Differentiable Function

Differentiable functions are crucial in machine learning because they allow for efficient gradient-based optimization, enabling the training of complex models. Current research focuses on developing differentiable approximations of non-differentiable functions, such as dynamic time warping and the L0 norm, to improve the efficiency and applicability of algorithms in various domains, including image registration, energy-efficient deep learning, and reinforcement learning. These approximations are often implemented using neural networks, particularly convolutional and Siamese architectures, leading to faster and more accurate solutions for tasks like multimodal image analysis and time series analysis. This work has significant implications for advancing the capabilities and efficiency of machine learning models across a wide range of applications.

Papers

July 19, 2023

DISA: DIfferentiable Similarity Approximation for Universal Multimodal Registration
Matteo Ronchetti, Wolfgang Wein, Nassir Navab, Oliver Zettinig, Raphael Prevost
Splitting Algorithm Registration Algorithm Modality Pair Differentiable Function

July 1, 2023

Minimizing Energy Consumption of Deep Learning Models by Energy-Aware Training
Dario Lazzaro, Antonio Emanuele Cinà, Maura Pintor, Ambra Demontis, Battista Biggio, Fabio Roli, Marcello Pelillo
Deep Neural Network Deep Learning Model Energy Efficiency Gradient Based Algorithm Energy Efficient Training Differentiable Function

January 30, 2023

Approximating DTW with a convolutional neural network on EEG data
Hugo Lerogeron, Romain Picot-Clemente, Alain Rakotomamonjy, Laurent Heutte
Convolutional Neural Network Time Series EEG Data EEG Datasets Differentiable Function

October 9, 2022

A Differentiable Distance Approximation for Fairer Image Classification
Nicholas Rosa, Tom Drummond, Mehrtash Harandi
Absolute Stance Bias Artificial Intelligence Model AI Model Fairness Measure Differentiable Function

October 3, 2022

Offline Reinforcement Learning with Differentiable Function Approximation is Provably Efficient
Ming Yin, Mengdi Wang, Yu-Xiang Wang
Reinforcement Learning Offline Reinforcement Learning Q Learning Differentiable Function

February 10, 2022

Off-Policy Fitted Q-Evaluation with Differentiable Function Approximators: Z-Estimation and Inference Theory
Ruiqi Zhang, Xuezhou Zhang, Chengzhuo Ni, Mengdi Wang
Policy Evaluation Policy Estimation Q Evaluation Differentiable Function