On the Practicality of Integrity Attacks on Document-Level Sentiment Analysis

By Andrew Newell, Rahul Potharaju, Luojie Xiang, Cristina Nita-Rotaru. 2014

59 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Networks and Communications, Information Systems, Safety, Robustness & Interpretability, Sentiment & Opinion Analysis

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Networks and Communications

In Information Systems

In Safety, Robustness & Interpretability

In Sentiment & Opinion Analysis

Citations in the corpus

Stronger data poisoning attacks break data sanitization defenses (2018, methodology): “This is typical in the literature: attackers might be optimized to act within an attack budget (Mei & Zhu, 2015b) and to only add points that belong to the input domain (e.g., word counts in a document should be integer-valued (Nelson et al., 2008; Newell et al., 2014), but not to evade defenses.”
Sever: A Robust Meta-Algorithm for Stochastic Optimization (2018, background): “Recent work has shown that for high-dimensional datasets, even a small fraction of outliers can substantially degrade the learned model [BNL12, NPXNR14, KL17, SKL17].”
Certified Defenses for Data Poisoning Attacks (2017, methodology): “Beyond the the sphere and slab constraints, a valid attack on text data must satisfy additional integrity constraints (Newell et al., 2014) — if we encode input text t by a vector x = φ ( t ) of indicator features, then each entry of φ ( t ) is a non-negative integer.”
Weight Poisoning Attacks on Pretrained Models (2020, background): “…by Gu et al. (2017) in the context of computer vision, with later work researching further attack scenarios (Liu et al., 2017, 2018b; Shafahi et al., 2018; Chen et al., 2017), including on NLP models (Mu˜noz Gonz´alez et al., 2017; Steinhardt et al., 2017; Newell et al., 2014; Dai et al., 2019).”
Label-Consistent Backdoor Attacks (2019, background): “In these attacks, the adversary introduces outliers that force the model to learn a suboptimal decision boundary (Biggio et al., 2012; Xiao et al., 2012, 2015; Newell et al., 2014; Mei & Zhu, 2015; Steinhardt et al., 2017).”
A Comprehensive Survey on Poisoning Attacks and Countermeasures in Machine Learning (2022, methodology): “, feature extraction) on the raw data before they are input into the model, and the data involved in the model training is usually low dimension feature vectors [5, 90, 132].”
Clean-Label Backdoor Attacks (2018, background): “The most natural form of data poisoning aims to use the malicious samples aim to reduce the test accuracy of the resulting model (Xiao et al., 2012; 2015; Newell et al., 2014; Mei & Zhu, 2015; Burkard & Lagesse, 2017).”
Reflection Backdoor: A Natural Backdoor Attack on Deep Neural Networks (2020, background): “The objective for data poisoning attack is to disrupt the proper training of DNN models by reducing their test-time performance on all or a specific subset of test examples [5,7,31,46,47,69].”
Back to the Drawing Board: A Critical Evaluation of Poisoning Attacks on Production Federated Learning (2021, methodology): “We build on the classic label flipping data poisoning attack [45], [64], [65] designed for centralized ML.”
Online Data Poisoning Attack (2019, background): “A variety of attacks against other learners have been developed, including neural networks [20, 12], autoregressive models [7, 1], linear and stochastic bandits [17, 11], collaborative ﬁltering [14], and sentiment analysis [21], etc.”