scholarly journals Lifelong Learning with a Changing Action Set

2020 ◽  
Vol 34 (04) ◽  
pp. 3373-3380
Author(s):  
Yash Chandak ◽  
Georgios Theocharous ◽  
Chris Nota ◽  
Philip Thomas
Keyword(s):  
Lifelong Learning ◽  
Real World ◽  
Large Scale ◽  
Learning Problems ◽  
Sequential Decision Making ◽  
Sequential Decision ◽  
Transition Dynamics ◽  
Unknown Structure ◽  
Changes Over Time ◽  
Over Time

In many real-world sequential decision making problems, the number of available actions (decisions) can vary over time. While problems like catastrophic forgetting, changing transition dynamics, changing rewards functions, etc. have been well-studied in the lifelong learning literature, the setting where the size of the action set changes remains unaddressed. In this paper, we present first steps towards developing an algorithm that autonomously adapts to an action set whose size changes over time. To tackle this open problem, we break it into two problems that can be solved iteratively: inferring the underlying, unknown, structure in the space of actions and optimizing a policy that leverages this structure. We demonstrate the efficiency of this approach on large-scale real-world lifelong learning problems.

scholarly journals Opportunity Costs in the Game of Best Choice

10.37236/8322 ◽  
2019 ◽  
Vol 26 (1) ◽  
Author(s):  
Madeline Crews ◽  
Brant Jones ◽  
Kaitlyn Myers ◽  
Laura Taalman ◽  
Michael Urbanski ◽  
...  
Keyword(s):  
Decision Making ◽  
Classical Case ◽  
Secretary Problem ◽  
Sequential Decision Making ◽  
Opportunity Costs ◽  
Sequential Decision ◽  
Probability Of Success ◽  
Uniform Model ◽  
Weighted Model ◽  
Over Time

The game of best choice, also known as the secretary problem, is a model for sequential decision making with many variations in the literature. Notably, the classical setup assumes that the sequence of candidate rankings is uniformly distributed over time and that there is no expense associated with the candidate interviews. Here, we weight each ranking permutation according to the position of the best candidate in order to model costs incurred from conducting interviews with candidates that are ultimately not hired. We compare our weighted model with the classical (uniform) model via a limiting process. It turns out that imposing even infinitesimal costs on the interviews results in a probability of success that is about 28%, as opposed to 1/e (about 37%) in the classical case.

scholarly journals Concept Drift Detection in Dynamic Probabilistic Relational Models

2021 ◽  
Vol 34 (1) ◽  
Author(s):  
Nils Finke ◽  
Tanya Braun ◽  
Marcel Gehrke ◽  
Ralf Möller
Keyword(s):  
Real World ◽  
Joint Distribution ◽  
Concept Drift ◽  
Probability Distributions ◽  
Real World Data ◽  
Relational Models ◽  
World Data ◽  
Probabilistic Relational Models ◽  
Changes Over Time ◽  
Over Time

Dynamic probabilistic relational models, which are factorized w.r.t. a full joint distribution, are used to cater for uncertainty and for relational and temporal aspects in real-world data. While these models assume the underlying temporal process to be stationary, real-world data often exhibits non-stationary behavior where the full joint distribution changes over time. We propose an approach to account for non-stationary processes w.r.t. to changing probability distributions over time, an effect known as concept drift. We use factorization and compact encoding of relations to efficiently detect drifts towards new probability distributions based on evidence.

A new analytic framework for prediction of migration patterns and locations of historical manuscripts based on their script types

2019 ◽  
Vol 35 (2) ◽  
pp. 441-458
Author(s):  
Gila Prebor ◽  
Maayan Zhitomirsky-Geffet ◽  
Yitzchak Miller
Keyword(s):  
Large Scale ◽  
Western Europe ◽  
Machine Learning Algorithms ◽  
Supervised Machine Learning ◽  
13Th Century ◽  
Migration Patterns ◽  
Historical Manuscripts ◽  
European Jewish ◽  
Changes Over Time ◽  
Over Time

Abstract In this article, we utilized large-scale statistical analysis and data visualization techniques of the greatest collection in the world of Hebrew manuscript metadata records to develop a new methodology for quantitative investigation of the palaeographic, geographic, and temporal characteristics of historical manuscripts. The study aims to explore whether and to what extent the script type of the manuscript and its changes over time can be used to automatically predict and complete missing geospatial data of the manuscripts. To this end, various ontological entities were used as features to train supervised machine-learning algorithms to predict the places of writing of manuscripts which were often absent in the catalogue records. The obtained results show that while the script type as an only feature might not be sufficient for prediction of the location of the manuscript’s writing, its combination with temporal data of the manuscript yielded about 80% accuracy. Eventually, our system was able to complete the missing places of writing for over 60% of the manuscripts in the corpus. Moreover, we found that through typical and marginal script types in different regions and their changes over time, it is possible to draw the migration map of the Jewish communities over the centuries. This reinforces the findings of historical research on Jewish migration patterns and communal formation. For example, the waves of immigration from Western Europe can be seen clearly from the second half of the 13th century, which continued until the 17th century and greatly increased the Eastern European Jewish community.

Consumer Knowledge and Acceptance of Driving Automation: Changes Over Time and Across Age Groups

2021 ◽  
Vol 65 (1) ◽  
pp. 1395-1399
Author(s):  
Chaiwoo Lee ◽  
Pnina Gershon ◽  
Bryan Reimer ◽  
Bruce Mehler ◽  
Joseph F. Coughlin
Keyword(s):  
Large Scale ◽  
Age Groups ◽  
The United States ◽  
Consumer Knowledge ◽  
Safety Assurance ◽  
Vehicle Automation ◽  
Willingness To Use ◽  
Changes Over Time ◽  
Over Time ◽  
Knowledge And Acceptance

Increasing availability of and extensive investments toward automation in consumer vehicles call for a better understanding of public perceptions and acceptance. This study presents a five-year series of large-scale surveys (N=17,548, average 3,510 participants/year) on consumer knowledge and acceptance of vehicle automation in the United States. Results suggest a continued hesitance toward use of self-driving vehicles, with willingness to use increasing sharply under hypothetical conditions around inability to drive and added safety assurance. While drivers of all ages were most comfortable with driver assist level automation, acceptance of automation overall decreased with age. Findings also indicate that the public may have incorrect beliefs regarding the availability of self-driving vehicles. In conclusion, drivers’ acceptance of vehicle automation changes over time, is tied to factors beyond the current state of development and deployment, and may depend on a relative assessment of benefits and reliability in comparison to their own driving capabilities.

scholarly journals Controlling transient gas flow in real-world pipeline intersection areas

2020 ◽  
Author(s):  
Felix Hennings ◽  
Lovis Anderson ◽  
Kai Hoppmann-Baum ◽  
Mark Turner ◽  
Thorsten Koch
Keyword(s):  
Real World ◽  
Gas Flow ◽  
Mixed Integer ◽  
Real World Data ◽  
Operation Modes ◽  
Intersection Area ◽  
Demand Scenario ◽  
Changes Over Time ◽  
Ongoing Project ◽  
Over Time

Abstract Compressor stations are the heart of every high-pressure gas transport network. Located at intersection areas of the network, they are contained in huge complex plants, where they are in combination with valves and regulators responsible for routing and pushing the gas through the network. Due to their complexity and lack of data compressor stations are usually dealt with in the scientific literature in a highly simplified and idealized manner. As part of an ongoing project with one of Germany’s largest transmission system operators to develop a decision support system for their dispatching center, we investigated how to automatize the control of compressor stations. Each station has to be in a particular configuration, leading in combination with the other nearby elements to a discrete set of up to 2000 possible feasible operation modes in the intersection area. Since the desired performance of the station changes over time, the configuration of the station has to adapt. Our goal is to minimize the necessary changes in the overall operation modes and related elements over time while fulfilling a preset performance envelope or demand scenario. This article describes the chosen model and the implemented mixed-integer programming based algorithms to tackle this challenge. By presenting extensive computational results on real-world data, we demonstrate the performance of our approach.

Assessing sustainability knowledge of a student population

2014 ◽  
Vol 15 (4) ◽  
pp. 375-389 ◽  
Author(s):  
Adam Zwickle ◽  
Tomas M. Koontz ◽  
Kristina M. Slagle ◽  
Jeremy T. Bruskotter
Keyword(s):  
Large Scale ◽  
Online Survey ◽  
Assessment Tool ◽  
Multiple Choice Questions ◽  
Content Type ◽  
College And University ◽  
Expert Input ◽  
Changes Over Time ◽  
Practical Implications ◽  
Over Time

Purpose – The purpose of this article is to present a tool for assessing the sustainability knowledge of an undergraduate population. Design/methodology/approach – Multiple-choice questions were developed through soliciting expert input, focus groups, pilot testing, distribution via a large-scale online survey and analysis using item response theory. Findings – The final assessment consists of 16 questions from the environmental, economic and social domains, covering foundational concepts within the topic of sustainability. Research limitations/implications – This assessment represents an initial effort to quantify knowledge of the broad and abstract concept of sustainability. The authors plan to continue refining these questions to better differentiate between students with higher levels of knowledge and to replace those with answers that may change over time. Practical implications – With knowledge of sustainability concepts becoming increasingly included in institution-wide learning objectives, there is a growing demand for a way to measure progress in this area. Our assessment tool can easily be used (via a campus-wide survey or distributed at the classroom level) by institutions to gauge current levels of knowledge and track changes over time, as well as assess the effectiveness of courses and curricula at meeting sustainability knowledge goals. Originality/value – This assessment of sustainability knowledge is the first of its kind to include all three separate domains of sustainability, and we expect it to be useful across a variety of college and university contexts.

scholarly journals Fast Pareto Optimization for Subset Selection with Dynamic Cost Constraints

2021 ◽  
Author(s):  
Chao Bian ◽  
Chao Qian ◽  
Frank Neumann ◽  
Yang Yu
Keyword(s):  
Objective Function ◽  
Greedy Algorithm ◽  
Real World ◽  
Fundamental Problem ◽  
Subset Selection ◽  
Selection Strategy ◽  
Cost Constraints ◽  
Approximation Guarantee ◽  
Changes Over Time ◽  
Over Time

Subset selection with cost constraints is a fundamental problem with various applications such as influence maximization and sensor placement. The goal is to select a subset from a ground set to maximize a monotone objective function such that a monotone cost function is upper bounded by a budget. Previous algorithms with bounded approximation guarantees include the generalized greedy algorithm, POMC and EAMC, all of which can achieve the best known approximation guarantee. In real-world scenarios, the resources often vary, i.e., the budget often changes over time, requiring the algorithms to adapt the solutions quickly. However, when the budget changes dynamically, all these three algorithms either achieve arbitrarily bad approximation guarantees, or require a long running time. In this paper, we propose a new algorithm FPOMC by combining the merits of the generalized greedy algorithm and POMC. That is, FPOMC introduces a greedy selection strategy into POMC. We prove that FPOMC can maintain the best known approximation guarantee efficiently.

scholarly journals Pig-Posture Recognition Based on Computer Vision: Dataset and Exploration

Animals ◽  
2021 ◽  
Vol 11 (5) ◽  
pp. 1295
Author(s):  
Hongmin Shao ◽  
Jingyu Pu ◽  
Jiong Mu
Keyword(s):  
Large Scale ◽  
Disease Monitoring ◽  
Pathological Changes ◽  
Convolutional Network ◽  
Postural Changes ◽  
The World ◽  
Posture Recognition ◽  
Behavioral Activities ◽  
Changes Over Time ◽  
Over Time

Posture changes in pigs during growth are often precursors of disease. Monitoring pigs’ behavioral activities can allow us to detect pathological changes in pigs earlier and identify the factors threatening the health of pigs in advance. Pigs tend to be farmed on a large scale, and manual observation by keepers is time consuming and laborious. Therefore, the use of computers to monitor the growth processes of pigs in real time, and to recognize the duration and frequency of pigs’ postural changes over time, can prevent outbreaks of porcine diseases. The contributions of this article are as follows: (1) The first human-annotated pig-posture-identification dataset in the world was established, including 800 pictures of each of the four pig postures: standing, lying on the stomach, lying on the side, and exploring. (2) When using a deep separable convolutional network to classify pig postures, the accuracy was 92.45%. The results show that the method proposed in this paper achieves adequate pig-posture recognition in a piggery environment and may be suitable for livestock farm applications.

scholarly journals Diversity of Online Behaviours Associated with Physical Attendance in Lectures

2019 ◽  
Vol 6 (1) ◽  
Author(s):  
Amelia Brennan ◽  
Akshay Sharma ◽  
Pablo Munguia
Keyword(s):  
Large Scale ◽  
Online Course ◽  
Course Content ◽  
Face To Face ◽  
Course Materials ◽  
Use Of Technology ◽  
Online Behaviour ◽  
Online Access ◽  
Changes Over Time ◽  
Over Time

A common use of technology in higher education is the provision of online course materials, invoking an investigation of the ways in which students engage with online course content, and how their participation changes over time. This is particularly necessary in the context of high absenteeism from lectures, where online access may be the only way in which particular students are engaging with the course. In this study, we examine large-scale patterns of attendance in class, as well as four types of access to online materials. We define two online behavioural metrics — richness and evenness — to capture the distribution of online behaviours within 255 courses, and examine how these change over time. We find that both physical and online attendance decrease throughout the semester, but the fraction of students present online is considerably higher than the fraction present in lectures. Students adapt their online behaviour, and rare behaviours disappear over time. It is important to consider how we provide content, both face-to-face and online, in order to ensure that as many students as possible are accessing this content in ways that we intend.

scholarly journals On Merging and Dividing Social Graphs

2015 ◽  
Vol 5 (1) ◽  
pp. 23-49 ◽  
Author(s):  
Pascal Held ◽  
Alexander Dockhorn ◽  
Rudolf Kruse
Keyword(s):  
Parameter Estimation ◽  
Social Interaction ◽  
Real World ◽  
Extended Version ◽  
Graph Models ◽  
Advantages And Disadvantages ◽  
Graph Properties ◽  
Social Graphs ◽  
Changes Over Time ◽  
Over Time

Abstract Modeling social interaction can be based on graphs. However most models lack the flexibility of including larger changes over time. The Barabási-Albert-model is a generative model which already offers mechanisms for adding nodes. We will extent this by presenting four methods for merging and five for dividing graphs based on the Barabási- Albert-model. Our algorithms were motivated by different real world scenarios and focus on preserving graph properties derived from these scenarios. With little alterations in the parameter estimation those algorithms can be used for other graph models as well. All algorithms were tested in multiple experiments using graphs based on the Barabási- Albert-model, an extended version of the Barabási-Albert-model by Holme and Kim, the Watts-Strogatz-model and the Erdős-Rényi-model. Furthermore we concluded that our algorithms are able to preserve different properties of graphs independently from the used model. To support the choice of algorithm, we created a guideline which highlights advantages and disadvantages of discussed methods and their possible use-cases.

Sign in / Sign up

Export Citation Format

Share Document