Knowledge Distillation for Discourse Relation Analysis

Jiang, Congcong; Qian, Tieyun; Liu, Bing

doi:10.1145/3487553.3524245

Cited by 6 publications

(4 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This paper mainly adopts two categories of competitive baselines for the PDTB 2.0 dataset and the CoNLL-2016 shared task 2 . The first category is the previous state-of-the-art (SOTA) baselines, such as TransS , BMGF-RoBERTa (Liu et al, 2020), LDSGM (Wu et al, 2022), XLNet-large (Kim et al, 2020), OTMT (XLNet-large) (Jiang et al, 2022a), and Con-trastiveIDRR (Long and Webber, 2022). Two partitions of these SOTA baselines are highlighted for comparison with our method.…”

Section: Baselinesmentioning

confidence: 99%

“…• OTMT(XLNet) (Jiang et al, 2022a) to obtain better label embeddings, and then employ the label sequence decoder to output the predicted labels.…”

Section: A3 Baseline Modelsmentioning

confidence: 99%

“…Model Parameters BMGF-RoBERTa (Liu et al, 2020) 2.3M XLNet(base, cased) (Kim et al, 2020) 110M XLNet(large, cased) (Kim et al, 2020) 340M OTMT(XLNet-base) (Jiang et al, 2022a) 110M OTMT(XLNet-large) (Jiang et al, 2022a) 340M Fine-Tuning (T5-base) (Raffel et al, 2020) 220M Fine-Tuning (T5-large) (Raffel et al, 2020) 770M Prefix-Tuning (T5-base) (Li and Liang, 2021) 0.12M Prefix-Tuning (T5-large) (Li and Liang, 2021) 0.16M Prompt-Tuning (T5-base) (Lester et al, 2021) 0.12M Prompt-Tuning (T5-large) (Lester et al, 2021) 0…”

Section: Model Input and Outputmentioning

confidence: 99%

See 2 more Smart Citations

DiscoPrompt: Path Prediction Prompt Tuning for Implicit Discourse Relation Recognition

Chan¹,

Liu²,

Cheng³

et al. 2023

Findings of the Association for Computational Linguistics: ACL 2023

View full text Add to dashboard Cite

Implicit Discourse Relation Recognition (IDRR) is a sophisticated and challenging task to recognize the discourse relations between the arguments with the absence of discourse connectives. The sense labels for each discourse relation follow a hierarchical classification scheme in the annotation process (Prasad et al., 2008), forming a hierarchy structure. Most existing works do not well incorporate the hierarchy structure but focus on the syntax features and the prior knowledge of connectives in the manner of pure text classification. We argue that it is more effective to predict the paths inside the hierarchical tree (e.g., "Comparison -> Contrast -> however") rather than flat labels (e.g., Contrast) or connectives (e.g., however). We propose a prompt-based path prediction method to utilize the interactive information and intrinsic senses among the hierarchy in IDRR. This is the first work that injects such structure information into pre-trained language models via prompt tuning, and the performance of our solution shows significant and consistent improvement against competitive baselines.

show abstract

Section: Baselinesmentioning

confidence: 99%

“…• OTMT(XLNet) (Jiang et al, 2022a) to obtain better label embeddings, and then employ the label sequence decoder to output the predicted labels.…”

Section: A3 Baseline Modelsmentioning

confidence: 99%

Section: Model Input and Outputmentioning

confidence: 99%

See 1 more Smart Citation

DiscoPrompt: Path Prediction Prompt Tuning for Implicit Discourse Relation Recognition

Chan¹,

Liu²,

Cheng³

et al. 2023

Findings of the Association for Computational Linguistics: ACL 2023

View full text Add to dashboard Cite

show abstract

“…Being able to answer this question allows to also address another question of bigger theoretical depth and importance, namely how coherence and cohesion are realized in the discourse. Exploring discourse relations remains a topic of growing interest, and several rather sophisticated approaches have recently been proposed [7][8][9][10][11][12].…”

Section: Introductionmentioning

confidence: 99%

From Discourse Relations to Network Edges: A Network Theory Approach to Discourse Analysis

Tantos

Kosmidis

2023

Applied Sciences

View full text Add to dashboard Cite

In this paper, we argue that discourse representations can be mapped to networks and analyzed by tools provided in network theory so that deep properties of discourse structure are revealed. Two discourse-annotated corpora, C58 and STAC, that belong to different discourse types and languages were compared and analyzed. Various key network indices were used for the discourse representations of both corpora and show the different network profiles of the two discourse types. Moreover, both network motifs and antimotifs were discovered for the discourse networks in the two corpora that shed light on strong tendencies in building or avoiding to build discourse relations between utterances for permissible three-node discourse subgraphs. These results may lead to new types of discourse structure rules that draw on the properties of the networks that lie behind discourse representation. Another important aspect is that the second version of the STAC corpus, which includes nonlinguistic discourse units and their relations, exhibits similar trends in terms of network subgraphs compared to its first version. This suggests that the nonlinguistic context has a significant impact on discourse structure.

show abstract

Bias-Rectified Multi-way Learning with Data Augmentation for Implicit Discourse Relation Recognition

Zheng,

Liang,

Xiang

et al. 2024

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Knowledge Distillation for Discourse Relation Analysis

Cited by 6 publications

References 5 publications

DiscoPrompt: Path Prediction Prompt Tuning for Implicit Discourse Relation Recognition

DiscoPrompt: Path Prediction Prompt Tuning for Implicit Discourse Relation Recognition

From Discourse Relations to Network Edges: A Network Theory Approach to Discourse Analysis

Bias-Rectified Multi-way Learning with Data Augmentation for Implicit Discourse Relation Recognition

Contact Info

Product

Resources

About