Sainyam Galhotra

Fairness testing: testing software for discrimination

This paper defines software fairness and discrimination and develops a testing-based method for measuring if and how much software discriminates, focusing on causality in discriminatory behavior. Evidence of software discrimination has been found in modern software systems that recommend criminal sentences, grant access to financial products, and determine who is allowed to participate in promotions. Our approach, Themis, generates efficient test suites to measure discrimination. Given a schema describing valid system inputs, Themis generates discrimination tests automatically and does not require an oracle. We evaluate Themis on 20 software systems, 12 of which come from prior work with explicit focus on avoiding discrimination. We find that (1) Themis is effective at discovering software discrimination, (2) state-of-the-art techniques for removing discrimination from algorithms fail in many situations, at times discriminating against as much as 98% of an input subdomain, (3) Themis optimizations are effective at producing efficient test suites for measuring discrimination, and (4) Themis is more efficient on systems that exhibit more discrimination. We thus demonstrate that fairness testing is a critical aspect of the software development cycle in domains with possible discrimination and provide initial tools for measuring software discrimination.

Debunking the Myths of Influence Maximization: An In-Depth Benchmarking Study

Influence maximization (IM) on social networks is one of the most active areas of research in computer science. While various IM techniques proposed over the last decade have definitely enriched the field, unfortunately, experimental reports on existing techniques fall short in validity and integrity since many comparisons are not based on a common platform or merely discussed in theory. In this paper, we perform an in-depth benchmarking study of IM techniques on social networks. Specifically, we design a benchmarking platform, which enables us to evaluate and compare the existing techniques systematically and thoroughly under identical experimental conditions. Our benchmarking results analyze and diagnose the inherent deficiencies of the existing approaches and surface the open challenges in IM even after a decade of research. More fundamentally, we unearth and debunk a series of myths and establish that there is no single state-of-the-art technique in IM. At best, a technique is the state of the art in only one aspect.

ASIM: A Scalable Algorithm for Influence Maximization under the Independent Cascade Model

The steady growth of graph data from social networks has resulted in wide-spread research in finding solutions to the influence maximization problem. Although, TIM is one of the fastest existing algorithms, it cannot be deemed scalable owing to its exorbitantly high memory footprint.cIn this paper, we address the scalability aspect -- memory consumption and running time of the influence maximization problem. We propose ASIM, a scalable algorithm capable of running within practical compute times on commodity hardware. Empirically, ASIM is

Holistic Influence Maximization: Combining Scalability and Efficiency with Opinion-Aware Models

6-8

Optimal Radius for Connectivity in Duty-Cycled Wireless Sensor Networks

times faster when compared to CELF++ with similar memory consumption, while its memory footprint is

STAR: Real-time Spatio-Temporal Analysis and Prediction of Traffic Insights using Social Media

\approx 200

Diffusion driven instability to a drift driven one: Turing patterns in the presence of an electric field

times smaller when compared to TIM.

Turing-Hopf instabilities through a combination of diffusion, advection, and finite size effects

The steady growth of graph data from social networks has resulted in wide-spread research in finding solutions to the influence maximization problem. In this paper, we propose a holistic solution to the influence maximization (IM) problem. (1) We introduce an opinion-cum-interaction (OI) model that closely mirrors the real-world scenarios. Under the OI model, we introduce a novel problem of Maximizing the Effective Opinion (MEO) of influenced users. We prove that the MEO problem is NP-hard and cannot be approximated within a constant ratio unless P=NP. (2) We propose a heuristic algorithm OSIM to efficiently solve the MEO problem. To better explain the OSIM heuristic, we first introduce EaSyIM - the opinion-oblivious version of OSIM, a scalable algorithm capable of running within practical compute times on commodity hardware. In addition to serving as a fundamental building block for OSIM, EaSyIM is capable of addressing the scalability aspect - memory consumption and running time, of the IM problem as well. Empirically, our algorithms are capable of maintaining the deviation in the spread always within 5% of the best known methods in the literature. In addition, our experiments show that both OSIM and EaSyIM are effective, efficient, scalable and significantly enhance the ability to analyze real datasets.

Optimal radius for connectivity in duty-cycled wireless sensor networks

We investigate the condition on transmission radius needed to achieve connectivity in duty-cycled wireless sensor networks (briefly, DC-WSNs). First, we settle a conjecture of Das et al. [2012] and prove that the connectivity condition on random geometric graphs (RGGs), given by Gupta and Kumar [1989], can be used to derive a weakly sufficient condition to achieve connectivity in DC-WSNs. To find a stronger result, we define a new vertex-based random connection model that is of independent interest. Following a proof technique of Penrose [1991], we prove that when the density of the nodes approaches infinity, then a finite component of size greater than 1 exists with probability 0 in this model. We use this result to obtain an optimal condition on node transmission radius that is both necessary and sufficient to achieve connectivity and is hence optimal. The optimality of such a radius is also tested via simulation for two specific duty-cycle schemes, called the contiguous and the random selection duty-cycle schemes. Finally, we design a minimum-radius duty-cycling scheme that achieves connectivity with a transmission radius arbitrarily close to the one required in random geometric graphs. The overhead in this case is that we have to spend some time computing the schedule.

Tracking the conductance of rapidly evolving topic-subgraphs

The steady growth of data from social networks has resulted in wide-spread research in a host of application areas including transportation, health-care, customer-care and many more. Owing to the ubiquity and popularity of transportation (more recently) the growth in the number of problems reported by the masses has no bounds. With the advent of social media, reporting problems has become easier than before. In this paper, we address the problem of efficient management of transportation related woes by leveraging the information provided by social media sources such as -- Facebook, Twitter etc. We develop techniques for viral event detection, identify frequently co-occurring problem patterns and their root-causes and mine suggestions to solve the identified problems. We predict the occurrence of different problems, (with an accuracy of ≈ 80%) at different locations and times leveraging the analysis done above along with weather information and news reports. In addition, we design a feature-packed visualization that significantly enhances the ability to analyse data in real-time.