In practice, users of a Recommender System (RS) fall into a few clusters based on their preferenc... more In practice, users of a Recommender System (RS) fall into a few clusters based on their preferences. In this work, we conduct a systematic study on user-cluster targeted data poisoning attacks on Matrix Factorisation (MF) based RS, where an adversary injects fake users with falsely crafted user-item feedback to promote an item to a specific user cluster. We analyse how user and item feature matrices change after data poisoning attacks and identify the factors that influence the effectiveness of the attack on these feature matrices. We demonstrate that the adversary can easily target specific user clusters with minimal effort and that some items are more susceptible to attacks than others. Our theoretical analysis has been validated by the experimental results obtained from two real-world datasets. Our observations from the study could serve as a motivating point to design a more robust RS.
Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval, 2021
How to quickly and reliably learn the preferences of new users remains a key challenge in the des... more How to quickly and reliably learn the preferences of new users remains a key challenge in the design of recommender systems. In this paper we introduce a new type of online learning algorithm, cluster-based bandits, to address this challenge. This exploits the fact that users can often be grouped into clusters based on the similarity of their preferences, and this allows accelerated learning of new user preferences since the task becomes one of identifying which cluster a user belongs to and typically there are far fewer clusters than there are items to be rated. Clustering by itself is not enough however. Intra-cluster variability between users can be thought of as adding noise to user ratings. Deterministic methods such as decision-trees perform poorly in the presence of such noise. We identify so-called distinguisher items that are particularly informative for deciding which cluster a new user belongs to despite the rating noise. Using these items the cluster-based bandit algorithm is able to efficiently adapt to user responses and rapidly learn the correct cluster to assign to a new user.
Distributed Detection in Cognitive Radio Networks with Unknown Primary User's Traffic
TENCON 2019 - 2019 IEEE Region 10 Conference (TENCON)
We consider a distributed Cognitive Radio Network (CRN) in which a number of secondary users (SUs... more We consider a distributed Cognitive Radio Network (CRN) in which a number of secondary users (SUs) cooperatively sense the spectrum holes in a channel whose occupancy by the primary user (PU) follows a discrete time Markov chain. We consider a problem where the transition probability matrix (TPM) of the channel occupancy process followed by the PU is unknown. For this problem, we first propose a procedure to estimate the TPM. We then seek a distributed detection procedure with the estimated TPM. We show that our estimate of the TPM is very close to that of the actual TPM in the case of perfect sensing; in the case of imperfect sensing, the estimate has a little bias. This bias can be used to correct the estimate further. We study the detection and the false alarm characteristics of the distributed detection procedure that we propose (which uses the estimates on the channel occupancy probabilities). The proposed distributed detection procedure achieves a good Receiver Operating Characteristic (ROC) performance which is close to that of the centralised CRN.
In practice, users of a Recommender System (RS) fall into a few clusters based on their preferenc... more In practice, users of a Recommender System (RS) fall into a few clusters based on their preferences. In this work, we conduct a systematic study on user-cluster targeted data poisoning attacks on Matrix Factorisation (MF) based RS, where an adversary injects fake users with falsely crafted user-item feedback to promote an item to a specific user cluster. We analyse how user and item feature matrices change after data poisoning attacks and identify the factors that influence the effectiveness of the attack on these feature matrices. We demonstrate that the adversary can easily target specific user clusters with minimal effort and that some items are more susceptible to attacks than others. Our theoretical analysis has been validated by the experimental results obtained from two real-world datasets. Our observations from the study could serve as a motivating point to design a more robust RS.
Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval, 2021
How to quickly and reliably learn the preferences of new users remains a key challenge in the des... more How to quickly and reliably learn the preferences of new users remains a key challenge in the design of recommender systems. In this paper we introduce a new type of online learning algorithm, cluster-based bandits, to address this challenge. This exploits the fact that users can often be grouped into clusters based on the similarity of their preferences, and this allows accelerated learning of new user preferences since the task becomes one of identifying which cluster a user belongs to and typically there are far fewer clusters than there are items to be rated. Clustering by itself is not enough however. Intra-cluster variability between users can be thought of as adding noise to user ratings. Deterministic methods such as decision-trees perform poorly in the presence of such noise. We identify so-called distinguisher items that are particularly informative for deciding which cluster a new user belongs to despite the rating noise. Using these items the cluster-based bandit algorithm is able to efficiently adapt to user responses and rapidly learn the correct cluster to assign to a new user.
Distributed Detection in Cognitive Radio Networks with Unknown Primary User's Traffic
TENCON 2019 - 2019 IEEE Region 10 Conference (TENCON)
We consider a distributed Cognitive Radio Network (CRN) in which a number of secondary users (SUs... more We consider a distributed Cognitive Radio Network (CRN) in which a number of secondary users (SUs) cooperatively sense the spectrum holes in a channel whose occupancy by the primary user (PU) follows a discrete time Markov chain. We consider a problem where the transition probability matrix (TPM) of the channel occupancy process followed by the PU is unknown. For this problem, we first propose a procedure to estimate the TPM. We then seek a distributed detection procedure with the estimated TPM. We show that our estimate of the TPM is very close to that of the actual TPM in the case of perfect sensing; in the case of imperfect sensing, the estimate has a little bias. This bias can be used to correct the estimate further. We study the detection and the false alarm characteristics of the distributed detection procedure that we propose (which uses the estimates on the channel occupancy probabilities). The proposed distributed detection procedure achieves a good Receiver Operating Characteristic (ROC) performance which is close to that of the centralised CRN.
Uploads
Papers by Sulthana Shams