Author name cluster

M. N. Murty

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

3 papers

2 author rows

ECAI Conference 2020 Conference Paper

A Multilayered Informative Random Walk for Attributed Social Network Embedding

Sambaran Bandyopadhyay
Anirban Biswas
Harsh Kara
M. N. Murty

Network representation learning (also known as Graph embedding) is a technique to map the nodes of a network to a lower dimensional vector space. Random walk based representation techniques are found to be efficient as they can easily preserve different orders of proximities between the nodes in the embedding space. Most of the social networks now-a-days have some content (or attributes) associated with each node. These attributes can provide complementary information along with the link structure of the network. But in a real life network, the information carried by the link structure and that by the attributes vary significantly over the nodes. Most of the existing unsupervised attributed network embedding algorithms do not distinguish between the link structure and the attributes of a node depending on their informativeness. In this work, we propose an unsupervised node embedding technique that exploits both the structure and attributes by intelligently prioritizing one of them, in the random walk, for each node separately. We convert the network into a multi-layered graph and propose a novel random walk based on the informativeness of a node in different layers. This unified approach is simple and computationally fast, yet able to use the content as a complement to structure and viceversa. Experimental evaluations on four real world publicly available datasets show the merit of our approach (up to 168. 75% improvement) compared to the state-of-the-art algorithms in the domain. We make the source code available to download.

Details

AAAI Conference 2019 Conference Paper

Outlier Aware Network Embedding for Attributed Networks

Sambaran Bandyopadhyay
N. Lokesh
M. N. Murty

Attributed network embedding has received much interest from the research community as most of the networks come with some content in each node, which is also known as node attributes. Existing attributed network approaches work well when the network is consistent in structure and attributes, and nodes behave as expected. But real world networks often have anomalous nodes. Typically these outliers, being relatively unexplainable, affect the embeddings of other nodes in the network. Thus all the downstream network mining tasks fail miserably in the presence of such outliers. Hence an integrated approach to detect anomalies and reduce their overall effect on the network embedding is required. Towards this end, we propose an unsupervised outlier aware network embedding algorithm (ONE) for attributed networks, which minimizes the effect of the outlier nodes, and hence generates robust network embeddings. We align and jointly optimize the loss functions coming from structure and attributes of the network. To the best of our knowledge, this is the first generic network embedding approach which incorporates the effect of outliers for an attributed network without any supervision. We experimented on publicly available real networks and manually planted different types of outliers to check the performance of the proposed algorithm. Results demonstrate the superiority of our approach to detect the network outliers compared to the state-of-the-art approaches. We also consider different downstream machine learning applications on networks to show the efficiency of ONE as a generic network embedding technique. The source code is made available at https: //github. com/sambaranban/ONE.

PDF Details

AAAI Conference 2018 Conference Paper

Overlap-Robust Decision Boundary Learning for Within-Network Classification

Sharad Nandanwar
M. N. Murty

We study the problem of within network classiﬁcation, where given a partially labeled network, we infer the labels of the remaining nodes based on the link structure. Conventional loss functions penalize a node based on a function of its predicted label and target label. Such loss functions under-perform while learning on a network having overlapping classes. In relational setting, even though the ground truth is not known for the unlabeled nodes, some evidence is present in the form of labeling acquired by the nodes in their neighborhood. We propose a structural loss function for learning in networks based on the hypothesis that loss is induced when a node fails to acquire a label that is consistent with the labels of the majority of the nodes in its neighborhood. We further combine this with a novel semantic regularizer, which we call homophily regularizer, to capture the smooth transition of discriminatory power and behavior of semantically similar nodes. The proposed structural loss along with the regularizer permits relaxation labeling. Through extensive comparative study on different real-world datasets, we found that our method improves over the state-of-the-art approaches.

PDF Details