DataScienceUIBK commited on
Commit
c80d71a
·
verified ·
1 Parent(s): 724d781

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +80 -1
README.md CHANGED
@@ -1,3 +1,82 @@
1
  ---
2
  license: cc0-1.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: cc0-1.0
3
+ task_categories:
4
+ - question-answering
5
+ language:
6
+ - en
7
+ size_categories:
8
+ - 100M<n<1B
9
+ ---
10
+ # COMPLEXTEMPQA Dataset
11
+
12
+ COMPLEXTEMPQA is a large-scale dataset designed for complex temporal question answering (TQA). It consists of over 100 million question-answer pairs, making it one of the most extensive datasets available for TQA. The dataset is generated using data from Wikipedia and Wikidata and spans questions over a period of 36 years (1987-2023).
13
+
14
+ ## Dataset Description
15
+
16
+ COMPLEXTEMPQA categorizes questions into three main types:
17
+ - Attribute Questions
18
+ - Comparison Questions
19
+ - Counting Questions
20
+
21
+ These categories are further divided based on their relation to events, entities, or time periods.
22
+
23
+ ### Question Types and Counts
24
+
25
+ | Question Type | Subtype | Count |
26
+ |-----------------------|---------------------|---------------|
27
+ | Attribute | Event | 83,798 |
28
+ | Attribute | Entity | 84,079 |
29
+ | Attribute | Time | 9,454 |
30
+ | Comparison | Event | 25,353,340 |
31
+ | Comparison | Entity | 74,678,117 |
32
+ | Comparison | Time | 54,022,952 |
33
+ | Counting | Event | 18,325 |
34
+ | Counting | Entity | 10,798 |
35
+ | Counting | Time | 12,732 |
36
+ | Multi-Hop | | 76,933 |
37
+ | Unnamed Event | | 8,707,123 |
38
+ | **Total** | | **100,228,457**|
39
+
40
+ ### Metadata
41
+
42
+ Each question in the dataset is accompanied by detailed metadata, including:
43
+ - Type of question based on taxonomy
44
+ - Wikidata IDs of the questioned entities or events
45
+ - Country information for both questions and answers
46
+ - Difficulty rating (easy or hard)
47
+ - Time span related to the question
48
+
49
+ ## Dataset Characteristics
50
+
51
+ ### Size
52
+ COMPLEXTEMPQA comprises over 100 million question-answer pairs, focusing on events, entities, and time periods from 1987 to 2023.
53
+
54
+ ### Complexity
55
+ Questions require advanced reasoning skills, including multi-hop question answering, temporal aggregation, and across-time comparisons.
56
+
57
+ ### Taxonomy
58
+ The dataset follows a unique taxonomy categorizing questions into attributes, comparisons, and counting types, ensuring comprehensive coverage of temporal queries.
59
+
60
+ ### Evaluation
61
+ The dataset has been evaluated for readability, ease of answering before and after web searches, and overall clarity. Human raters have assessed a sample of questions to ensure high quality.
62
+
63
+ ## Usage
64
+
65
+ ### Evaluation and Training
66
+ COMPLEXTEMPQA can be used for:
67
+ - Evaluating the temporal reasoning capabilities of large language models (LLMs)
68
+ - Fine-tuning language models for better temporal understanding
69
+ - Developing and testing retrieval-augmented generation (RAG) systems
70
+
71
+ ### Research Applications
72
+ The dataset supports research in:
73
+ - Temporal question answering
74
+ - Information retrieval
75
+ - Language understanding
76
+
77
+ ### Adaptation and Continual Learning
78
+ COMPLEXTEMPQA's temporal metadata facilitates the development of online adaptation and continual training approaches for LLMs, aiding in the exploration of time-based learning and evaluation.
79
+
80
+ ## Access
81
+
82
+ The dataset and code are freely available at [https://github.com/DataScienceUIBK/ComplexTempQA](https://github.com/DataScienceUIBK/ComplexTempQA).