Janani Kalyanam

Leveraging Social Context for Modeling Topic Evolution

Topic discovery and evolution (TDE) has been a problem which has gained long standing interest in the research community. The goal in topic discovery is to identify groups of keywords from large corpora so that the information in those corpora are summarized succinctly. The nature of text corpora has changed dramatically in the past few years with the advent of social media. Social media services allow users to constantly share, follow and comment on posts from other users. Hence, such services have given a new dimension to the traditional text corpus. The new dimension being that todays corpora have a social context embedded in them in terms of the community of users interested in a particular post, their profiles etc. We wish to harness this social context that comes along with the textual content for TDE. In particular, our goal is to both qualitatively and quantitatively analyze when social context actually helps with TDE. Methodologically, we approach the problem of TDE by a proposing non-negative matrix factorization (NMF) based model that incorporates both the textual information and social context information. We perform experiments on large scale real world dataset of news articles, and use Twitter as the platform providing information about the social context of these news articles. We compare with and outperform several state-of-the-art baselines. Our conclusion is that using the social context information is most useful when faced with topics that are particularly difficult to detect.

Identification and compensation of Wiener-Hammerstein systems with feedback

Efficient operation of RF power amplifiers requires compensation strategies to mitigate nonlinear behavior. As bandwidth increases, memory effects become more pronounced, and Volterra series based compensation becomes onerous due to the exponential growth in the number of necessary coefficients. Behavioral models such as Wiener-Hammerstein systems with a parallel feedforward or feedback filter are more tractable but more difficult to identify. In this paper, we extend a Wiener-Hammerstein identification method to such systems showing that identification is possible (up to inherent model ambiguities) from single- and two-tone measurements. We also calculate the Cramér-Rao bound for the system parameters and compare to our identification method in simulation. Finally, we demonstrate equalization performance using measured data from a wideband GaN power amplifier.

Twitter-Based Detection of Illegal Online Sale of Prescription Opioid

Objectives To deploy a methodology accurately identifying tweets marketing the illegal online sale of controlled substances. Methods We first collected tweets from the Twitter public application program interface stream filtered for prescription opioid keywords. We then used unsupervised machine learning (specifically, topic modeling) to identify topics associated with illegal online marketing and sales. Finally, we conducted Web forensic analyses to characterize different types of online vendors. We analyzed 619 937 tweets containing the keywords codeine, Percocet, fentanyl, Vicodin, Oxycontin, oxycodone, and hydrocodone over a 5-month period from June to November 2015. Results A total of 1778 tweets (< 1%) were identified as marketing the sale of controlled substances online; 90% had imbedded hyperlinks, but only 46 were “live” at the time of the evaluation. Seven distinct URLs linked to Web sites marketing or illegally selling controlled substances online. Conclusions Our methodology can identify illegal online sale of prescription opioids from large volumes of tweets. Our results indicate that controlled substances are trafficked online via different strategies and vendors. Public Health Implications Our methodology can be used to identify illegal online sellers in criminal violation of the Ryan Haight Online Pharmacy Consumer Protection Act.

Prediction and Characterization of High-Activity Events in Social Media Triggered by Real-World News

On-line social networks publish information on a high volume of real-world events almost instantly, becoming a primary source for breaking news. Some of these real-world events can end up having a very strong impact on on-line social networks. The effect of such events can be analyzed from several perspectives, one of them being the intensity and characteristics of the collective activity that it produces in the social platform. We research 5,234 real-world news events encompassing 43 million messages discussed on the Twitter microblogging service for approximately 1 year. We show empirically that exogenous news events naturally create collective patterns of bursty behavior in combination with long periods of inactivity in the network. This type of behavior agrees with other patterns previously observed in other types of natural collective phenomena, as well as in individual human communications. In addition, we propose a methodology to classify news events according to the different levels of intensity in activity that they produce. In particular, we analyze the most highly active events and observe a consistent and strikingly different collective reaction from users when they are exposed to such events. This reaction is independent of an event’s reach and scope. We further observe that extremely high-activity events have characteristics that are quite distinguishable at the beginning stages of their outbreak. This allows us to predict with high precision, the top 8% of events that will have the most impact in the social network by just using the first 5% of the information of an event’s lifetime evolution. This strongly implies that high-activity events are naturally prioritized collectively by the social network, engaging users early on, way before they are brought to the mainstream audience.

Learning from unstructured multimedia data

Information in todays world is highly heterogeneous and unstructured. Learning and inferring from such data is challenging and is an active research topic. In this paper, we present and investigate an approach to learning from heterogeneous and unstructured multimedia data. Inspired by approaches in many fields including computer vision, we investigate a histogram based approach to represent multimodal unstructured data. While existing works have predominantly focused on histogram based approaches for unimodal data, we present a methodology to represent unstructured multimodal data. We explain how to discover the prototypical features or codewords over which these histograms are built. We present experimental results on classification and retrieval tasks performed on the histogram based representation.

Detection of illicit online sales of fentanyls via Twitter

A counterfeit fentanyl crisis is currently underway in the United States. Counterfeit versions of commonly abused prescription drugs laced with fentanyl are being manufactured, distributed, and sold globally, leading to an increase in overdose and death in countries like the United States and Canada. Despite concerns from the U.S. Drug Enforcement Agency regarding covert and overt sale of fentanyls online, no study has examined the role of the Internet and social media on fentanyl illegal marketing and direct-to-consumer access. In response, this study collected and analyzed five months of Twitter data (from June-November 2015) filtered for the keyword “fentanyl” using Amazon Web Services. We then analyzed 28,711 fentanyl-related tweets using text filtering and a machine learning approach called a Biterm Topic Model (BTM) to detect underlying latent patterns or “topics” present in the corpus of tweets. Using this approach we detected a subset of 771 tweets marketing the sale of fentanyls online and then filtered this down to nine unique tweets containing hyperlinks to external websites. Six hyperlinks were associated with online fentanyl classified ads, 2 with illicit online pharmacies, and 1 could not be classified due to traffic redirection. Importantly, the one illicit online pharmacy detected was still accessible and offered the sale of fentanyls and other controlled substances direct-to-consumers with no prescription required at the time of publication of this study. Overall, we detected a relatively small sample of Tweets promoting illegal online sale of fentanyls. However, the detection of even a few online sellers represents a public health danger and a direct violation of law that demands further study.

A Review of Digital Surveillance Methods and Approaches to Combat Prescription Drug Abuse

Purpose of ReviewThe use of social media to conduct digital surveillance to address different health challenges is growing. This multidisciplinary review assesses the current state of methods and applied research used to conduct digital surveillance for prescription drug abuse.Recent FindingsFifteen studies met our inclusion criteria from the databases reviewed (PubMed, IEEE Xplore, and ACM Digital Library). The articles were characterized based on their overarching goals and aims, data collection and dataset attributes, and analysis approaches. Overall, reviewed studies grouped into two overarching categories as either being method-focused (advancing novel methodologies using social media data), applied-focused (generating new information on prescription drug abuse behavior), or having both elements. The social media platform most predominantly used was Twitter, with wide variation in sample size and duration of data collection. Several data analysis strategies were employed, including machine learning, temporal analysis, rule-based approaches, and statistical analysis.SummaryOur review indicates that the field of prescription drug abuse digital surveillance is still maturing. Though many studies captured large volumes of data, the majority did not analyze data to characterize user behavior, a critical step needed in order to better explain the underlying risk environment for prescription drug abuse. Future studies need to better translate method-based approaches into applied research, use data generated from social media platforms other than Twitter, and take advantage of emerging data analysis strategies, including deep learning and multimodal approaches.

Physical layer considerations for wideband cognitive radio

Next generation cognitive radios will benefit from the capability of transmitting and receiving communications waveforms across many disjoint frequency channels spanning hundreds of megahertz of bandwidth. The information theoretic advantages of multi-channel operation for cognitive radio (CR), however, come at the expense of stringent linearity requirements on the analog transmit and receive hardware. This paper presents the quantitative advantages of multi-channel operation for next generation CR, and the advanced digital compensation algorithms to mitigate transmit and receive nonlinearities that enable broadband multi-channel operation. Laboratory measurements of the improvement in the performance of a multi-channel CR communications system operating below 2 GHz in over 500 MHz of instantaneous bandwidth are presented.

From event detection to storytelling on microblogs

The problem of detecting events from content published on microblogs has garnered much interest in recent times. In this paper, we address the questions of what happens after the outbreak of an event in terms of how the event gradually progresses and attains each of its milestones, and how it eventually dissipates. We propose a model based approach to capture the gradual unfolding of an event over time. This enables the model to automatically produce entire timeline trajectories of events from the time of their outbreak to their disappearance. We apply our model on the Twitter messages collected about Ebola during the 2014 outbreak and obtain the progression timelines of several events that occurred during the outbreak. We also compare our model to several existing topic modeling and event detection baselines in literature to demonstrate its efficiency.