-
On Image Segmentation With Noisy Labels: Characterization and Volume Properties of the Optimal Solutions to Accuracy and Dice
Authors:
Marcus Nordström,
Henrik Hult,
Jonas Söderberg,
Fredrik Löfman
Abstract:
We study two of the most popular performance metrics in medical image segmentation, Accuracy and Dice, when the target labels are noisy. For both metrics, several statements related to characterization and volume properties of the set of optimal segmentations are proved, and associated experiments are provided. Our main insights are: (i) the volume of the solutions to both metrics may deviate sign…
▽ More
We study two of the most popular performance metrics in medical image segmentation, Accuracy and Dice, when the target labels are noisy. For both metrics, several statements related to characterization and volume properties of the set of optimal segmentations are proved, and associated experiments are provided. Our main insights are: (i) the volume of the solutions to both metrics may deviate significantly from the expected volume of the target, (ii) the volume of a solution to Accuracy is always less than or equal to the volume of a solution to Dice and (iii) the optimal solutions to both of these metrics coincide when the set of feasible segmentations is constrained to the set of segmentations with the volume equal to the expected volume of the target.
△ Less
Submitted 31 March, 2023; v1 submitted 13 June, 2022;
originally announced June 2022.
-
Socially Intelligent Interfaces for Increased Energy Awareness in the Home
Authors:
Jussi Karlgren,
Lennart E. Fahlén,
Anders Wallberg,
Pär Hansson,
Olov Ståhl,
Jonas Söderberg,
Karl-Petter Åkesson
Abstract:
This paper describes how home appliances might be enhanced to improve user awareness of energy usage. Households wish to lead comfortable and manageable lives. Balancing this reasonable desire with the environmental and political goal of reducing electricity usage is a challenge that we claim is best met through the design of interfaces that allows users better control of their usage and unobtrusi…
▽ More
This paper describes how home appliances might be enhanced to improve user awareness of energy usage. Households wish to lead comfortable and manageable lives. Balancing this reasonable desire with the environmental and political goal of reducing electricity usage is a challenge that we claim is best met through the design of interfaces that allows users better control of their usage and unobtrusively informs them of the actions of their peers. A set of design principles along these lines is formulated in this paper. We have built a fully functional prototype home appliance with a socially aware interface to signal the aggregate usage of the users peer group according to these principles, and present the prototype in the paper.
△ Less
Submitted 29 June, 2021;
originally announced June 2021.
-
Categorizing Online Shop** Behavior from Cosmetics to Electronics: An Analytical Framework
Authors:
Sohini Roychowdhury,
Wenxi Li,
Ebrahim Alareqi,
Akhilesh Pandita,
Ao Liu,
Joakim Soderberg
Abstract:
A success factor for modern companies in the age of Digital Marketing is to understand how customers think and behave based on their online shop** patterns. While the conventional method of gathering consumer insights through questionnaires and surveys still form the bases of descriptive analytics for market intelligence units, we propose a machine learning framework to automate this process. In…
▽ More
A success factor for modern companies in the age of Digital Marketing is to understand how customers think and behave based on their online shop** patterns. While the conventional method of gathering consumer insights through questionnaires and surveys still form the bases of descriptive analytics for market intelligence units, we propose a machine learning framework to automate this process. In this paper we present a modular consumer data analysis platform that processes session level interaction records between users and products to predict session level, user journey level and customer behavior specific patterns leading towards purchase events. We explore the computational framework and provide test results on two Big data sets-cosmetics and consumer electronics of size 2GB and 15GB, respectively. The proposed system achieves 97-99% classification accuracy and recall for user-journey level purchase predictions and categorizes buying behavior into 5 clusters with increasing purchase ratios for both data sets. Thus, the proposed framework is extendable to other large e-commerce data sets to obtain automated purchase predictions and descriptive consumer insights.
△ Less
Submitted 6 October, 2020;
originally announced October 2020.
-
Betti numbers of graded modules and the Multiplicity Conjecture in the non-Cohen-Macaulay case
Authors:
Mats Boij,
Jonas Soderberg
Abstract:
We use the results by Eisenbud and Schreyer to prove that any Betti diagram of a graded module over a standard graded polynomial ring is a positive linear combination Betti diagrams of modules with a pure resolution. This implies the Multiplicity Conjecture of Herzog, Huneke and Srinivasan for modules that are not necessarily Cohen-Macaulay. We give a combinatorial proof of the convexity of the…
▽ More
We use the results by Eisenbud and Schreyer to prove that any Betti diagram of a graded module over a standard graded polynomial ring is a positive linear combination Betti diagrams of modules with a pure resolution. This implies the Multiplicity Conjecture of Herzog, Huneke and Srinivasan for modules that are not necessarily Cohen-Macaulay. We give a combinatorial proof of the convexity of the simplicial fan spanned by the pure diagrams.
△ Less
Submitted 11 March, 2008;
originally announced March 2008.
-
Graded Betti numbers and $h$-vectors of level modules
Authors:
Jonas Söderberg
Abstract:
We study $h$-vectors and graded Betti numbers of level modules up to multiplication by a rational number. Assuming a conjecture on the possible graded Betti numbers of Cohen-Macaulay modules we get a description of the possible $h$-vectors of level modules up to multiplication by a rational number. We also determine, again up to multiplication by a rational number, the cancellable $h$-vectors an…
▽ More
We study $h$-vectors and graded Betti numbers of level modules up to multiplication by a rational number. Assuming a conjecture on the possible graded Betti numbers of Cohen-Macaulay modules we get a description of the possible $h$-vectors of level modules up to multiplication by a rational number. We also determine, again up to multiplication by a rational number, the cancellable $h$-vectors and the $h$-vectors of level modules with the weak Lefschetz property. Furthermore, we prove that level modules of codimension three satisfy the upper bound of the Multiplicity conjecture of Herzog, Huneke and Srinivasan, and that the lower bound holds if the module, in addition, has the weak Lefschetz property.
△ Less
Submitted 4 December, 2006; v1 submitted 4 December, 2006;
originally announced December 2006.
-
Graded Betti numbers of Cohen-Macaulay modules and the Multiplicity conjecture
Authors:
Mats Boij,
Jonas Söderberg
Abstract:
We give conjectures on the possible graded Betti numbers of Cohen-Macaulay modules up to multiplication by positive rational numbers. The idea is that the Betti diagrams should be non-negative linear combinations of pure diagrams. The conjectures are verified in the cases where the structure of resolutions are known, i.e., for modules of codimension two, for Gorenstein algebras of codimension th…
▽ More
We give conjectures on the possible graded Betti numbers of Cohen-Macaulay modules up to multiplication by positive rational numbers. The idea is that the Betti diagrams should be non-negative linear combinations of pure diagrams. The conjectures are verified in the cases where the structure of resolutions are known, i.e., for modules of codimension two, for Gorenstein algebras of codimension three and for complete intersections. The motivation for the conjectures comes from the Multiplicity conjecture of Herzog, Huneke and Srinivasan.
△ Less
Submitted 21 March, 2007; v1 submitted 3 November, 2006;
originally announced November 2006.
