-
Brain MRI Screening Tool with Federated Learning
Authors:
Roman Stoklasa,
Ioannis Stathopoulos,
Efstratios Karavasilis,
Efstathios Efstathopoulos,
Marek Dostál,
Miloš Keřkovský,
Michal Kozubek,
Luigi Serio
Abstract:
In clinical practice, we often see significant delays between MRI scans and the diagnosis made by radiologists, even for severe cases. In some cases, this may be caused by the lack of additional information and clues, so even the severe cases need to wait in the queue for diagnosis. This can be avoided if there is an automatic software tool, which would supplement additional information, alerting…
▽ More
In clinical practice, we often see significant delays between MRI scans and the diagnosis made by radiologists, even for severe cases. In some cases, this may be caused by the lack of additional information and clues, so even the severe cases need to wait in the queue for diagnosis. This can be avoided if there is an automatic software tool, which would supplement additional information, alerting radiologists that the particular patient may be a severe case.
We are presenting an automatic brain MRI Screening Tool and we are demonstrating its capabilities for detecting tumor-like pathologies. It is the first version on the path toward a robust multi-pathology screening solution. The tool supports Federated Learning, so multiple institutions may contribute to the model without disclosing their private data.
△ Less
Submitted 23 November, 2023;
originally announced November 2023.
-
Federated Learning Enables Big Data for Rare Cancer Boundary Detection
Authors:
Sarthak Pati,
Ujjwal Baid,
Brandon Edwards,
Micah Sheller,
Shih-Han Wang,
G Anthony Reina,
Patrick Foley,
Alexey Gruzdev,
Deepthi Karkada,
Christos Davatzikos,
Chiharu Sako,
Satyam Ghodasara,
Michel Bilello,
Suyash Mohan,
Philipp Vollmuth,
Gianluca Brugnara,
Chandrakanth J Preetha,
Felix Sahm,
Klaus Maier-Hein,
Maximilian Zenk,
Martin Bendszus,
Wolfgang Wick,
Evan Calabrese,
Jeffrey Rudie,
Javier Villanueva-Meyer
, et al. (254 additional authors not shown)
Abstract:
Although machine learning (ML) has shown promise in numerous domains, there are concerns about generalizability to out-of-sample data. This is currently addressed by centrally sharing ample, and importantly diverse, data from multiple sites. However, such centralization is challenging to scale (or even not feasible) due to various limitations. Federated ML (FL) provides an alternative to train acc…
▽ More
Although machine learning (ML) has shown promise in numerous domains, there are concerns about generalizability to out-of-sample data. This is currently addressed by centrally sharing ample, and importantly diverse, data from multiple sites. However, such centralization is challenging to scale (or even not feasible) due to various limitations. Federated ML (FL) provides an alternative to train accurate and generalizable ML models, by only sharing numerical model updates. Here we present findings from the largest FL study to-date, involving data from 71 healthcare institutions across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, utilizing the largest dataset of such patients ever used in the literature (25,256 MRI scans from 6,314 patients). We demonstrate a 33% improvement over a publicly trained model to delineate the surgically targetable tumor, and 23% improvement over the tumor's entire extent. We anticipate our study to: 1) enable more studies in healthcare informed by large and diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further quantitative analyses for glioblastoma via performance optimization of our consensus model for eventual public release, and 3) demonstrate the effectiveness of FL at such scale and task complexity as a paradigm shift for multi-site collaborations, alleviating the need for data sharing.
△ Less
Submitted 25 April, 2022; v1 submitted 22 April, 2022;
originally announced April 2022.
-
Fuzzy Generalised Quantifiers for Natural Language in Categorical Compositional Distributional Semantics
Authors:
Matej Dostal,
Mehrnoosh Sadrzadeh,
Gijs Wijnholds
Abstract:
Recent work on compositional distributional models shows that bialgebras over finite dimensional vector spaces can be applied to treat generalised quantifiers for natural language. That technique requires one to construct the vector space over powersets, and therefore is computationally costly. In this paper, we overcome this problem by considering fuzzy versions of quantifiers along the lines of…
▽ More
Recent work on compositional distributional models shows that bialgebras over finite dimensional vector spaces can be applied to treat generalised quantifiers for natural language. That technique requires one to construct the vector space over powersets, and therefore is computationally costly. In this paper, we overcome this problem by considering fuzzy versions of quantifiers along the lines of Zadeh, within the category of many valued relations. We show that this category is a concrete instantiation of the compositional distributional model. We show that the semantics obtained in this model is equivalent to the semantics of the fuzzy quantifiers of Zadeh. As a result, we are now able to treat fuzzy quantification without requiring a powerset construction.
△ Less
Submitted 23 September, 2021;
originally announced September 2021.
-
Moss' logic for ordered coalgebras
Authors:
Marta Bílková,
Matěj Dostál
Abstract:
We present a finitary version of Moss' coalgebraic logic for $T$-coalgebras, where $T$ is a locally monotone endofunctor of the category of posets and monotone maps. The logic uses a single cover modality whose arity is given by the least finitary subfunctor of the dual of the coalgebra functor $T_ω^\partial$, and the semantics of the modality is given by relation lifting. For the semantics to wor…
▽ More
We present a finitary version of Moss' coalgebraic logic for $T$-coalgebras, where $T$ is a locally monotone endofunctor of the category of posets and monotone maps. The logic uses a single cover modality whose arity is given by the least finitary subfunctor of the dual of the coalgebra functor $T_ω^\partial$, and the semantics of the modality is given by relation lifting. For the semantics to work, $T$ is required to preserve exact squares. For the finitary setting to work, $T_ω^\partial$ is required to preserve finite intersections. We develop a notion of a base for subobjects of $T_ωX$. This in particular allows us to talk about the finite poset of subformulas for a given formula. The notion of a base is introduced generally for a category equipped with a suitable factorisation system.
We prove that the resulting logic has the Hennessy-Milner property for the notion of similarity based on the notion of relation lifting. We define a sequent proof system for the logic, and prove its completeness.
△ Less
Submitted 6 August, 2022; v1 submitted 19 January, 2019;
originally announced January 2019.
