verification -> Claim Verification
Classification -> Verification -> Sufficient Information Verification
: Verify whether a text contains sufficient information to answer a questionClassification -> Verification -> Grammar Verification
: Verify whether a text is grammaticalClassification -> Verification -> Relevance Verification
Classification -> Verification -> Answer Verification
: Verify whether a text answers the question
Combinatorics
Command Execution
Coreference
Coreference -> Entity Coreference
Coreference -> Pronoun Disambiguation
Counting
: Count an attribute of input e.g. a task to count number of vowels in a given wordDialogue Understanding
Document Understanding
Entity Detection
Ethical Judgement
Explanation Generation
Fake News Detection
Grammar Error
Grammar Error -> Grammar Error Correction
Grammar Error -> Grammar Error Detection
Hallucination
: Given a context, generate imaginary content e.g. given a sentence, generate a story/poem.Hate Speech Detection
Hypernym Discovery
Intent Detection
Language Identification
Mathematics
Mathematics -> Algebra
Mathematics -> Arithmetic
Mathematics -> Geometry
Named Entity Recognition
Order Generation
: Given a set of elements, find their order (e.g. monotonically increasing/decreasing numbers, increasing/decreasing size in case of objects)Paragraph Generation
Paraphrasing
Parts-of-speech
Question Decomposition
Reasoning
Reasoning -> Abductive Reasoning
Reasoning -> Analogical Reasoning
Reasoning -> Causal Reasoning
Reasoning -> Commonsense Reasoning
: Tasks related to activities humans do in daily life e.g. eating breakfast in the morning, sleeping during night etc.Reasoning -> Deductive Reasoning
Reasoning -> Logical Reasoning
Reasoning -> Multihop Reasoning
Reasoning -> Numerical Commonsense Reasoning
: Tasks which requires numerical commonsense knowledge e.g. a car has 4 wheels.Reasoning -> Numerical Reasoning
Reasoning -> Physical Reasoning
: Tasks involving physical interactions with objects e.g. a knife (and not a paper) is used to cut objectsReasoning -> Planning
: Tasks which need some sort of planning e.g. how to go to Hawaii?Reasoning -> Qualitative Reasoning
Reasoning -> Reasoning on Actions
Reasoning -> Reasoning on Social Interactions
Reasoning -> Reasoning with Symbols
: Tasks where symbols represent various things e.g. if X is the number of apples in the freeze today morning and Y is the number remaining after I ate a few apples, X-Y is the number of apples I ate.Reasoning -> Spatial Reasoning
Reasoning -> Temporal Reasoning
Relation Prediction
Relevancy Estimation
Review Generation
Role Labelling
Semantic Parsing
Sentence Generation
Sentiment Analysis
Sorting
Stance Detection
Structured Text Generation
: Generate structured text in the output e.g. a task that converts questions in natural language to SQL queriesStyle Transfer
Summarization
Tabular Text Operation
Tabular Text Operation -> Column Matching
: Given two sets in the input, generate a mapping between them e.g. given a set of countries and their capitals in the input, generate an output that maps countries to capitals.Tabular Text Operation -> Question Answering
Text Comparison
Text Modification
Text Simplification
Text Span Selection
Textual Entailment
Title Generation
Topic Generation
Translation
Weblink Generation
Word Sense Disambiguation
Accounting
Anthropology
Architecture
Art
Astronomy
Biology
Biology -> Anatomy
Biology -> Clinical Knowledge
Biology -> Human Biology
Biology -> Virology
Business Ethics
Chemistry
Computer Science
Computer Science -> Machine Learning
Computer Security
Dialogue
Econometrics
Electrical Engineering
Fiction
Formal Fallacy
Formal logic
Geography
Global Facts
Government and Politics
Pop Culture
History
History -> European History
History -> 9/11 Reports
Human Sexuality
International Law
Jurisprudence
Justice
Law
Macroeconomics
Management
Marketing
Mathematics
Medical Genetics
Medicine
Microeconomics
Moral Scenarios
Movies
Music
News
Nutrition
Personal Narratives
Philosophy
Physics
Prehistory
Psychology
Public Relations
School Science Textbooks
Natural Science
Security: Environmental Security
Security: National Security
Social Media
Sociology
Sports
Sports -> NFL
Statistics
US Foreign Policy
Wikipedia
World Religions
Commonsense
Commonsense -> Social Commonsense
: a situation involving two same gender people with contrasting attributes, emotions, social roles, etc.Commonsense -> Physical Commonsense
: a context involving two physical objects with contrasting properties, usage, locations, etc.
- https://en.wikipedia.org/wiki/List_of_ISO_639-1_codes: ISO language name column