Search

Estimation of query difficulty is an attempt to quantify the quality of results returned by the search system for a query. Ideally, a system that can predict difficult queries can adapt parameters or change algorithms to suit the query. Such systems could give feedback to the user, for example, by reporting confidence scores for results, and report to the system administrator regarding topics that are of increasing interest to users but are not answered well by the system.

Brain-computer interfaces are devices for enabling patients with severe motor disorders to communicate with the world. One method for operating such devices is to use movement-related potentials that are generated in the brain when the patient moves, or imagines a movement of, one of his limbs. An algorithm for detecting movement-related potentials using a small number of EEG channels was developed. This algorithm is a combination of the matched filter, a non-linear transformation previously developed as part of a similar detector, and a classifier. The algorithm was compared with previous designs of similar detectors by both theoretic analysis and off-line evaluation of performance on data recorded from five subjects. It is shown that the performance of the algorithm was superior to that of previous methods, improving the area under the receiver operating characteristic curve to 87.8%, an improvement of 25% compared with the best previously suggested detection method. Finally, the probable sources for false detections were identified, and possible ways to minimise them are proposed.

Classification of movement-related potentials recorded from the scalp to their corresponding limb is a crucial task in brain-computer interfaces based on such potentials. Many features can be extracted from raw electroencephalographic signals to be used for classification, but the utilization of irrelevant or superfluous features is detrimental to the performance of classification algorithms. It is, therefore, necessary to select a small number of relevant features for the classification task. This paper demonstrates the use of two feature selection methods to choose a small number (10-20) of relevant features from a bank containing upward of 1000 features. One method is based on information theory and the other on the use of genetic algorithms. We show that the former is poorly suited for the aforementioned classification task and discuss the probable reasons for this. However, using a genetic algorithm on data recorded from five subjects we demonstrate that it is possible to differentiate between the movements of two limbs with a classification accuracy of 87% using as little as 10 features without subject training. With the addition of a simple coding scheme, this method can be applied to multiple limb classification and a 63% classification accuracy rate can be reached when attempting to distinguish between three limbs.

Background: Cannabis is one of the most widely used psychoactive substances worldwide, but adverse drug reactions (ADRs) associated with its use are difficult to study because of its prohibited status in many countries. Objective: Internet search engine queries have been used to investigate ADRs in pharmaceutical drugs. In this proof-of-concept study, we tested whether these queries can be used to detect the adverse reactions of cannabis use. Methods: We analyzed anonymized queries from US-based users of Bing, a widely used search engine, made over a period of 6 months and compared the results with the prevalence of cannabis use as reported in the US National Survey on Drug Use in the Household (NSDUH) and with ADRs reported in the Food and Drug Administration’s Adverse Drug Reporting System. Predicted prevalence of cannabis use was estimated from the fraction of people making queries about cannabis, marijuana, and 121 additional synonyms. Predicted ADRs were estimated from queries containing layperson descriptions to 195 ICD-10 symptoms list. Results: Our results indicated that the predicted prevalence of cannabis use at the US census regional level reaches an R2 of .71 NSDUH data. Queries for ADRs made by people who also searched for cannabis reveal many of the known adverse effects of cannabis (eg, cough and psychotic symptoms), as well as plausible unknown reactions (eg, pyrexia). Conclusions: These results indicate that search engine queries can serve as an important tool for the study of adverse reactions of illicit drugs, which are difficult to study in other settings.

Background: Regular physical activity is known to be beneficial for people with type 2 diabetes. Nevertheless, most of the people who have diabetes lead a sedentary lifestyle. Smartphones create new possibilities for helping people to adhere to their physical activity goals through continuous monitoring and communication, coupled with personalized feedback. Objective: The aim of this study was to help type 2 diabetes patients increase the level of their physical activity. Methods: We provided 27 sedentary type 2 diabetes patients with a smartphone-based pedometer and a personal plan for physical activity. Patients were sent short message service messages to encourage physical activity between once a day and once per week. Messages were personalized through a Reinforcement Learning algorithm so as to improve each participant’s compliance with the activity regimen. The algorithm was compared with a static policy for sending messages and weekly reminders. Results: Our results show that participants who received messages generated by the learning algorithm increased the amount of activity and pace of walking, whereas the control group patients did not. Patients assigned to the learning algorithm group experienced a superior reduction in blood glucose levels (glycated hemoglobin [HbA1c]) compared with control policies, and longer participation caused greater reductions in blood glucose levels. The learning algorithm improved gradually in predicting which messages would lead participants to exercise. Conclusions: Mobile phone apps coupled with a learning algorithm can improve adherence to exercise in diabetic patients. This algorithm can be used in large populations of diabetic patients to improve health and glycemic control. Our results can be expanded to other areas where computer-led health coaching of humans may have a positive impact. Summary of a part of this manuscript has been previously published as a letter in Diabetes Care, 2016. [J Med Internet Res 2017;19(10):e338]

The social influence of people on their peers in the selection of products and services is frequently modeled as a diffusion process. Recently, such processes have been successfully applied as a tool for predicting customer turnover, or churn, in mobile communication carriers. These predictions are most accurate when specific social ties are used in the diffusion process, and are primarily useful when they provide a long forecast horizon, so as to enable a service provider to take mitigating actions. Here, we investigate several measures of social affinity and compare their performances for churn prediction, using data from two large mobile phone carriers. Our analysis demonstrates that the various measures of social ties capture different calling and texting patterns, and that a significant improvement in the accuracy of prediction is reached by combining them. We study the predictive horizon of diffusion processes and show that it deteriorates significantly as the horizon increases. Our findings underline the usefulness of diffusion processes for enhancing churn prediction while providing insights to their limitations.

BackgroundMood disorders affect a significant portion of the general population. Cycling mood disorders are characterized by intermittent episodes (or events) of the disease. ObjectiveUsing anonymized Web search logs, we identify a population of people with significant interest in mood stabilizing drugs (MSD) and seek evidence of mood swings in this population. MethodsWe extracted queries to the Microsoft Bing search engine made by 20,046 Web searchers over six months, separately explored searcher demographics using data from a large external panel of users, and sought supporting information from people with mood disorders via a survey. We analyzed changes in information needs over time relative to searches on MSD. ResultsQueries for MSD focused on side effects and their relation to the disease. We found evidence of significant changes in search behavior and interests coinciding with days that MSD queries are made. These include large increases (>100%) in the access of nutrition information, commercial information, and adult materials. A survey of patients diagnosed with mood disorders provided evidence that repeated queries on MSD may come with exacerbations of mood disorder. A classifier predicting the occurrence of such queries one day before they are observed obtains strong performance (AUC=0.78). ConclusionsObserved patterns in search behavior align with known behaviors and those highlighted by survey respondents. These observations suggest that searchers showing intensive interest in MSD may be patients who have been prescribed these drugs. Given behavioral dynamics, we surmise that the days on which MSD queries are made may coincide with commencement of mania or depression. Although we do not have data on mood changes and whether users have been diagnosed with bipolar illness, we see evidence of cycling in people who show interest in MSD and further show that we can predict impending shifts in behavior and interest.

Online providers frequently offer a variety of services, ranging from news to mail. These providers endeavour to keep users accessing and interacting with the sites offering these services, that is to engage users by spending time on these sites and returning regularly to them. The standard approach to evaluate engagement with a site is by measuring engagement metrics of each site separately. However, when assessing engagement within a network of sites, it is crucial to take into account the traffic between sites. This paper proposes a methodology for studying networked used engagement. We represent sites (nodes) and user traffic (edges) between them as a network, and apply complex network metrics to study networked user engagement. We demonstrate the value of these metrics when applied to 728 services offered by Yahoo!~with a sample of 2M users and a total of 25M online sessions.

Objectives : A method by which potentials related to voluntary movement can be recorded noninvasively from the human spinal cord is presented. Methods : A novel signal processing technique performed on signals recorded by surface electrodes placed over the spinal column was used to filter time-locked back muscle noise, so that the only remaining signals were the spinal movement-related potentials from the brain to the limbs and vice versa. Results : The signals obtained from 7 subjects using this technique are shown and temporally compared with movement-related cortical potentials (MRCP) and muscle electromyogram. It is demonstrated that the spinal signal starts approximately 600 ms before the actual movement, and that some features of this signal correspond to changes in cortical potentials. Conclusions : These findings imply that the spinal cord is not a simple command-carrying medium from the brain to the limbs, and implies that some computational activities take place at the spinal cord level.

This workshop brings together researchers and practitioners from industry and academia to discuss search and discovery in the medi-cal domain. The event focuses on ways to make medical and health information more accessible to laypeople (including enhancements to ranking algorithms and search interfaces), and how we can dis-cover new medical facts and phenomena from information sought online, as evidenced in query streams and other sources such as social media. This domain also offers many opportunities for appli-cations that monitor and improve quality of life of those affected by medical conditions, by providing tools to support their health-related information behavior.

During unexpected events such as natural disasters, individuals rely on the information generated by news outlets to form their understanding of these events. This information, while often voluminous, is frequently degraded by the inclusion of unimportant, duplicate, or wrong information. It is important to be able to present users with only the novel, important information about these events as they develop. We present the problem of updating users about time critical news events, and focus on the task of deciding which information to select for updating users as an event develops. We propose a solution to this problem which incorporates techniques from information retrieval and multi-document summarization and evaluate this approach on a set of historic events using a large stream of news documents. We also introduce an evaluation method which is significantly less expensive than traditional approaches to temporal summarization.

Internet users notoriously take an assumed identity or masquerade as someone else, for reasons such as financial profit or social benefit. But often the converse is also observed, where people choose to reveal true features of their identity, including deeply intimate details. This work attempts to explore several of the conditions that allow this to happen by analyzing the content generated by these users. We examine multiple social media on the Web, specifically focusing on Yahoo! Answers, encompassing more than a billion answers posted since 2006. Our analysis covers discussions of personal topics such as body measurements and income, and of socially sensitive subjects such as sexual behaviors. We offer quantitative proof that people are aware of the fact that they are posting sensitive information, and yet provide accurate information to fulfill specific information needs. Our analysis further reveals that on community question answering sites, when users are truthful, their expectation of an accurate answer is met.

In the online world, user engagement refers to the phenomena associated with being captivated by a web application and wanting to use it longer and frequently. Nowadays, many providers operate multiple content sites, very different from each other. Due to their extremely varied content, these are usually studied and optimized separately. However, user engagement should be examined not only within individual sites, but also across sites, that is the entire content provider network. In previous work, we investigated networked user engagement, by defining a global measure of engagement that captures the effect that sites have on the engagement on other sites within the same browsing session. Here, we look at the effect of links on networked user engagement, as these are commonly used by online content providers to increase user engagement.

Our research goal is to provide a better understanding of how users engage with online services, and how to measure this engagement. We should not speak of one main approach to measure user engagement --- e.g. through one fixed set of metrics --- because engagement depends on the online services at hand. Instead, we should be talking of models of user engagement. As a first step, we analysed a number of online services, and show that it is possible to derive effectively simple models of user engagement, for example, accounting for user types and temporal aspects. This paper provides initial insights into engagement patterns, allowing for a better understanding of the important characteristics of how users repeatedly interact with a service or group of services.

The information needs of users and the documents which answer it are frequently contingent on the different characteristics of users. This is especially evident during natural disasters, such as earthquakes and violent weather incidents, which create a strong transient information need. In this paper we investigate how the information need of users is affected by their physical detachment, as estimated by their physical location in relation to that of the event, and by their social detachment, as quantified by the number of their acquaintances who may be affected by the event. Drawing on large-scale data from three major events, we show that social and physical detachment levels of users are a major influence on their information needs, as manifested by their search engine queries. We demonstrate how knowing social and physical detachment levels can assist in improving retrieval for two applications: identifying search queries related to events and ranking results in response to event-related queries. We find that the average precision in identifying relevant search queries improves by approximately 18%, and that the average precision of ranking that uses detachment information improves by 10%.

People nowadays can obtain information on current news events through media outlets, social media, and by actively seeking information using search engines. In this paper we investigate the temporal relationship between news coverage by media outlets, social media, and query logs and show that social media frequently precedes other information sources. Additionally, we demonstrate that there is strong negative correlation between the probability for reporting of an event and the distance of the information source from the event.

This work deals with the task of predicting the popularity of user-generated content. We demonstrate how the novelty of newly published content plays an important role in affecting its popularity. More specifically, we study three dimensions of novelty. The first one, termed contemporaneous novelty , models the relative novelty embedded in a new post with respect to contemporary content that was generated by others. The second type of novelty, termed self novelty , models the relative novelty with respect to the user’s own contribution history. The third type of novelty, termed discussion novelty , relates to the novelty of the comments associated by readers with respect to the post content. We demonstrate the contribution of the new novelty measures to estimating blog-post popularity by predicting the number of comments expected for a fresh post. We further demonstrate how novelty based measures can be utilized for predicting the citation volume of academic papers.

We consider the dynamics of rapid propagation of information in complex social networks focusing on mobile phone networks. We introduce two models for an information propagation process. The first model describes the temporal behavior of people which leads to the emergence of information propagation events and is based on the existence of two types of subscribers: regular subscribers and subscribers that tend to spread information. The second model describes the topology of paths in which the information propagates from one subscriber to another. We further introduce an efficient algorithm for identification of information propagation events. We then apply our algorithm to a large-scale mobile phone network and demonstrate the correspondence between theoretical expectations and the actual results.

In this work we propose a novel framework for bookmark weighting which allows us to estimate the effectiveness of each of the bookmarks individually. We show that by weighting bookmarks according to their estimated quality we can significantly improve search effectiveness. Using empirical evaluation on real data gathered from two large bookmarking systems, we demonstrate the effectiveness of the new framework for search enhancement.

Concurrent computer programs are fast becoming prevalent in many critical applications. Unfortunately, these programs are especially difficult to test and debug. Recently, it has been suggested that injecting random timing noise into many points within a program can assist in eliciting bugs within the program. Upon eliciting the bug, it is necessary to identify a minimal set of points that indicate the source of the bug to the programmer. In this paper, we pose this problem as an active feature selection problem. We propose an algorithm called the iterative group sampling algorithm that iteratively samples a lower dimensional projection of the program space and identifies candidate relevant points. We analyze the convergence properties of this algorithm. We test the proposed algorithm on several real-world programs and show its superior performance. Finally, we show the algorithms' performance on a large concurrent program.

This work shows how the content of a digital library can be enhanced to better satisfy its users' needs. Missing content is identified by finding missing content topics in the system's query log or in a pre-defined taxonomy of required knowledge. The collection is then enhanced with new relevant knowledge, which is extracted from external sources that satisfy those missing content topics. Experiments we conducted measure the precision of the system before and after content enhancement. The results demonstrate a significant improvement in the system effectiveness as a result of content enhancement and the superiority of the missing content enhancement policy over several other possible policies.

Heuristics in compilers are often designed by manually analyzing sample programs. Recent advances have successfully applied machine learning to automatically generate heuristics. The typical format of these approaches reduces the input loops, functions or programs to a finite vector of features. A machine learning algorithm then learns a mapping from these features to the desired heuristic parameters. Choosing the right features is important and requires expert knowledge since no machine learning tool will work well with poorly chosen features.This paper introduces a novel mechanism to generate features. Grammars describing languages of features are defined and from these grammars sentences are randomly produced. The features are then evaluated over input data and computed values are given to machine learning tools.We propose the construction of domain specific feature languages for different purposes in different parts of the compiler. Using these feature languages, complex, machine generated features are extracted from program code. Using our observation that some functions can benefit from setting different compiler options, while others cannot, we demonstrate the use of a decision tree classifier to automatically identify the former using the automatically generated features.We show that our method outperform human generated features on problems of loop unrolling and phase ordering, achieving a statistically significant decrease in run-time compared to programs compiled using GCC’s heuristics.

Heterogeneous clusters and grid infrastructures are becoming increasingly popular. In these computing infrastructures, machines have different resources, including memory sizes, disk space, and installed software packages. These differences give rise to a problem of over-provisioning, that is, sub-optimal utilization of a cluster due to users requesting resource capacities greater than what their jobs actually need. Our analysis of a real workload file (LANL CM5) revealed differences of up to two orders of magnitude between requested memory capacity and actual memory usage. This paper presents an algorithm to estimate actual resource capacities used by batch jobs. Such an algorithm reduces the need for users to correctly predict the resources required by their jobs, while at the same time managing the scheduling system to obtain superior utilization of available hardware. The algorithm is based on the Reinforcement Learning paradigm; it learns its estimation policy on-line and dynamically modifies it according to the overall cluster load. The paper includes simulation results which indicate that our algorithm can yield an improvement of over 30% in utilization (overall throughput) of heterogeneous clusters.

International audience; Tuning hardwired compiler optimizations for rapidly evolving hardware makes porting an optimizing compiler for each new platform extremely challenging. Our radical approach is to develop a modular, extensible, self-optimizing compiler that automatically learns the best optimization heuristics based on the behavior of the platform. In this paper we describe MILEPOST GCC, a machine-learning-based compiler that automatically adjusts its optimization heuristics to improve the execution time, code size, or compilation time of specific programs on different architectures. Our preliminary experimental results show that it is possible to considerably reduce execution time of the MiBench benchmark suite on a range of platforms entirely automatically.

As concurrent and distributive applications are becoming more common and debugging such applications is very difficult, practical tools for automatic debugging of concurrent applications are in demand. In previous work, we applied automatic debugging to noise-based testing of concurrent programs. The idea of noise-based testing is to increase the probability of observing the bugs by adding, using instrumentation, timing "noise" to the execution of the program. The technique of finding a small subset of points that causes the bug to manifest can be used as an automatic debugging technique. Previously, we showed that Delta Debugging can be used to pinpoint the bug location on some small programs.In the work reported in this paper, we create and evaluate two algorithms for automatically pinpointing program locations that are in the vicinity of the bugs on a number of industrial programs. We discovered that the Delta Debugging algorithms do not scale due to the non-monotonic nature of the concurrent debugging problem. Instead we decided to try a machine learning feature selection algorithm. The idea is to consider each instrumentation point as a feature, execute the program many times with different instrumentations, and correlate the features (instrumentation points) with the executions in which the bug was revealed. This idea works very well when the bug is very hard to reveal using instrumentation, correlating to the case when a very specific timing window is needed to reveal the bug. However, in the more common case, when the bugs are easy to find using instrumentation points ranked high by the feature selection algorithm is not high enough. We show that for these cases, the important value is not the absolute value of the evaluation of the feature but the derivative of that value along the program execution path.As a number of groups expressed interest in this research, we built an open infrastructure for automatic debugging algorithms for concurrent applications, based on noise injection based concurrent testing using instrumentation. The infrastructure is described in this paper.

This paper describes a new discrete event simulation model built using a mathematical tool (Matlab) to investigate the simulation of the programming and the testing phases of a software development project. In order to show how the model can be used and to provide some preliminary concrete results, we give three examples of how this model can be utilized to examine the effect of adopting different strategies for coding and testing a new software system. Specifically, we provide results of simulation runs intended to simulate the effects on the coding and testing phases of different testing strategies, the adoption of pair programming in an otherwise-unchanged process, and the automation of testing. The model source code is available for downloading at http://qp.research.ibm.com/concurrency_testing, and we invite researchers and practitioners to use and modify the model.

The lynx Lynx canadensis is a common predator in the boreal forests of North America. Its population fluctuates during a 9- to 11-year cycle in synchrony with the population size of its main prey, the snowshoe hare Lepus americanus. Using adult museum specimens, we studied changes in skull (and hence body) size of the lynx in Alaska during the second half of the 20th century. The population cycle in Alaska averaged 9 years, similar to that reported in the neighbouring Yukon. Using harvest data of lynx as an estimate of population size, we found that skull size was negatively related to population size. This relationship was strongest not for the population density in the year of death (X), but for year X-3, a carry-over effect from the first year (or years) of life, indicating that conditions during the fast-growth years are determining body size. We suggest that the density-dependent effect is probably due to changes in food supply, either resulting from the adverse effects of competition or a possible diminished availability of food. Two skull parameters decreased significantly during the second half of the 20th century. We do not know the cause for the year effect and suggest that it might be due to a long-term change in the availability of prey. Canine size did not change during the study period, probably an indication that snowshoe hares maintained their status as the main prey of the lynx throughout the study period.

In this article we present novel learning methods for estimating the quality of results returned by a search engine in response to a query. Estimation is based on the agreement between the top results of the full query and the top results of its sub-queries. We demonstrate the usefulness of quality estimation for several applications, among them improvement of retrieval, detecting queries for which no relevant content exists in the document collection, and distributed information retrieval. Experiments on TREC data demonstrate the robustness and the effectiveness of our learning algorithms.

A brain-computer interface (BCI) is a system for direct communication between brain and computer. The BCI developed in this work is based on a BCI described by Farwell and Donchin in 1988, which allows a subject to communicate one of 36 symbols presented on a 6 /spl times/ 6 matrix. The system exploits the P300 component of event-related brain potentials (ERP) as a medium for communication. The processing methods distinguish this work from Donchin's work. In this work, independent component analysis (ICA) was used to separate the P300 source from the background noise. A matched filter was used together with averaging and threshold techniques for detecting the existence of P300s. The processing method was evaluated offline on data recorded from six healthy subjects. The method achieved a communication rate of 5.45 symbols/min with an accuracy of 92.1% compared to 4.8 symbols/min with an accuracy of 90% in Donchin's work. The online interface was tested with the same six subjects. The average communication rate achieved was 4.5 symbols/min with an accuracy of 79.5% as apposed to the 4.8 symbols/min with an accuracy of 56% in Donchin's work. The presented BCI achieves excellent performance compared to other existing BCIs, and allows a reasonable communication rate, while maintaining a low error rate.

Pattern classification is the field devoted to the study of methods designed to categorize data into distinct classes. This categorization can be either distinct labeling of the data (supervised learning), division of the data into classes (unsupervised learning), selection of the most significant features of the data (feature selection), or a combination of more than one of these tasks.

Hand amputees would highly benefit from a robotic prosthesis, which would allow the movement of a number of fingers. In this paper we propose using the electromyographic signals recorded by two pairs of electrodes placed over the arm for operating such prosthesis. Multiple features from these signals are extracted whence the most relevant features are selected by a genetic algorithm as inputs for a simple classifier. This method results in a probability of error of less than 2%.

We present a case study of involuntary motion in one part of the body induced by a voluntary motion in another part. The case investigated was that of a steady-state motion of the fingers at various frequencies, which induced a similar kind of motion in the legs. These phenomena were observed in a subject practicing sphincter gymnastics, a physical therapy method in which repetitive motion of different muscles is exercised, and is known to induce motions in other muscles of the body. We show that a linear model, for which we give an experimental transfer function and demonstrate its validity, can describe the generation of the induced periodic motions by the voluntary motions. We compare the results of the transfer function with those of Von Holst's statistical analysis.

BackgroundPostmarket drug safety surveillance largely depends on spontaneous reports by patients and health care providers; hence, less common adverse drug reactions—especially those caused by long-term exposure, multidrug treatments, or those specific to special populations—often elude discovery. ObjectiveHere we propose a low cost, fully automated method for continuous monitoring of adverse drug reactions in single drugs and in combinations thereof, and demonstrate the discovery of heretofore-unknown ones. MethodsWe used aggregated search data of large populations of Internet users to extract information related to drugs and adverse reactions to them, and correlated these data over time. We further extended our method to identify adverse reactions to combinations of drugs. ResultsWe validated our method by showing high correlations of our findings with known adverse drug reactions (ADRs). However, although acute early-onset drug reactions are more likely to be reported to regulatory agencies, we show that less acute later-onset ones are better captured in Web search queries. ConclusionsOur method is advantageous in identifying previously unknown adverse drug reactions. These ADRs should be considered as candidates for further scrutiny by medical regulatory authorities, for example, through phase 4 trials.

BackgroundThere is widespread use of the Internet to promote anorexia as a lifestyle choice. Pro-anorexia content can be harmful for people affected or at risk of having anorexia. That movement is actively engaged in sharing photos on social networks such as Flickr. ObjectiveTo study the characteristics of the online communities engaged in disseminating content that encourages eating disorders (known as “pro-anorexia”) and to investigate if the posting of such content is discouraged by the posting of recovery-oriented content. MethodsThe extraction of pro-anorexia and pro-recovery photographs from the photo sharing site Flickr pertaining to 242,710 photos from 491 users and analyzing four separate social networks therein. ResultsPro-anorexia and pro-recovery communities interact to a much higher degree among themselves than what is expected from the distribution of contacts (only 59-72% of contacts but 74-83% of comments are made to members inside the community). Pro-recovery users employ similar words to those used by pro-anorexia users to describe their photographs, possibly in order to ensure that their content appears when pro-anorexia users search for images. Pro-anorexia users who are exposed to comments from the opposite camp are less likely to cease posting pro-anorexia photographs than those who do not receive such comments (46% versus 61%), and if they cease, they do so approximately three months later. Our observations show two highly active communities, where most interaction is within each community. However, the pro-recovery community takes steps to ensure that their content is visible to the pro-anorexia community, both by using textual descriptions of their photographs that are similar to those used by the pro-anorexia group and by commenting to pro-anorexia content. The latter activity is, however, counterproductive, as it entrenches pro-anorexia users in their stance. ConclusionsOur results highlight the nature of pro-anorexia and pro-recovery photo sharing and accentuate the need for clinicians to be aware of such content and its effect on their patients. Our findings suggest that some currently used interventions are not useful in helping pro-anorexia users recover. Thus, future work should focus on new intervention methods, possibly tailored to individual characteristics.

Although traditionally the primary information sources for cancer patients have been the treating medical team, patients and their relatives increasingly turn to the Internet, though this source may be misleading and confusing. We assess Internet searching patterns to understand the information needs of cancer patients and their acquaintances, as well as to discern their underlying psychological states. We screened 232,681 anonymous users who initiated cancer-specific queries on the Yahoo Web search engine over three months, and selected for study users with high levels of interest in this topic. Searches were partitioned by expected survival for the disease being searched. We compared the search patterns of anonymous users and their contacts. Users seeking information on aggressive malignancies exhibited shorter search periods, focusing on disease- and treatment-related information. Users seeking knowledge regarding more indolent tumors searched for longer periods, alternated between different subjects, and demonstrated a high interest in topics such as support groups. Acquaintances searched for longer periods than the proband user when seeking information on aggressive (compared to indolent) cancers. Information needs can be modeled as transitioning between five discrete states, each with a unique signature representing the type of information of interest to the user. Thus, early phases of information-seeking for cancer follow a specific dynamic pattern. Areas of interest are disease dependent and vary between probands and their contacts. These patterns can be used by physicians and medical Web site authors to tailor information to the needs of patients and family members.

Users of social networks often focus on specific areas of that network, leading to the well-known "filter bubble" effect. Connecting people to a new area of the network in a way that will cause them to become active in that area could help alleviate this effect and improve social welfare. Here we present preliminary analysis of network referrals, that is, attempts by users to connect peers to other areas of the network. We classify these referrals by their efficiency, i.e., the likelihood that a referral will result in a user becoming active in the new area of the network. We show that by using features describing past experience of the referring author and the content of their messages we are able to predict whether referral will be effective, reaching an AUC of 0.87 for those users most experienced in writing efficient referrals. Our results represent a first step towards algorithmically constructing efficient referrals with the goal of mitigating the "filter bubble" effect pervasive in on line social networks., Comment: Accepted to the 2018 Web conference (WWW2018)