scispace - formally typeset
Search or ask a question
Topic

Task analysis

About: Task analysis is a research topic. Over the lifetime, 10432 publications have been published within this topic receiving 283481 citations.


Papers
More filters
Journal ArticleDOI
TL;DR: This survey provides a well-rounded view on state-of-the-art deep learning approaches for MTL in computer vision, explicitly emphasizing on dense prediction tasks.
Abstract: With the advent of deep learning, many dense prediction tasks, i.e. tasks that produce pixel-level predictions, have seen significant performance improvements. The typical approach is to learn these tasks in isolation, that is, a separate neural network is trained for each individual task. Yet, recent multi-task learning (MTL) techniques have shown promising results w.r.t. performance, computations and/or memory footprint, by jointly tackling multiple tasks through a learned shared representation. In this survey, we provide a well-rounded view on state-of-the-art deep learning approaches for MTL in computer vision, explicitly emphasizing on dense prediction tasks. Our contributions concern the following. First, we consider MTL from a network architecture point-of-view. We include an extensive overview and discuss the advantages/disadvantages of recent popular MTL models. Second, we examine various optimization methods to tackle the joint learning of multiple tasks. We summarize the qualitative elements of these works and explore their commonalities and differences. Finally, we provide an extensive experimental evaluation across a variety of dense prediction benchmarks to examine the pros and cons of the different methods, including both architectural and optimization based strategies.

320 citations

Book
01 Oct 2006
TL;DR: This book discusses task-based language teaching in science education and vocational training, TBLT and ICT, and a box full of feelings: promoting infants' second language acquisition all day long.
Abstract: Chapter 1 Introduction: Task-based language teaching in a nutshell, Chapter 2 From needs to tasks: Language learning needs in a task-based approach, Chapter 3 Tasks for absolute beginners ... and beyond: Developing and sequencing tasks at basic proficiency levels, Chapter 4 Developing language tasks for primary and secondary education, Chapter 5 Task-based language teaching in science education and vocational training, 6 TBLT and ICT: Developing and assessing interactive multimedia for task-based language teaching, Chapter 7 Developing and introducing task-based language tests, Chapter 8 The role of the teacher in TBLT, Chapter 9 A box full of feelings: Promoting infants' second language acquisition all day long, Chapter 10 Teacher training: Task-based as well?

317 citations

Journal ArticleDOI
TL;DR: It is argued that while promoting noticing is important, promoting the quality of that noticing is a more important issue to be addressed in L2 writing pedagogy.

315 citations

Journal Article
TL;DR: This article focuses on the selection and creation of mathematical tasks, drawing on QUASAR's research on mathematical tasks and on the own experiences with teachers and teacher educators.

314 citations

Journal ArticleDOI
TL;DR: The author describes and evaluates four types of CPR, including basic designs that include quantitative process–outcome, qualitative helpful factors, and microanalytic sequential process; the fourth refers to methods such as task analysis and comprehensive process analysis that integrate the first three.
Abstract: Change process research (CPR) is the study of the processes by which change occurs in psychotherapy and is a necessary complement to randomized clinical trials and other forms of efficacy research. In this article the author describes and evaluates four types of CPR. The first three are basic designs and include quantitative process–outcome, qualitative helpful factors, and microanalytic sequential process; the fourth, the significant events approach, refers to methods such as task analysis and comprehensive process analysis that integrate the first three. The strengths and weaknesses of each design are described and summarized using both causal and practical criteria as part of an overall argument for systematic methodological pluralism.

314 citations


Network Information
Related Topics (5)
Feature extraction
111.8K papers, 2.1M citations
78% related
Robustness (computer science)
94.7K papers, 1.6M citations
78% related
User interface
85.4K papers, 1.7M citations
78% related
The Internet
213.2K papers, 3.8M citations
77% related
Deep learning
79.8K papers, 2.1M citations
77% related
Performance
Metrics
No. of papers in the topic in previous years
YearPapers
202328
202264
2021665
2020819
2019737
2018834