首页 > AI前沿 > DLS: Dual-Logic Graph Neural Networks with Chaos-Fuzzy Residuals for Robust Link Prediction

DLS: Dual-Logic Graph Neural Networks with Chaos-Fuzzy Residuals for Robust Link Prediction

arXiv机器学习 2026-11-17 19:22 11 阅读 查看原文
用户需要我将一段纯文本的新闻(实际上是学术论文摘要)排版成语义化的HTML片段。 规则如下: 1. 只输出HTML片段本身,不要Markdown围栏、不要解释。 2. 每个自然段落用`

`包裹,长文用`

`/`

`划分小节。 3. 列表内容用`
  • `,引用用`
    `,代码用`
    `,重要句子可用``。
    4. 保留原文所有信息,不得增删改内容。
    5. 不要输出``/``/``等文档级标签。
    
    正文内容:
    "Link prediction is a fundamental task in graph mining, critical for applications ranging from biological interaction discovery to recommender systems. While Message Passing Neural Networks (MPNNs) have achieved state-of-the-art results, they are fundamentally limited by the inadvertent suppression of high-frequency signals, as the recursive aggregation that stabilizes structural learning inherently leads to over-smoothing, rendering node representations indistinguishable in deep layers. This limitation is particularly severe in heterophilous graphs, where structural proximity does not imply feature similarity. In this paper, we propose DLS (Dual-Logic SAGE), a novel inductive framework designed to break the over-smoothing barrier via a "controlled instability" mechanism. Unlike conventional methods relying on static structural heuristics, DLS introduces a dynamic residual pipeline. First, a Chaotic Injection Module utilizes the Logistic Map to diversify the feature manifold, exploiting sensitivity to initial conditions to separate topologically similar but semantically distinct nodes. Second, to mitigate the aleatoric uncertainty introduced by chaotic dynamics, a Fuzzy-Logic Gating Mechanism applies soft-thresholding to filter noise and retain only high-confidence signals. Experimental results on diverse benchmarks demonstrate that DLS effectively balances feature discrimination with robustness, significantly outperforming state-of-the-art baselines in heterophilous settings while maintaining competitive efficiency on standard homophilous graphs."
    
    分析正文结构:
    这是一段完整的学术摘要,可以分为几个逻辑部分,但原文是一整段。根据规则“每个自然段落用

    包裹,长文用

    /

    划分小节”,我可以将这段长文划分为几个小节,或者就把它当成一个长段落,加几个小标题。 不过,原文并没有明显的小标题。如果强行加小标题可能会被视为“增删改内容”。规则说“保留原文所有信息,不得增删改内容”,如果我添加了原文没有的小标题文字,就违反了规则。 因此,最安全的做法是:不添加任何原文没有的小标题文字。可以将原文分为几个自然段吗?原文是一整段,如果拆成多个`

    `,算不算“增删改”?通常拆分段落如果只是换行,不改变文字,是可以的。但原文本没有段落分隔。 或者,我可以直接把整段放在一个`

    `里,然后对一些重要的句子或术语使用``。 为了体现“长文用

    /

    划分小节”,但又不能增删内容,这似乎是一个矛盾。如果我不加标题,就不使用`

    /

    `。规则是“长文用

    /

    划分小节”,如果没有标题,我就直接用`

    `。 或者,我可以把原文按照逻辑分成几个段落,每段用`

    `包裹。 段落1:Link prediction is a fundamental task... feature similarity. 段落2:In this paper, we propose DLS... dynamic residual pipeline. 段落3:First, a Chaotic Injection Module... high-confidence signals. 段落4:Experimental results on diverse benchmarks... homophilous graphs. 这样拆分段落不改变文字内容,只是增加了段落结构,符合“每个自然段落用

    包裹”。 重要句子/词汇可以用``,例如: - Link prediction - Message Passing Neural Networks (MPNNs) - over-smoothing - heterophilous graphs - DLS (Dual-Logic SAGE) - "controlled instability" - Chaotic Injection Module - Fuzzy-Logic Gating Mechanism 让我们仔细拆分并加粗:

    Link prediction is a fundamental task in graph mining, critical for applications ranging from biological interaction discovery to recommender systems. While Message Passing Neural Networks (MPNNs) have achieved state-of-the-art results, they are fundamentally limited by the inadvertent suppression of high-frequency signals, as the recursive aggregation that stabilizes structural learning inherently leads to over-smoothing, rendering node representations indistinguishable in deep layers. This limitation is particularly severe in heterophilous graphs, where structural proximity does not imply feature similarity.

    In this paper, we propose DLS (Dual-Logic SAGE), a novel inductive framework designed to break the over-smoothing barrier via a "controlled instability" mechanism. Unlike conventional methods relying on static structural heuristics, DLS introduces a dynamic residual pipeline.

    First, a Chaotic Injection Module utilizes the Logistic Map to diversify the feature manifold, exploiting sensitivity to initial conditions to separate topologically similar but semantically distinct nodes. Second, to mitigate the aleatoric uncertainty introduced by chaotic dynamics, a Fuzzy-Logic Gating Mechanism applies soft-thresholding to filter noise and retain only high-confidence signals.

    Experimental results on diverse benchmarks demonstrate that DLS effectively balances feature discrimination with robustness, significantly outperforming state-of-the-art baselines in heterophilous settings while maintaining competitive efficiency on standard homophilous graphs.

    检查规则: 1. 只输出HTML片段本身,不要Markdown围栏、不要解释。 -> 好的。 2. 每个自然段落用

    包裹,长文用

    /

    划分小节。 -> 这里没有小节标题,所以只用

    。如果非要用h2/h3,只能加原文没有的字,这违反了规则4。所以不用h2/h3是合理的。 3. 列表内容用

    • ,引用用
      ,代码用
      ,重要句子可用。 -> 使用了。
      4. 保留原文所有信息,不得增删改内容。 -> 文字完全一致,只是加了HTML标签和分段。
      5. 不要输出//等文档级标签。 -> 好的。
      
      再核对一遍文字,确保没有遗漏或修改:
      原文:
      Link prediction is a fundamental task in graph mining, critical for applications ranging from biological interaction discovery to recommender systems. While Message Passing Neural Networks (MPNNs) have achieved state-of-the-art results, they are fundamentally limited by the inadvertent suppression of high-frequency signals, as the recursive aggregation that stabilizes structural learning inherently leads to over-smoothing, rendering node representations indistinguishable in deep layers. This limitation is particularly severe in heterophilous graphs, where structural proximity does not imply feature similarity. In this paper, we propose DLS (Dual-Logic SAGE), a novel inductive framework designed to break the over-smoothing barrier via a "controlled instability" mechanism. Unlike conventional methods relying on static structural heuristics, DLS introduces a dynamic residual pipeline. First, a Chaotic Injection Module utilizes the Logistic Map to diversify the feature manifold, exploiting sensitivity to initial conditions to separate topologically similar but semantically distinct nodes. Second, to mitigate the aleatoric uncertainty introduced by chaotic dynamics, a Fuzzy-Logic Gating Mechanism applies soft-thresholding to filter noise and retain only high-confidence signals. Experimental results on diverse benchmarks demonstrate that DLS effectively balances feature discrimination with robustness, significantly outperforming state-of-the-art baselines in heterophilous settings while maintaining competitive efficiency on standard homophilous graphs.
      
      我的HTML:
      

      Link prediction is a fundamental task in graph mining, critical for applications ranging from biological interaction discovery to recommender systems. While Message Passing Neural Networks (MPNNs) have achieved state-of-the-art results, they are fundamentally limited by the inadvertent suppression of high-frequency signals, as the recursive aggregation that