• Corpus ID: 59729978

Real-Time Detection of Task Switches of Desktop Users

@inproceedings{Shen2007RealTimeDO,
  title={Real-Time Detection of Task Switches of Desktop Users},
  author={Jianqiang Shen and Lida Li and Thomas G. Dietterich},
  booktitle={IJCAI},
  year={2007}
}
Desktop users commonly work on multiple tasks. The TaskTracer system provides a convenient, low-cost way for such users to define a hierarchy of tasks and to associate resources with those tasks. With this information, TaskTracer then supports the multi-tasking user by configuring the computer for the current task. To do this, it must detect when the user switches the task and identify the user's current task at all times. This problem of "task switch detection" is a special case of the general… 

Tables from this paper

Detecting and correcting user activity switches: algorithms and interfaces

TLDR
This paper presents TaskPredictor2, a complete redesign of the activity predictor in TaskTracer and its notification user interface that applies a novel online learning algorithm that is able to incorporate a richer set of features than the authors' previous predictors.

Detecting Developers’ Task Switches and Types

TLDR
This work examines the use of automatically collected computer interaction data for detecting developers’ task switches as well as task types in the field with high accuracy and semantic features and discusses their practical value for a wide range of applications in real work settings.

From documents to tasks: deriving user tasks from document usage patterns

TLDR
This paper proposes a new approach for automatically estimating a user's tasks from document interactions in an unsupervised manner and results indicate that the approach is better than previous approaches that use content.

EXPLOITING THE USER INTERACTION CONTEXT FOR AUTOMATIC TASK DETECTION

TLDR
An ontology-based user interaction context model that can be automatically populated by capturing simple user interaction events on the computer desktop and applying rule-based and information extraction mechanisms and isolate a combination of contextual features that present a significantly better discriminative power than classical ones is presented.

Improving Intelligent Assistants for Desktop Activities

TLDR
Reducing the number of interruptions to the user, a cost-sensitive approach to task prediction and notification in the user interface is needed, and improving the accuracy of predictions.

Automatic identification and description of software developers tasks

TLDR
Two novel approaches are introduced that capture the contents of a developer’s active window at regular intervals to create vector and visual representations of the work in a particular time interval and automatically detects the times at which developers switch tasks.

ABSTRACT OF DISSERTATION

TLDR
This thesis proposes a novel online learning algorithm that is able to incorporate a richer set of features than the authors' previous predictors and proves an error bound for the algorithm and presents experimental results that show improved accuracy and a 180-fold speedup on real user data.

Switch detector: an activity spotting system for desktop

TLDR
This paper presents a new technique that takes advantage of temporal aspects of user activity behavior to infer when it is most likely that an activity switch is occurring and describes an interactive switch notification system embodying these ideas.

UICO: an ontology-based user interaction context model for automatic task detection on the computer desktop

TLDR
An ontology-based user interaction context model (UICO) is proposed that enhances the performance of task detection on the user's computer desktop by utilizing rule-based, information extraction and machine learning approaches.

References

SHOWING 1-10 OF 29 REFERENCES

A hybrid learning system for recognizing user tasks from desktop activities and email messages

TLDR
This paper introduces TaskPredictor, a machine learning system that attempts to predict the user's current activity, and provides experimental results on data collected from TaskTracer users.

TaskTracer: a desktop environment to support multi-tasking knowledge workers

TLDR
A novel Publisher-Subscriber architecture for collecting and processing users' activity data is presented, several different user interfaces tried with TaskTracer are described, and the possibility of applying machine learning techniques to recognize/predict users' tasks is discussed.

"Constant, constant, multi-tasking craziness": managing multiple working spheres

TLDR
Empirical results that suggest that people organize their work in terms of much larger and thematically connected units of work are presented, and it is argued that design of information technology needs to support people's continual switching between working spheres.

A Comparison of Algorithms for Maximum Entropy Parameter Estimation

TLDR
A number of algorithms for estimating the parameters of ME models are considered, including iterative scaling, gradient ascent, conjugate gradient, and variable metric methods.

Maximum Entropy Markov Models for Information Extraction and Segmentation

TLDR
A new Markovian sequence model is presented that allows observations to be represented as arbitrary overlapping features (such as word, capitalization, formatting, part-of-speech), and defines the conditional probability of state sequences given observation sequences.

Statistical Models for Text Segmentation

TLDR
Assessment of the approach on quantitative and qualitative grounds demonstrates its effectiveness in two very different domains, Wall Street Journal news articles and television broadcast news story transcripts, using a new probabilistically motivated error metric.

A Simple-Transition Model for Relational Sequences

TLDR
A simple-transition cost model is introduced, which is parameterized by weighted constraints and a statetransition cost and gives a practical approach to SSM based on logical reasoning and bounded search for relational Horn constraints.

Event detection from time series data

TLDR
An iterative algorithm is proposed that fits a model to a time segment, and uses a likelihood criterion to determine if the segment should be partitioned further, i.e. if it contains a new changepoint.

Speaker, Environment and Channel Change Detection and Clustering via the Bayesian Information Criterion

TLDR
The segmentation algorithm can successfully detect acoustic changes; the clustering algorithm can produce clusters with high purity, leading to improvements in accuracy through unsupervised adaptation as much as the ideal clustering by the true speaker identities.