Nested and Balanced Entity Recognition using Multi-Task Learning

Andreas Waldis; Luca Mazzola

RECORD

Nested and Balanced Entity Recognition using Multi-Task Learning

Andreas Waldis, Luca Mazzola

Computation and Language · Submitted 11 Jun 2021

abstract compiled · 2026-09-23 · no language-model review

The authors' main claim Our approach achieves state-of-the-art NER performances, while it outperforms previous CR approaches.

how it is stated · asserted 1.00 read the record →

Comment (0) Read on arXiv ↗

Cite ▾

Claim paper

The compiled record

The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Open the compiled record

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-09-23: 1 request, 6,832 input tokens.

What the abstract itself states, as probabilities: claims first 0.06 · states limitation 0.29 · human participants 0.03 · controlled comparison 0.38 · randomized assignment 0.01 · states quantitative result 0.06 · compares with prior results 0.95 · mentions code or data availability 0.02.

The abstract, sentence by sentence

arXiv listing abstract v1. Results and the main claim are underlined; show every role.

Entity Recognition (ER) within a text is a fundamental exercise in Natural Language Processing, enabling further depending tasks such as Knowledge Extraction, Text Summarisation, or Keyphrase Extraction. An entity consists of single words or of a consecutive sequence of terms, constituting the basic building blocks for communication. Mainstream ER approaches are mainly limited to flat structures, concentrating on the outermost entities while ignoring the inner ones. This paper introduces a partly-layered network architecture that deals with the complexity of overlapping and nested cases. The proposed architecture consists of two parts: (1) a shared Sequence Layer and (2) a stacked component with multiple Tagging Layers. The adoption of such an architecture has the advantage of preventing overfit to a specific word-length, thus maintaining performance for longer entities despite their lower frequency. To verify the proposed architecture's effectiveness, we train and evaluate this architecture to recognise two kinds of entities - Concepts (CR) and Named Entities (NER). Our approach achieves state-of-the-art NER performances, while it outperforms previous CR approaches. Considering these promising results, we see the possibility to evolve the architecture for other cases such as the extraction of events or the detection of argumentative components.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#	sentence	role	the authors' own work	states the main conclusion	strength
s0	Entity Recognition (ER) within a text is a fundamental exercise in Natural Language Processing, enabling further depending tasks such as Knowledge Extraction, Text Summarisation, or Keyphrase Extraction.	background 1.00	0.04	0.02	asserted 1.00
s1	An entity consists of single words or of a consecutive sequence of terms, constituting the basic building blocks for communication.	background 1.00	0.04	0.01	asserted 1.00
s2	Mainstream ER approaches are mainly limited to flat structures, concentrating on the outermost entities while ignoring the inner ones.	gap 0.86	0.05	0.02	asserted 1.00
s3	This paper introduces a partly-layered network architecture that deals with the complexity of overlapping and nested cases.	contribution 0.98	0.83	0.06	asserted 1.00
s4	The proposed architecture consists of two parts: (1) a shared Sequence Layer and (2) a stacked component with multiple Tagging Layers.	method 0.98	0.68	0.03	asserted 1.00
s5	The adoption of such an architecture has the advantage of preventing overfit to a specific word-length, thus maintaining performance for longer entities despite their lower frequency.	method 0.50	0.33	0.05	asserted 0.95
s6	To verify the proposed architecture's effectiveness, we train and evaluate this architecture to recognise two kinds of entities - Concepts (CR) and Named Entities (NER).	method 0.89	0.96	0.04	asserted 0.96
s7	Our approach achieves state-of-the-art NER performances, while it outperforms previous CR approaches.	result 1.00	0.95	0.73	asserted 1.00
s8	Considering these promising results, we see the possibility to evolve the architecture for other cases such as the extraction of events or the detection of argumentative components.	future work 0.99	0.21	0.10	hedged possibility 0.97

What kind of paper the classifier reports this is: method or tool 1.00 · evidence from observation or experiment 0.94.

What was examined, and what was not

sources read

arXiv listing abstract v1

not examined

full text
figures and tables
references and citing sentences
relations to other papers

levels run

P abstract

not run

S sections
T body sentences
C citations
R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

The paper

Original abstract

Entity Recognition (ER) within a text is a fundamental exercise in Natural Language Processing, enabling further depending tasks such as Knowledge Extraction, Text Summarisation, or Keyphrase Extraction. An entity consists of single words or of a consecutive sequence of terms, constituting the basic building blocks for communication. Mainstream ER approaches are mainly limited to flat structures, concentrating on the outermost entities while ignoring the inner ones. This paper introduces a partly-layered network architecture that deals with the complexity of overlapping and nested cases. The proposed architecture consists of two parts: (1) a shared Sequence Layer and (2) a stacked component with multiple Tagging Layers. The adoption of such an architecture has the advantage of preventing overfit to a specific word-length, thus maintaining performance for longer entities despite their lower frequency. To verify the proposed architecture's effectiveness, we train and evaluate this architecture to recognise two kinds of entities - Concepts (CR) and Named Entities (NER). Our approach achieves state-of-the-art NER performances, while it outperforms previous CR approaches. Considering these promising results, we see the possibility to evolve the architecture for other cases such as the extraction of events or the detection of argumentative components.

Human review

How Pith checks papers → · Full audit of this paper →

How the literature uses it

What this paper does with the arXiv papers it cites

Each citing sentence is taken from the citing paper's LaTeX source and typed by Pith's classifier. A label appears at probability 0.8 or more (0.95 for support), a contest at 0.7 or more; the quoted words are the citing authors'. How each label was checked · JSON

What it cites · 4 sentences without a confident label

Uses the method or tool of 1 paper

As loss function, we use the cross-entropy and as optimiser the AdamW [ref] .

Decoupled Weight Decay Regularization · Functional Facets · 0.99

Uses data or values from 1 paper

As showed in [ref] , this imbalance can rise to a ratio of 168:1.

Dice Loss for Data-imbalanced NLP Tasks · Related Work · 0.88

Compares with 3 papers

This comparison considers the overall, word-length, and nested-level performance by comparing the different variations and recent approaches (BiFlaG [ref] , BERT-MRC [ref] , SecondBest [ref] , Layered-CRF [ref] ).

Nested Named Entity Recognition via Second-best Sequence Learning and Decoding · Named Entity Recognition · 1.00
This comparison considers the overall, word-length, and nested-level performance by comparing the different variations and recent approaches (BiFlaG [ref] , BERT-MRC [ref] , SecondBest [ref] , Layered-CRF [ref] ).

A Unified MRC Framework for Named Entity Recognition · Named Entity Recognition · 1.00
This comparison considers the overall, word-length, and nested-level performance by comparing the different variations and recent approaches (BiFlaG [ref] , BERT-MRC [ref] , SecondBest [ref] , Layered-CRF [ref] ).

Bipartite Flat-Graph Network for Nested Named Entity Recognition · Named Entity Recognition · 1.00

Cites as background 5 papers

Recent approaches encode these levels with a specific tagging schema [ref] , or using special architectures like a ensemble of networks [ref] , a layered network [ref] , anchor-region networks [ref] , or graph-based networks [ref] .

Sequence-to-Nuggets: Nested Entity Mention Detection via Anchor-Region Networks · Related Work · 0.88
Recent approaches encode these levels with a specific tagging schema [ref] , or using special architectures like a ensemble of networks [ref] , a layered network [ref] , anchor-region networks [ref] , or graph-based networks [ref] .

Neural Architectures for Nested NER through Linearization · Related Work · 0.88
Recent approaches encode these levels with a specific tagging schema [ref] , or using special architectures like a ensemble of networks [ref] , a layered network [ref] , anchor-region networks [ref] , or graph-based networks [ref] .

Bipartite Flat-Graph Network for Nested Named Entity Recognition · Related Work · 0.88

Discussion

Know something about this paper? Leave a comment, or highlight any passage above to comment on it.