Sign In

Communications of the ACM

121 - 130 of 1,487 for bentley

FT-iSort: efficient fault tolerance for introsort

Introspective sorting is a ubiquitous sorting algorithm which underlies many large scale distributed systems. Hardware-mediated soft errors can result in comparison and memory errors, and thus cause introsort to generate incorrect output, which in turn disrupts systems built upon introsort; hence, it is critical to incorporate fault tolerance capability within introsort. This paper proposes the first theoretically-sound, practical fault tolerant introsort with negligible overhead: FT-iSort. To tolerate comparison errors, we use minimal TMR protection via exploiting the properties of the effects of soft errors on introsort. This algorithm-based selective protection incurs far less overhead than naïve TMR protection designed to protect an entire application. To tolerate memory errors that escape DRAM error correcting code, we propose XOR-based re-execution. We incorporate our fault tolerance method into the well-known parallel sorting implementation HykSort, and we find that fault tolerant HykSort incurs negligible overhead and obtains nearly the same scalability as unprotected HykSort.

Self-Adapting the Brownian Radius in a Differential Evolution Algorithm for Dynamic Environments

Several algorithms aimed at dynamic optimisation problems have been developed. This paper reports on the incorporation of a self-adaptive Brownian radius into competitive differential evolution (CDE). Four variations of a novel technique to achieving the self-adaptation is suggested and motivated. An experimental investigation over a large number of benchmark instances is used to determine the most effective of the four variations. The new algorithm is compared to its base algorithm on an extensive set of benchmark problems and its performance analysed. Finally, the new algorithm is compared to other algorithms by means of reported results found in the literature. The results indicate that CDE is improved the the incorporation of the self-adaptive Brownian radius and that the new algorithm compares well with other algorithms.

Dynamic Data Structures for Document Collections and Graphs

In the dynamic indexing problem, we must maintain a changing collection of text documents so that we can efficiently support insertions, deletions, and pattern matching queries. We are especially interested in developing efficient data structures that store and query the documents in compressed form. All previous compressed solutions to this problem rely on answering rank and select queries on a dynamic sequence of symbols. Because of the lower bound in [Fredman and Saks, 1989], answering rank queries presents a bottleneck in compressed dynamic indexing. In this paper we show how this lower bound can be circumvented using our new framework. We demonstrate that the gap between static and dynamic variants of the indexing problem can be almost closed. Our method is based on a novel framework for adding dynamism to static compressed data structures. Our framework also applies more generally to dynamizing other problems. We show, for example, how our framework can be applied to develop compressed representations of dynamic graphs and binary relations.

Reveal: Investigating Proactive Location-Based Reminiscing with Personal Digital Photo Repositories

Recording experiences and memories is an important role for digital photography, with smartphone cameras leading to individuals taking increasing numbers of pictures of everyday experiences. Increasingly, these are automatically stored in personal, cloud-backed, photo repositories. However, such experiences can be forgotten quickly, with images 'lost' within the user's library, loosing their role in supporting reminiscing. We investigate how users might be provoked to view these images and the benefits they bring through the development and evaluation of a proactive, location-based reminiscing tool, called Reveal. We outline how a location-based approach allowed participants to reflect more widely on their photo practice, and the potential of such reminiscing tools to support effective management and curation of individual's increasingly large personal photo collections.

MessageOnTap: A Suggestive Interface to Facilitate Messaging-related Tasks

Text messages are sometimes prompts that lead to information related tasks, e.g. checking one's schedule, creating reminders, or sharing content. We introduce MessageOnTap, a suggestive inter-face for smartphones that uses the text in a conversation to suggest task shortcuts that can streamline likely next actions. When activated, MessageOnTap uses word embeddings to rank relevant external apps, and parameterizes associated task shortcuts using key phrases mentioned in the conversation, such as times, persons, or events. MessageOnTap also tailors the auto-complete dictionary based on text in the conversation, to streamline any text input.We first conducted a month-long study of messaging behaviors(N=22) that informed our design. We then conducted a lab study to evaluate the effectiveness of MessageOnTap's suggestive interface, and found that participants can complete tasks 3.1x faster withMessageOnTap than their typical task flow.

Sochiatrist: Signals of Affect in Messaging Data

Messaging is a common mode of communication, with conversations written informally between individuals. Interpreting emotional affect from messaging data can lead to a powerful form of reflection or act as a support for clinical therapy. Existing analysis techniques for social media commonly use LIWC and VADER for automated sentiment estimation. We correlate LIWC, VADER, and ratings from human reviewers with affect scores from 25 participants. We explore differences in how and when each technique is successful. Results show that human review does better than VADER, the best automated technique, when humans are judging positive affect ($r_s=0.45$ correlation when confident, $r_s=0.30$ overall). Surprisingly, human reviewers only do slightly better than VADER when judging negative affect ($r_s=0.38$ correlation when confident, $r_s=0.29$ overall). Compared to prior literature, VADER correlates more closely with PANAS scores for private messaging than public social media. Our results indicate that while any technique that serves as a proxy for PANAS scores has moderate correlation at best, there are some areas to improve the automated techniques by better considering context and timing in conversations.

Smartwatch in vivo

In recent years, the smartwatch has returned as a form factor for mobile computing with some success. Yet it is not clear how smartwatches are used and integrated into everyday life differently from mobile phones. For this paper, we used wearable cameras to record twelve participants' daily use of smartwatches, collecting and analysing incidents where watches were used from over 34 days of user recording. This allows us to analyse in detail 1009 watch uses. Using the watch as a timepiece was the most common use, making up 50% of interactions, but only 14% of total watch usage time. The videos also let us examine why and how smartwatches are used for activity tracking, notifications, and in combination with smartphones. In discussion, we return to a key question in the study of mobile devices: how are smartwatches integrated into everyday life, in both the actions that we take and the social interactions we are part of?

Challenges and opportunities in data mining contact lists for inferring relationships

The smartphone contact list has the potential to be a valuable source of data about personal relationships. To understand how we might data mine the information that people store in their contact lists, we collected the contact lists of 54 participants. Initially we found that the majority of contact list features were unused. However, a further examination of the "name" field revealed a broad variety of contact-naming behaviors. We observed contact "name" fields that included affiliations, relationship role labels, multiple names, phone types, and references to companies/services/places. People's appropriation and usage of contact lists have implications for automated attempts to merge or mine contact lists that assume people use the features and structure of the contact list tool as intended. They also offer new opportunities for data mining to better describe relationships between users and their contacts.

Privacy in Context: Critically Engaging with Theory to Guide Privacy Research and Design

Privacy has been a key research theme in the CSCW and HCI communities, but the term is often used in an ad hoc and fragmented way. This is likely due to the fact that privacy is a complex and multi-faceted concept. This one-day workshop will facilitate discourse around key privacy theories and frameworks that can inform privacy research with the goal of producing guidelines for privacy researchers on how and when to incorporate which theories into various aspects of their empirical privacy research. This will lay the groundwork to move the privacy field forward.

Bridging a Bridge: Bringing Two HCI Communities Together

ACM SIGCHI is the largest association for professionals in HCI that bridges computer science, information science, as well as the social and psychological sciences. Meanwhile, a parallel HCI community was formed in 2001 within the Association of Information Systems (AIS SIGHCI) community. While some researchers have already bridged these two HCI sub-disciplines, the history and core values of these respective fields are quite different, offering new insights for how we can move forward together to sustain the future of HCI research. The main goal of this workshop is to begin building a bridge between these two communities to maximize the relevance, rigor, and generalizability of HCI research.