User Input
Analysed Document
User Input
Analysed Document
Document ID
user_input
Text
Scholarly Argumentation Mining (SAM) has recently gained attention due to its potential to help scholars with the rapid growth of published scientific literature. It comprises two subtasks: argumentative discourse unit recognition (ADUR) and argumentative relation extraction (ARE), both of which are challenging since they require e.g. the integration of domain knowledge, the detection of implicit statements, and the disambiguation of argument structure. While previous work focused on dataset construction and baseline methods for specific document sections, such as abstract or results, full-text scholarly argumentation mining has seen little progress. In this work, we introduce a sequential pipeline model combining ADUR and ARE for full-text SAM, and provide a first analysis of the performance of pretrained language models (PLMs) on both subtasks. We establish a new SotA for ADUR on the Sci-Arg corpus, outperforming the previous best reported result by a large margin (+7% F1). We also present the first results for ARE, and thus for the full AM pipeline, on this benchmark dataset. Our detailed error analysis reveals that non-contiguous ADUs as well as the interpretation of discourse connectors pose major challenges and that data annotation needs to be more consistent.
Model Configuration
▼
argumentation structure
▼
Argumentation Model Configuration
Load Argumentation Model
retriever
▼
Retriever Configuration
Load Retriever
Device (e.g. 'cuda' or 'cpu')
cpu
Regex to partition the text
\n\n\n+
Analyse
Document ID
Render Options
▼
Render with
Render Arguments
Re-render
See plain result ...
▼
Fetch annotated document as JSON
Model Output
Retrieval
Import Documents
Retrieval
Import Documents
Indexed Documents
▼
id
num_adus
num_relations
id
num_adus
num_relations
Data Snapshot:
Download
Upload
ID (hover)
ID (selected)
Selected ADU
Relevant ADUs from other documents
▼
relation
adu
reference_adu
doc_id
sim_score
rel_score
relation
adu
reference_adu
doc_id
sim_score
rel_score
Retrieval Configuration
▼
Minimum Similarity
↺
0
1
Top K
↺
2
50
Retrieve *similar* ADUs for *selected* ADU
doc_id
adu_id
score
text
doc_id
adu_id
score
text
Retrieve *similar* ADUs for *all* ADUs in the document
doc_id
query_adu_id
adu_id
score
text
doc_id
query_adu_id
adu_id
score
text
Retrieve *relevant* ADUs for *all* ADUs in the document
doc_id
adu_id
score
text
query_span_id
doc_id
adu_id
score
text
query_span_id
Textbox
Batch Analyse Texts
Batch Analyse PDFs
ACL Anthology Venues
Import from ACL Anthology
Import text from arXiv
▼
arXiv paper ID
abstract only
Load & Analyse from arXiv
Import argument structure annotated PIE dataset
▼
Parameters for Loading the PIE Dataset
Load & Embed PIE Dataset