Tracking State Changes in Procedural Text: a Challenge Dataset and Models for Process Paragraph Comprehension
@inproceedings{Dalvi2018TrackingSC, title={Tracking State Changes in Procedural Text: a Challenge Dataset and Models for Process Paragraph Comprehension}, author={Bhavana Dalvi and Lifu Huang and Niket Tandon and Wen-tau Yih and Peter E. Clark}, booktitle={NAACL-HLT}, year={2018} }
We present a new dataset and models for comprehending paragraphs about processes (e.g., photosynthesis), an important genre of text describing a dynamic world. The new dataset, ProPara, is the first to contain natural (rather than machine-generated) text about a changing world along with a full annotation of entity states (location and existence) during those changes (81k datapoints). The end-task, tracking the location and existence of entities through the text, is challenging because the… CONTINUE READING
Figures, Tables, and Topics from this paper.
Citations
Publications citing this paper.
SHOWING 1-10 OF 20 CITATIONS
Building Dynamic Knowledge Graphs from Text using Machine Reading Comprehension
VIEW 15 EXCERPTS
CITES BACKGROUND & METHODS
HIGHLY INFLUENCED
Be Consistent! Improving Procedural Text Comprehension using Label Consistency
VIEW 4 EXCERPTS
CITES METHODS, RESULTS & BACKGROUND
HIGHLY INFLUENCED
Commonsense Reasoning for Natural Language Understanding: A Survey of Benchmarks, Resources, and Approaches
VIEW 4 EXCERPTS
CITES BACKGROUND
HIGHLY INFLUENCED
Effective Use of Transformer Networks for Entity Tracking
VIEW 11 EXCERPTS
CITES BACKGROUND & METHODS
HIGHLY INFLUENCED
Everything Happens for a Reason: Discovering the Purpose of Actions in Procedural Text
VIEW 9 EXCERPTS
CITES BACKGROUND & METHODS
Predicting State Changes in Procedural Text using Analogical Question Answering
VIEW 7 EXCERPTS
CITES BACKGROUND & METHODS
HIGHLY INFLUENCED
Tracking Discrete and Continuous Entity State for Process Understanding
VIEW 9 EXCERPTS
CITES METHODS & BACKGROUND
HIGHLY INFLUENCED
Reasoning about Actions and State Changes by Injecting Commonsense Knowledge
VIEW 9 EXCERPTS
CITES BACKGROUND & METHODS
HIGHLY INFLUENCED
Step Semantics : Representations for State Changes in Natural Language
VIEW 4 EXCERPTS
CITES RESULTS, BACKGROUND & METHODS
HIGHLY INFLUENCED
WIQA: A dataset for "What if..." reasoning over procedural text
VIEW 4 EXCERPTS
CITES BACKGROUND
References
Publications referenced by this paper.
SHOWING 1-10 OF 27 REFERENCES
ADADELTA: An Adaptive Learning Rate Method
VIEW 4 EXCERPTS
HIGHLY INFLUENTIAL
SQuAD: 100, 000+ Questions for Machine Comprehension of Text
VIEW 4 EXCERPTS
HIGHLY INFLUENTIAL
Simpler Context-Dependent Logical Forms via Model Projections
VIEW 7 EXCERPTS
HIGHLY INFLUENTIAL
Memory Networks
VIEW 4 EXCERPTS
HIGHLY INFLUENTIAL
Modeling Biological Processes for Reading Comprehension
VIEW 6 EXCERPTS
HIGHLY INFLUENTIAL
Long Short-Term Memory
VIEW 11 EXCERPTS
HIGHLY INFLUENTIAL
Bidirectional Attention Flow for Machine Comprehension
VIEW 17 EXCERPTS
HIGHLY INFLUENTIAL
Query-Reduction Networks for Question Answering
VIEW 16 EXCERPTS
HIGHLY INFLUENTIAL
Tracking the World State with Recurrent Entity Networks
VIEW 5 EXCERPTS
HIGHLY INFLUENTIAL
Dropout: a simple way to prevent neural networks from overfitting
VIEW 2 EXCERPTS
HIGHLY INFLUENTIAL