Loss Function
Loss functions are crucial components of machine learning models, guiding the learning process by quantifying the difference between predicted and actual values. Current research emphasizes developing loss functions tailored to specific challenges, such as class imbalance in classification (addressed through asymmetric losses and hyperparameter distributions) and robustness to noise and outliers (using bounded and smooth alternatives to standard functions like mean squared error). These advancements improve model accuracy, efficiency, and generalizability across diverse applications, including medical image analysis, time series prediction, and physics-informed neural networks. The ongoing exploration of loss function design directly impacts the performance and reliability of machine learning models in various scientific and engineering domains.
Papers
Lessons from Generalization Error Analysis of Federated Learning: You May Communicate Less Often!
Milad Sefidgaran, Romain Chor, Abdellatif Zaidi, Yijun Wan
Acoustic Scene Clustering Using Joint Optimization of Deep Embedding Learning and Clustering Iteration
Yanxiong Li, Mingle Liu, Wucheng Wang, Yuhan Zhang, Qianhua He
Fair Patient Model: Mitigating Bias in the Patient Representation Learned from the Electronic Health Records
Sonish Sivarajkumar, Yufei Huang, Yanshan Wang
Linear Distance Metric Learning with Noisy Labels
Meysam Alishahi, Anna Little, Jeff M. Phillips
On Tail Decay Rate Estimation of Loss Function Distributions
Etrit Haxholli, Marco Lorenzi
Loss-Optimal Classification Trees: A Generalized Framework and the Logistic Case
Tommaso Aldinucci, Matteo Lapucci
Pairwise Ranking Losses of Click-Through Rates Prediction for Welfare Maximization in Ad Auctions
Boxiang Lyu, Zhe Feng, Zachary Robertson, Sanmi Koyejo
Robust T-Loss for Medical Image Segmentation
Alvaro Gonzalez-Jimenez, Simone Lionetti, Philippe Gottfrois, Fabian Gröger, Marc Pouly, Alexander Navarini
OWAdapt: An adaptive loss function for deep learning using OWA operators
Sebastián Maldonado, Carla Vairetti, Katherine Jara, Miguel Carrasco, Julio López
On the Choice of Perception Loss Function for Learned Video Compression
Sadaf Salehkalaibar, Buu Phan, Jun Chen, Wei Yu, Ashish Khisti
Asymptotic Characterisation of Robust Empirical Risk Minimisation Performance in the Presence of Outliers
Matteo Vilucchio, Emanuele Troiani, Vittorio Erba, Florent Krzakala
A Probabilistic Rotation Representation for Symmetric Shapes With an Efficiently Computable Bingham Loss Function
Hiroya Sato, Takuya Ikeda, Koichi Nishiwaki
When Does Optimizing a Proper Loss Yield Calibration?
Jarosław Błasiok, Parikshit Gopalan, Lunjia Hu, Preetum Nakkiran