Parametric Activation Functions for Neural Networks: A Tutorial Survey

LS Pusztaházi, G Eigner, O Csiszár - IEEE Access, 2024‏ - ieeexplore.ieee.org
Activation functions are pivotal in neural networks, determining the output of each neuron.
Traditionally, functions like sigmoid and ReLU have been static and deterministic. However …

Short-Term Forecasting of Daily Reference Crop Evapotranspiration Based on Calibrated Hargreaves–Samani Equation at Regional Scale

S Baber, K Ullah - Earth Systems and Environment, 2024‏ - Springer
This study focuses on enhancing real-time irrigation decisions and stream flow forecasts
using short-term daily forecasts of reference evapotranspiration (ETo). While conventional …

[HTML][HTML] Uninorm-like parametric activation functions for human-understandable neural models

O Csiszár, LS Pusztaházi, L Dénes-Fazakas… - Knowledge-Based …, 2023‏ - Elsevier
We present a deep learning model for finding human-understandable connections between
input features. Our approach uses a parameterized, differentiable activation function, based …

Esc-rules: Explainable, semantically constrained rule sets

M Glauer, R West, S Michie, J Hastings - ar** of linguistic variable in membership functions
MAH Ali, S Mekhilef, N Yusoff, B Abd Razak - Journal of King Saud …, 2022‏ - Elsevier
A new approach for Fuzzification and Defuzzification processes of a high degree of
overlap** between the linguistic variables through proportional and relative-dynamic …

Parametric activation functions modelling fuzzy connectives for better explainability of neural models

LS Pusztaházi, G Csiszár, MS Gashler… - 2022 IEEE 20th …, 2022‏ - ieeexplore.ieee.org
In this work, a neuro-fuzzy hybrid deep learning model is presented for finding human-
readable relationships between input features with the help of nilpotent fuzzy logic and multi …

Optimizing neural networks through activation function discovery and automatic weight initialization

G Bingham - arxiv preprint arxiv:2304.03374, 2023‏ - arxiv.org
Automated machine learning (AutoML) methods improve upon existing models by
optimizing various aspects of their design. While present methods focus on hyperparameters …

[ספר][B] Parameterizing and aggregating activation functions in deep neural networks

LB Godfrey - 2018‏ - search.proquest.com
The nonlinear activation functions applied by each neuron in a neural network are essential
for making neural networks powerful representational models. If these are omitted, even …

Computing with words—A framework for human-computer interaction

D Tamir, S Neumann, N Rishe, A Kandel… - … , AC 2019, Held as Part of …, 2019‏ - Springer
In this paper we explore the possibility of using computation with words (CWW) systems and
CWW-based human-computer interface (HCI) and interaction to enable efficient computation …

Logical activation functions: logit-space equivalents of Boolean operators

SC Lowe, R Earle, J d'Eon, T Trappenberg, S Oore - 2021‏ - openreview.net
Neuronal representations within artificial neural networks are commonly understood as
logits, representing the log-odds score of presence (versus absence) of features within the …