Scholarlink

<p>Sarcasm detection is challenging in natural language processing due to contextual, emotional, and cultural nuances. This paper proposes a dual-encoder BERT (D-BERT) architecture addressing contextual diversity and sentiment reversal challenges in sarcasm detection. We developed D-BERT with two specialized stages: the first BERT extracts deep contextual embeddings while the second performs classification. The model was trained on 149,480 samples from Twitter, Reddit, SemEval, and The Onion, representing diverse communication styles. D-BERT achieved validation accuracies of 91% (95% CI(Confidence Interval): [0.902, 0.918]) on multi-domain data, 86.4% on Twitter, and 83.6% on Reddit, with F1-scores of 0.91, 0.86, and 0.84 respectively. Statistical validation confirmed 13% improvement over single-encoder BERT (McNemar’s <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\varvec{\chi }^{\varvec{2}} = \varvec{178.4}\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\varvec{p}\varvec{&lt;} \varvec{0.0001}\)</EquationSource> </InlineEquation>, Cohen’s <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(\varvec{h = 0.82}\)</EquationSource> </InlineEquation>), representing large practical effect. Performance variation across platforms highlights domain-specific challenges. Adversarial evaluation relies on 15 qualitative test cases rather than systematic benchmarks. No LLM comparison was conducted. Evaluation demonstrates within-distribution generalization, not cross-domain transfer. Doubled computational cost (24.7ms vs 12.3ms) limits scalability. D-BERT effectively captures semantic and contextual features through dual-stage processing, advancing sarcasm detection while acknowledging scope boundaries for future research.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Dual Encoder Architecture for Robust, Adversarial Aware Sarcasm Detection across Heterogeneous Text Corpora

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月7日

查看全文
  • Ramakrishna Bodige,
  • Rameshbabu Akarapu,
  • Pramod Kumar P

摘要

Sarcasm detection is challenging in natural language processing due to contextual, emotional, and cultural nuances. This paper proposes a dual-encoder BERT (D-BERT) architecture addressing contextual diversity and sentiment reversal challenges in sarcasm detection. We developed D-BERT with two specialized stages: the first BERT extracts deep contextual embeddings while the second performs classification. The model was trained on 149,480 samples from Twitter, Reddit, SemEval, and The Onion, representing diverse communication styles. D-BERT achieved validation accuracies of 91% (95% CI(Confidence Interval): [0.902, 0.918]) on multi-domain data, 86.4% on Twitter, and 83.6% on Reddit, with F1-scores of 0.91, 0.86, and 0.84 respectively. Statistical validation confirmed 13% improvement over single-encoder BERT (McNemar’s \(\varvec{\chi }^{\varvec{2}} = \varvec{178.4}\) , \(\varvec{p}\varvec{<} \varvec{0.0001}\) , Cohen’s \(\varvec{h = 0.82}\) ), representing large practical effect. Performance variation across platforms highlights domain-specific challenges. Adversarial evaluation relies on 15 qualitative test cases rather than systematic benchmarks. No LLM comparison was conducted. Evaluation demonstrates within-distribution generalization, not cross-domain transfer. Doubled computational cost (24.7ms vs 12.3ms) limits scalability. D-BERT effectively captures semantic and contextual features through dual-stage processing, advancing sarcasm detection while acknowledging scope boundaries for future research.