15,180 research outputs found
LiBRe: Label-Wise Selection of Base Learners in Binary Relevance for Multi-label Classification
No full textMaking Learners (More) Monotone
No full textLearning performance can show non-monotonic behavior. That is, more data does not necessarily lead to better models, even on average. We propose three algorithms that take a supervised learning model and make it perform more monotone. We prove consistency and monotonicity with high probability, and evaluate the algorithms on scenarios where non-monotone behaviour occurs. Our proposed algorithm MTHT makes less than 1% non-monotone decisions on MNIST while staying competitive in terms of error rate compared to several baselines. Our code is available at https://github.com/tomviering/monotone.Virtual/online event due to COVID-19Pattern Recognition and BioinformaticsInteractive Intelligenc
A Distribution Dependent and Independent Complexity Analysis of Manifold Regularization
No full textManifold regularization is a commonly used technique in semi-supervised learning. It enforces the classification rule to be smooth with respect to the data-manifold. Here, we derive sample complexity bounds based on pseudo-dimension for models that add a convex data dependent regularization term to a supervised learning process, as is in particular done in Manifold regularization. We then compare the bound for those semi-supervised methods to purely supervised methods, and discuss a setting in which the semi-supervised method can only have a constant improvement, ignoring logarithmic terms. By viewing Manifold regularization as a kernel method we then derive Rademacher bounds which allow for a distribution dependent analysis. Finally we illustrate that these bounds may be useful for choosing an appropriate manifold regularization parameter in situations with very sparsely labeled data.Virtual/online event due to COVID-19Interactive IntelligencePattern Recognition and Bioinformatic
Link stability estimation based on link connectivity changes in mobile ad-hoc networks
No full textDear Wang,
Re: Link Stability Estimation Based on Link Connectivity Changes in Mobile Ad-hoc Networks
I have not been able to assess if this is an author version peer-reviewed or is it an author version non peer reviewed. Could you please clarify this so I can proceed to add your paper to Spiral. Spiral digital repository only accept peer-reviewed papers.
30/11/12 author has confirmed peer reviewe
Orometric methods in bounded metric data
No full textA large amount of data accommodated in knowledge graphs (KG) is metric. For example, the Wikidata KG contains a plenitude of metric facts about geographic entities like cities or celestial objects. In this paper, we propose a novel approach that transfers orometric (topographic) measures to bounded metric spaces. While these methods were originally designed to identify relevant mountain peaks on the surface of the earth, we demonstrate a notion to use them for metric data sets in general. Notably, metric sets of items enclosed in knowledge graphs. Based on this we present a method for identifying outstanding items using the transferred valuations functions isolation and prominence. Building up on this we imagine an item recommendation process. To demonstrate the relevance of the valuations for such processes, we evaluate the usefulness of isolation and prominence empirically in a machine learning setting. In particular, we find structurally relevant items in the geographic population distributions of Germany and France. © 2020, The Author(s)
Angle-Based Crowding Degree Estimation for Many-Objective Optimization
No full text© 2020, The Author(s). Many-objective optimization, which deals with an optimization problem with more than three objectives, poses a big challenge to various search techniques, including evolutionary algorithms. Recently, a meta-objective optimization approach (called bi-goal evolution, BiGE) which maps solutions from the original high-dimensional objective space into a bi-goal space of proximity and crowding degree has received increasing attention in the area. However, it has been found that BiGE tends to struggle on a class of many-objective problems where the search process involves dominance resistant solutions, namely, those solutions with an extremely poor value in at least one of the objectives but with (near) optimal values in some of the others. It is difficult for BiGE to get rid of dominance resistant solutions as they are Pareto nondominated and far away from the main population, thus always having a good crowding degree. In this paper, we propose an angle-based crowding degree estimation method for BiGE (denoted as aBiGE) to replace distance-based crowding degree estimation in BiGE. Experimental studies show the effectiveness of this replacement
Comparing bagging and boosting for natural language processing tasks, a typicality approach
No full textMulti-relational data mining, using UML for ILP
No full textsponsorship: Department of Information Management, Tilburg Universitystatus: Publishe
- …
