Skip to content

ai_assistance

Shared AI assistance components used across different review stages

URI: https://open-and-sustainable.github.io/revaise-model/schema/objects/ai_assistance

Name: ai_assistance

Classes

Class Description
Affiliation Institutional affiliation
AgreementMetrics Metrics for measuring agreement between reviewers/extractors
AIAssistance Configuration for AI assistance in any review stage
AIParameter AI model parameter configuration
AIPrompt Prompt configuration for AI interaction
AISession Session details for AI-assisted work
Author Author of a review or publication
        Participant A person participating in the review process
ConfidenceInterval Confidence interval specification
DatasetRef External dataset reference
DateRange A simple date range representation
FieldAgreement Agreement metrics for a specific field or item
HumanModification Record of human modification to AI output
ParticipantWorkload Workload statistics for a participant
PerformanceMetrics General performance metrics
        AIPerformanceMetrics Performance metrics for AI assistance, extending general performance metrics ...
ProcessMetrics Metrics related to process efficiency and completion
QualityControl Quality control measures and checks
StageExecution Base class for executing a specific stage of the review
StageOutput Output file or log produced by a review stage (results, logs, code, visualiza...
StageOutputRef Reference to a stage output in this or another RevAIse bundle

Slots

Slot Description
accuracy Overall accuracy
address Street address of the institution
affiliations Institutional affiliations of the author
agreement_rate Agreement rate for this field
agreement_sample_size Number of items compared
ai_assistance_config AI assistance configuration for a review stage
ai_confidence_threshold Minimum confidence score for accepting AI outputs
ai_config_id Reference to the AIAssistance configuration used
ai_error_handling Strategy for handling AI errors or failures
ai_id Unique identifier for AI assistance configuration
ai_model Name of the AI model
ai_parameters Model parameters and settings
ai_performance_metrics Performance metrics for the AI assistance
ai_prompts Prompts used to interact with the AI
ai_provider AI service provider
ai_purpose Purposes for which AI is used
ai_session_id Unique identifier for AI session
ai_training_description Description of training data or fine-tuning
ai_validation_rules Rules for validating AI outputs
ai_version Version of the AI model
api_calls Number of API calls made
area_under_curve Area under the ROC curve
assigned_items_count Number of items assigned to this participant
average_items_per_day Average items processed per day
backlog_size Number of items remaining
balanced_accuracy Average of sensitivity and specificity
checks_performed Specific checks performed
checksum Checksum of the output file
city City where the institution is located
cohen_kappa Cohen's kappa for inter-rater agreement
completed_items Number of items completed
completed_items_count Number of items completed by this participant
completion_rate Proportion of items completed
confidence_calibration Calibration of AI confidence scores
confidence_interval Confidence interval for the agreement rate
confidence_level Confidence level (e
conflict_of_interest_declared Whether conflicts of interest were declared
corrective_actions Corrective actions taken
cost Estimated cost
country Country where the institution is located
dataset_name Name of the dataset
department Department or division within the institution
diagnostic_odds_ratio Ratio of odds of positive test in diseased vs non-diseased
disagreement_count Number of disagreements
disagreement_types Types of disagreements observed
email Email address of the author
ended_at End time
errors_encountered Errors encountered during the session
expertise_areas Areas of expertise relevant to the review
f1_score Harmonic mean of precision and recall
false_discovery_rate Expected proportion of false discoveries
false_negative_rate Type II error rate
false_omission_rate Proportion of false negatives among negative calls
false_positive_rate Type I error rate
family_name Family/last name of the author
field_identifier Identifier of the field or item
field_level_agreements Agreement metrics for individual fields
first_item_date Date of first item processed
fleiss_kappa Fleiss' kappa for multiple raters
format File format or MIME type
given_name Given/first name of the author
human_agreement_rate Rate of agreement with human reviewers
human_modifications Modifications made by humans to AI outputs
human_oversight_level Level of human review required
inputs Input artifacts
institution_name Name of the affiliated institution
issues_found Number of issues identified
issues_resolved Number of issues resolved
item_ids_processed IDs of items processed in this session
item_modified Item or field that was modified
items_per_day Average items processed per day
items_processed_count Total items processed
kind Kind or type of stage output
krippendorff_alpha Krippendorff's alpha reliability coefficient
last_item_date Date of last item processed
license License governing the dataset
lower_bound Lower bound of the interval
matthews_correlation Matthews correlation coefficient
median_time_per_item Median time per item in minutes
modification_id Unique identifier for modification
modification_reason Reason for the modification
modification_timestamp When the modification was made
modified_value Value after human modification
modifier_id ID of the person who made the modification
name Full name of the author
negative_predictive_value Negative predictive value
notes Additional notes
orcid ORCID identifier for the author
original_ai_value Original value produced by AI
output_created_at Creation timestamp
output_label Human-readable label
outputs Output artifacts
overall_agreement Overall agreement rate across all items
parameter_description Description of what this parameter controls
parameter_name Name of the parameter
parameter_value Value of the parameter
participant_avg_time_per_item Average time spent per item (minutes)
participant_id Reference to the participant
participant_notes Additional notes about the participant
participant_role Role(s) of the participant in the review
peak_items_per_day Maximum items processed in a single day
percent_agreement Simple percentage agreement
precision Positive predictive value
process_avg_time_per_item Average time per item in minutes
process_total_time_hours Total time spent in hours
processing_speed Average processing time per item (seconds)
prompt_examples Few-shot examples included in the prompt
prompt_id Unique identifier for prompt
prompt_text The actual prompt text
prompt_type Type of prompting strategy
prompt_version Version of this prompt
qc_id Unique identifier for this quality control check
qc_pass_rate Proportion of items passing QC
qc_performed_by Who performed the QC
qc_sample_size Size of the sample checked
qc_timestamp When QC was performed
qc_type Type of quality control performed
range_description Description of a date range
range_end End of a date range
range_start Start of a date range
resource_uri URI locating the resource
ror_id Research Organization Registry identifier for the institution
sample_method Method used for sampling
sensitivity True positive rate (recall)
session_duration Duration of the session in seconds
session_timestamp When this AI session occurred
snapshot_hash Checksum of the dataset snapshot
specificity True negative rate
stage_description Detailed description
stage_label Human-readable label
stage_type Type of stage executed
started_at Start time
target_field Field or task this prompt targets
throughput_trend Trend in processing speed
time_spent_hours Total time spent in hours
token_usage Number of tokens used
total_comparisons Total number of comparisons made
total_items Total number of items to process
training_completed Training or calibration exercises completed
upper_bound Upper bound of the interval

Enumerations

Enumeration Description
AIPurpose Purpose for which AI is being used
DatasetKind Types of datasets used in systematic reviews
HumanOversightLevel Level of human oversight required
ModificationReason Reason for modifying AI output
ParticipantRole Roles that participants can have in the review process
PromptType Type of prompting strategy
QualityControlType Types of quality control checks
SamplingMethod Methods for sampling items for quality control
StageOutputKind Types of outputs that can be produced by review stages
StageType Logical stages of a systematic review (typical flow)

Types

Type Description
Boolean A binary (true or false) value
Curie a compact URI
Date a date (year, month and day) in an idealized calendar
DateOrDatetime Either a date or a datetime
Datetime The combination of a date and time
Decimal A real number with arbitrary precision that conforms to the xsd:decimal speci...
Double A real number that conforms to the xsd:double specification
Float A real number that conforms to the xsd:float specification
Integer An integer
Jsonpath A string encoding a JSON Path
Jsonpointer A string encoding a JSON Pointer
Ncname Prefix part of CURIE
Nodeidentifier A URI, CURIE or BNODE that represents a node in a model
Objectidentifier A URI or CURIE that represents an object in the model
Sparqlpath A string encoding a SPARQL Property Path
String A character string
Time A time object represents a (local) time of day, independent of any particular...
Uri a complete URI
Uriorcurie a URI or a CURIE

Subsets

Subset Description