scholarly journals Cost-Effective Active Learning for Hierarchical Multi-Label Classification

2018 ◽  
Author(s):  
Yi-Fan Yan ◽  
Sheng-Jun Huang
Keyword(s):  
Active Learning ◽  
Subset Selection ◽  
Cost Effective ◽  
Selection Method ◽  
Learning Approach ◽  
Potential Contribution ◽  
Annotation Costs ◽  
Model Training ◽  
Coarse To Fine ◽  
Different Levels

Active learning reduces the labeling cost by actively querying labels for the most valuable data. It is particularly important for multi-label learning, where the annotation cost is rather high because each instance may have multiple labels simultaneously. In many multi-label tasks, the labels are organized into hierarchies from coarse to fine. The labels at different levels of the hierarchy contribute differently to the model training, and also have diverse annotation costs. In this paper, we propose a multi-label active learning approach to exploit the label hierarchies for cost-effective queries. By incorporating the potential contribution of ancestor and descendant labels, a novel criterion is proposed to estimate the informativeness of each candidate query. Further, a subset selection method is introduced to perform active batch selection by balancing the informativeness and cost of each instance-label pair. Experimental results validate the effectiveness of both the proposed criterion and the selection method.

scholarly journals Actively Searching: Inverse Design of Novel Molecules with Simultaneously Optimized Properties

2021 ◽  
Author(s):  
Nicolae C. Iovanac ◽  
Robert MacKnight ◽  
Brett Savoie
Keyword(s):  
Active Learning ◽  
Quantum Chemistry ◽  
Chemical Space ◽  
Cost Effective ◽  
Generative Models ◽  
Learning Approach ◽  
The Novel ◽  
Practical Applications ◽  
Chemical Models ◽  
And Shifts

<p>Combining quantum chemistry characterizations with generative machine learning models has the potential to accelerate molecular searches in chemical space. In this paradigm, quantum chemistry acts as a relatively cost-effective oracle for evaluating the properties of particular molecules while generative models provide a means of sampling chemical space based on learned structure-function relationships. For practical applications, multiple potentially orthogonal properties must be optimized in tandem during a discovery workflow. This carries additional difficulties associated with specificity of the targets and the ability for the model to reconcile all properties simultaneously. Here we demonstrate an active learning approach to improve the performance of multi-target generative chemical models. We first demonstrate the effectiveness of a set of baseline models trained on single property prediction tasks in generating novel compounds with various property targets, including both interpolative and extrapolative generation scenarios. For property ranges where accurate targeting proves difficult, the novel compounds suggested by the model are characterized using quantum chemistry to obtain the true values, and these new molecules closest to expressing the desired properties are fed back into the generative model for additional training. This gradually improves the generative models’ understanding of unknown areas of chemical space and shifts the distribution of generated compounds towards the targeted values. We then demonstrate the effectiveness of this active learning approach in generating compounds with multiple chemical constraints, including vertical ionization potential, electron affinity, and dipole moment targets, and validate the results at the B97X-D3/def2-TZVP level. This method requires no modifications to extant generative approaches, but rather utilizes their inherent generative and predictive aspects for self-refinement, and can be applied to situations where any number of properties with varying degrees of correlation must be optimized simultaneously.</p>

scholarly journals Actively Searching: Inverse Design of Novel Molecules with Simultaneously Optimized Properties

2021 ◽  
Author(s):  
Nicolae C. Iovanac ◽  
Robert MacKnight ◽  
Brett Savoie
Keyword(s):  
Active Learning ◽  
Quantum Chemistry ◽  
Chemical Space ◽  
Cost Effective ◽  
Generative Models ◽  
Learning Approach ◽  
The Novel ◽  
Practical Applications ◽  
Chemical Models ◽  
And Shifts

<p>Combining quantum chemistry characterizations with generative machine learning models has the potential to accelerate molecular searches in chemical space. In this paradigm, quantum chemistry acts as a relatively cost-effective oracle for evaluating the properties of particular molecules while generative models provide a means of sampling chemical space based on learned structure-function relationships. For practical applications, multiple potentially orthogonal properties must be optimized in tandem during a discovery workflow. This carries additional difficulties associated with specificity of the targets and the ability for the model to reconcile all properties simultaneously. Here we demonstrate an active learning approach to improve the performance of multi-target generative chemical models. We first demonstrate the effectiveness of a set of baseline models trained on single property prediction tasks in generating novel compounds with various property targets, including both interpolative and extrapolative generation scenarios. For property ranges where accurate targeting proves difficult, the novel compounds suggested by the model are characterized using quantum chemistry to obtain the true values, and these new molecules closest to expressing the desired properties are fed back into the generative model for additional training. This gradually improves the generative models’ understanding of unknown areas of chemical space and shifts the distribution of generated compounds towards the targeted values. We then demonstrate the effectiveness of this active learning approach in generating compounds with multiple chemical constraints, including vertical ionization potential, electron affinity, and dipole moment targets, and validate the results at the B97X-D3/def2-TZVP level. This method requires no modifications to extant generative approaches, but rather utilizes their inherent generative and predictive aspects for self-refinement, and can be applied to situations where any number of properties with varying degrees of correlation must be optimized simultaneously.</p>

scholarly journals Searching for NEOs using Lowell observatorys Discovery Channel Telescope (DCT)

2006 ◽  
Vol 2 (S236) ◽  
pp. 363-370
Author(s):  
Edward Bowell ◽  
Robert L. Millis ◽  
Edward W. Dunham ◽  
Bruce W. Koehn ◽  
Byron W. Smith
Keyword(s):  
Cost Effective ◽  
Low Risk ◽  
Potential Contribution ◽  
Detection Program ◽  
Search Program ◽  
Different Levels

AbstractWe discuss the potential contribution of the Discovery Channel Telescope (or a clone) to a detection program aimed at discovering 90% of potentially hazardous objects (PHOs) larger than 140 m in diameter. Three options are described, each involving different levels of investment. We believe that LSST, Pan-STARRS, and DCT, working in a coordinated fashion, offer a cost-effective, low-risk way to accomplish the objectives of the extended NEO search program.

Peningkatan Hasil Belajar Fisika Melalui Pendekatan Numbered Head Together Dengan Pemberian Reward di SMP Negeri 5 Kota Jambi

2012 ◽  
Vol 1 (1) ◽  
Author(s):  
Delismar Delismar
Keyword(s):  
Action Research ◽  
Active Learning ◽  
Learning Outcomes ◽  
Learning Process ◽  
Learning Outcome ◽  
Class Action ◽  
Learning Approach ◽  
Physics Learning ◽  
Lecture Method ◽  
Student Observation

In classical learning approach, conventional lecture method is commonly used by teachers in implementing learning process in classes.  The teacher becomes the main source of learning.  The current student’s habit that tends to be passive and individualistic resulted in a passive and monotone learning.      To overcome these problems, I was interested to implement the model of numbered heads together in learning Physics in the Class VII B of SMP Negeri 5 Kota Jambi. The purpose of this learning approach is to enable students to develop cooperative skill and more active learning of physics and to improve learning results. This research is a class action research, which were performed in two cycles.  All students’ activities in the class were observed and recorded in observation sheet, consisting of teacher observation sheet and student observation sheet. To find out the learning outcomes, formative test was performed using a written instrument form.  The results show the increase of students’ discipline, cooperation, liveliness, timeliness in learning Physics.  In addition, the learning model also increases the students’ learning outcomes. The average learning results increased to 75.38 (increase 3.25 points).  To conclude, the implementation of Number Head Together increase students’ discipline, cooperation, activities, and timeliness.  The model also increase the Physics learning outcome of student in SMP Negeri 5 Kota  Jambi.

An Active Learning Approach to Teaching Social Determinants of Health

2020 ◽  
Vol 6 (4) ◽  
pp. 266-273
Author(s):  
Jeanita W. Richardson
Keyword(s):  
Active Learning ◽  
Learning Style ◽  
Social Determinants ◽  
Learning Approach ◽  
Learning Style Preferences ◽  
Open Discussion ◽  
Approach To Teaching ◽  
The Individual ◽  
Style Preferences ◽  
The Impact

This active learning exercise is designed to deconstruct the impact of social determinants through the assumption of randomly selected personas. As an active learning exercise, it provides opportunities for discussion, problem solving, writing, and synthesis, while incorporating multiple learning style preferences. Part 1 involves assessing the individual social determinants at work. Part 2 involves exploring ways said determinants can enhance community health through collaboration. Assumption of personas unlike one’s own facilitates an open discussion of social position and ranges of factors influential to health without potentially evoking a sense of defensiveness associated with personal privilege (or the lack thereof).

A Temporal Dependency Based Multi-modal Active Learning Approach for Audiovisual Event Detection

2017 ◽  
Vol 48 (2) ◽  
pp. 709-732 ◽  
Author(s):  
Patrick Thiam ◽  
Sascha Meudt ◽  
Günther Palm ◽  
Friedhelm Schwenker
Keyword(s):  
Active Learning ◽  
Event Detection ◽  
Learning Approach
Sign in / Sign up

Export Citation Format

Share Document