Publications (21)
A Characterization of the Combined Effects of Overlap and Imbalance on the SVM Classifier
Misha Denil, Thomas Trappenberg
In this paper we demonstrate that two common problems in Machine Learning---imbalanced and overlapping data distributions---do not have independent effects on the performance of SV…
Variance-Gated Ensembles: An Epistemic-Aware Framework for Uncertainty Estimation
H. Martin Gillis, Isaac Xu, Thomas Trappenberg
Machine learning applications require fast and reliable per-sample uncertainty estimation. A common approach is to use predictive distributions from Bayesian or approximation metho…
An Information-Geometric Formulation of Pattern Separation and Evaluation of Existing Indices
Harvey Wang, Selena Singh, Thomas Trappenberg +1
Pattern separation is a computational process by which dissimilar neural patterns are generated from similar input patterns. We present an information-geometric formulation of patt…
Multiplicative Decomposition of Heterogeneity in Mixtures of Continuous Distributions
Abraham Nunes, Martin Alda, Thomas Trappenberg
A system's heterogeneity (\textit{diversity}) is the effective size of its event space, and can be quantified using the Rényi family of indices (also known as Hill numbers in ecol…
Improving Detection of Rare Nodes in Hierarchical Multi-Label Learning
Isaac Xu, Martin Gillis, Ayushi Sharma +3
In hierarchical multi-label classification, a persistent challenge is enabling model predictions to reach deeper levels of the hierarchy for more detailed or fine-grained classific…
Covariance Last-Layer Ensembles: Function-Space Diversity for Efficient Uncertainty Quantification
H. Martin Gillis, Isaac Xu, Gabriel Spadon +1
A Last-Layer Ensemble (LLE), linear units on one shared frozen feature map, is an efficient single-pass approach to the disagreement-based epistemic uncertainty for out-of-dist…
BenthicNet: A global compilation of seafloor images for deep learning applications
Scott C. Lowe, Benjamin Misiuk, Isaac Xu +26
Advances in underwater imaging enable collection of extensive seafloor image datasets necessary for monitoring important benthic ecosystems. The ability to collect seafloor imagery…
Skin cancer detection based on deep learning and entropy to detect outlier samples
Andre G. C. Pacheco, Abder-Rahman Ali, Thomas Trappenberg
We describe our methods that achieved the 3rd and 4th places in tasks 1 and 2, respectively, at ISIC challenge 2019. The goal of this challenge is to provide the diagnostic for ski…
Representational Rényi heterogeneity
Abraham Nunes, Martin Alda, Timothy Bardouille +1
A discrete system's heterogeneity is measured by the Rényi heterogeneity family of indices (also known as Hill numbers or Hannah--Kay indices), whose units are {the numbers equiva…
Label-free Monitoring of Self-Supervised Learning Progress
Isaac Xu, Scott Lowe, Thomas Trappenberg
Self-supervised learning (SSL) is an effective method for exploiting unlabelled data to learn a high-level embedding space that can be used for various downstream tasks. However, e…
Classifier with Hierarchical Topographical Maps as Internal Representation
Thomas Trappenberg, Paul Hollensen, Pitoyo Hartono
In this study we want to connect our previously proposed context-relevant topographical maps with the deep learning community. Our architecture is a classifier with hidden layers t…
Hierarchical Multi-Label Classification with Missing Information for Benthic Habitat Imagery
Isaac Xu, Benjamin Misiuk, Scott C. Lowe +3
In this work, we apply state-of-the-art self-supervised learning techniques on a large dataset of seafloor imagery, \textit{BenthicNet}, and study their performance for a complex h…
Uncertainty Estimation using Variance-Gated Distributions
H. Martin Gillis, Isaac Xu, Thomas Trappenberg
Evaluation of per-sample uncertainty quantification from neural networks is essential for decision-making involving high-risk applications. A common approach is to use the predicti…
Masked strategies for images with small objects
H. Martin Gillis, Ming Hill, Paul Hollensen +2
The hematology analytics used for detection and classification of small blood components is a significant challenge. In particular, when objects exists as small pixel-sized entitie…
Platelet enumeration in dense aggregates
H. Martin Gillis, Yogeshwar Shendye, Paul Hollensen +2
Identifying and counting blood components such as red blood cells, various types of white blood cells, and platelets is a critical task for healthcare practitioners. Deep learning…
Logical Activation Functions: Logit-space equivalents of Probabilistic Boolean Operators
Scott C. Lowe, Robert Earle, Jason d'Eon +2
The choice of activation functions and their motivation is a long-standing issue within the neural network community. Neuronal representations within artificial neural networks are…
LogAvgExp Provides a Principled and Performant Global Pooling Operator
Scott C. Lowe, Thomas Trappenberg, Sageev Oore
We seek to improve the pooling operation in neural networks, by applying a more theoretically justified operator. We demonstrate that LogSumExp provides a natural OR operator for l…
Last-layer committee machines for uncertainty estimations of benthic imagery
H. Martin Gillis, Isaac Xu, Benjamin Misiuk +2
Automating the annotation of benthic imagery (i.e., images of the seafloor and its associated organisms, habitats, and geological features) is critical for monitoring rapidly chang…
A Generalized Transformer-based Radio Link Failure Prediction Framework in 5G RANs
Kazi Hasan, Thomas Trappenberg, Israat Haque
Radio link failure (RLF) prediction system in Radio Access Networks (RANs) is critical for ensuring seamless communication and meeting the stringent requirements of high data rates…
Spontaneous symmetry breaking on the lattice generated by Yukawa interaction
Wolfgang Bock, Asit K. De, Christoph Frick +2
We study by numerical simulation a lattice Yukawa model with naive fermions at intermediate values of the Yukawa coupling when the nearest neighbour coupling $\kp$ of the scala…
Uncertainty quantification for trustworthy deep learning: Methods and measures
H. Martin Gillis, Thomas Trappenberg
The paper surveys methods for quantifying uncertainty in deep neural networks, focusing on ensemble-based and approximate Bayesian approaches and how their outputs are measured.