Data Poisoning Attacks on Multi-Task Relationship Learning

By Mengchen Zhao, Bo An, Yaodong Yu, Sulin Liu, Sinno Jialin Pan. 2018

46 citations recorded by the source corpus.

Multi-task learning (MTL) is a machine learning paradigm that improves the performance of each task by exploiting useful information contained in multiple related tasks. However, the relatedness of tasks can be exploited by attackers to launch data poisoning attacks, which has been demonstrated a big threat to single-task learning. In this paper, we provide the first study on the vulnerability of MTL. Specifically, we focus on multi-task relationship learning (MTRL) models, a popular subclass of MTL models where task relationships are quantized and are learned directly from training data. We formulate the problem of computing optimal poisoning attacks on MTRL as a bilevel program that is adaptive to arbitrary choice of target tasks and attacking tasks. We propose an efficient algorithm called PATOM for computing optimal attack strategies. PATOM leverages the optimality conditions of the subproblem of MTRL to compute the implicit gradients of the upper level objective function. Experimental results on real-world datasets show that MTRL models are very sensitive to poisoning attacks and the attacker can significantly degrade the performance of target tasks, by either directly poisoning the target tasks or indirectly poisoning the related tasks exploiting the task relatedness. We also found that the tasks being attacked are always strongly correlated, which provides a clue for defending against such attacks.

Subfields: Artificial Intelligence, Clinical Psychology, Safety, Robustness & Interpretability, Privacy & Security

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 Safety, Robustness & Interpretability

In Privacy & Security

Eric Wallace

Citations in the corpus

Poisoning Language Models During Instruction Tuning (2023, background): “For example, some work poisons traditional multi-task or federated learning models (Zhao et al., 2018; Sun et al., 2021) and others poison the pre-training or self-supervision stages and affect downstream ﬁne-tuning results (Liu et al., 2022; Carlini, 2021; Carlini & Terzis, 2022).”
Adversarial Attacks on Neural Networks for Graph Data (2018, background): “Even [34], which considers relations between different tasks for multi-task relationship learning, still deals with the classical scenario of i.”
Data Poisoning Attacks on Federated Machine Learning (2020, methodology): “, [5], [21]), and design a projected stochastic gradient ascent based algorithm that efficiently maximally increases the empirical loss of target nodes, and further damages their classification/regression performance.”
Adversarial Attacks on Node Embeddings via Graph Poisoning (2018, background): “Zhao et al. (2018) study poisoning attacks on multi-task relationship learning, although they exploit relations between tasks, they still deal with the classic scenario of i.i.d. instances within each task.”