Artikel ; Online: Improving Robustness of Intent Detection Under Adversarial Attacks: A Geometric Constraint Perspective.
IEEE transactions on neural networks and learning systems
2024 Band 35, Heft 5, Seite(n) 6133–6144
Abstract: Deep neural networks (DNNs)-based natural language processing (NLP) systems are vulnerable to being fooled by adversarial examples presented in recent studies. Intent detection tasks in dialog systems are no exception, however, relatively few works have ... ...
Abstract | Deep neural networks (DNNs)-based natural language processing (NLP) systems are vulnerable to being fooled by adversarial examples presented in recent studies. Intent detection tasks in dialog systems are no exception, however, relatively few works have been attempted on the defense side. The combination of linear classifier and softmax is widely used in most defense methods for other NLP tasks. Unfortunately, it does not encourage the model to learn well-separated feature representations. Thus, it is easy to induce adversarial examples. In this article, we propose a simple, yet efficient defense method from the geometric constraint perspective. Specifically, we first propose an M-similarity metric to shrink variances of intraclass features. Intuitively, better geometric conditions of feature space can bring lower misclassification probability (MP). Therefore, we derive the optimal geometric constraints of anchors within each category from the overall MP (OMP) with theoretical guarantees. Due to the nonconvex characteristic of the optimal geometric condition, it is hard to satisfy the traditional optimization process. To this end, we regard such geometric constraints as manifold optimization processes in the Stiefel manifold, thus naturally avoiding the above challenges. Experimental results demonstrate that our method can significantly improve robustness compared with baselines, while retaining the excellent performance on normal examples. |
---|---|
Sprache | Englisch |
Erscheinungsdatum | 2024-05-02 |
Erscheinungsland | United States |
Dokumenttyp | Journal Article |
ISSN | 2162-2388 |
ISSN (online) | 2162-2388 |
DOI | 10.1109/TNNLS.2023.3267460 |
Datenquelle | MEDical Literature Analysis and Retrieval System OnLINE |
Zusatzmaterialien
Kategorien
Über subito bestellen
Dieser Service ist kostenpflichtig (siehe Lieferbedingungen von subito). Bestellungen, die einen Artikel nebst Supplementary Material umfassen, werden grundsätzlich wie mehrfache Bestellungen bearbeitet. Gebühren fallen in diesen Fällen für jede einzelne Bestellung an.
Fernleihe an ZB MED
Sie können sich den gewünschten Titel als lokale Nutzerin oder lokaler Nutzer von ZB MED direkt an den Standort Köln schicken lassen.