Researcher profile

Ravi Sundaram

Ravi Sundaram contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate

Trust snapshot

Quick read

Trust 15 - UnverifiedVerification L1Unclaimed author
3works
0followers
3topics
4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Identity and collaboration

How to connect with this researcher

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Log in to claim

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

3 published item(s)

preprint2020arXiv

Attention improves concentration when learning node embeddings

We consider the problem of predicting edges in a graph from node attributes in an e-commerce setting. Specifically, given nodes labelled with search query text, we want to predict links to related queries that share products. Experiments with a range of deep neural architectures show that simple feedforward networks with an attention mechanism perform best for learning embeddings. The simplicity of these models allows us to explain the performance of attention. We propose an analytically tractable model of query generation, AttEST, that views both products and the query text as vectors embedded in a latent space. We prove (and empirically validate) that the point-wise mutual information (PMI) matrix of the AttEST query text embeddings displays a low-rank behavior analogous to that observed in word embeddings. This low-rank property allows us to derive a loss function that maximizes the mutual information between related queries which is used to train an attention network to learn query embeddings. This AttEST network beats traditional memory-based LSTM architectures by over 20% on F-1 score. We justify this out-performance by showing that the weights from the attention mechanism correlate strongly with the weights of the best linear unbiased estimator (BLUE) for the product vectors, and conclude that attention plays an important role in variance reduction.

preprint2011arXiv

Scheduler Vulnerabilities and Attacks in Cloud Computing

In hardware virtualization a hypervisor provides multiple Virtual Machines (VMs) on a single physical system, each executing a separate operating system instance. The hypervisor schedules execution of these VMs much as the scheduler in an operating system does, balancing factors such as fairness and I/O performance. As in an operating system, the scheduler may be vulnerable to malicious behavior on the part of users seeking to deny service to others or maximize their own resource usage. Recently, publically available cloud computing services such as Amazon EC2 have used virtualization to provide customers with virtual machines running on the provider's hardware, typically charging by wall clock time rather than resources consumed. Under this business model, manipulation of the scheduler may allow theft of service at the expense of other customers, rather than merely reallocating resources within the same administrative domain. We describe a flaw in the Xen scheduler allowing virtual machines to consume almost all CPU time, in preference to other users, and demonstrate kernel-based and user-space versions of the attack. We show results demonstrating the vulnerability in the lab, consuming as much as 98% of CPU time regardless of fair share, as well as on Amazon EC2, where Xen modifications protect other users but still allow theft of service. In case of EC2, following the responsible disclosure model, we have reported this vulnerability to Amazon; they have since implemented a fix that we have tested and verified (See Appendix B). We provide a novel analysis of the necessary conditions for such attacks, and describe scheduler modifications to eliminate the vulnerability. We present experimental results demonstrating the effectiveness of these defenses while imposing negligible overhead.

preprint2007arXiv

A bounded-degree network formation game

Motivated by applications in peer-to-peer and overlay networks we define and study the \emph{Bounded Degree Network Formation} (BDNF) game. In an $(n,k)$-BDNF game, we are given $n$ nodes, a bound $k$ on the out-degree of each node, and a weight $w_{vu}$ for each ordered pair $(v,u)$ representing the traffic rate from node $v$ to node $u$. Each node $v$ uses up to $k$ directed links to connect to other nodes with an objective to minimize its average distance, using weights $w_{vu}$, to all other destinations. We study the existence of pure Nash equilibria for $(n,k)$-BDNF games. We show that if the weights are arbitrary, then a pure Nash wiring may not exist. Furthermore, it is NP-hard to determine whether a pure Nash wiring exists for a given $(n,k)$-BDNF instance. A major focus of this paper is on uniform $(n,k)$-BDNF games, in which all weights are 1. We describe how to construct a pure Nash equilibrium wiring given any $n$ and $k$, and establish that in all pure Nash wirings the cost of individual nodes cannot differ by more than a factor of nearly 2, whereas the diameter cannot exceed $O(\sqrt{n \log_k n})$. We also analyze best-response walks on the configuration space defined by the uniform game, and show that starting from any initial configuration, strong connectivity is reached within $Θ(n^2)$ rounds. Convergence to a pure Nash equilibrium, however, is not guaranteed. We present simulation results that suggest that loop-free best-response walks always exist, but may not be polynomially bounded. We also study a special family of \emph{regular} wirings, the class of Abelian Cayley graphs, in which all nodes imitate the same wiring pattern, and show that if $n$ is sufficiently large no such regular wiring can be a pure Nash equilibrium.