Prediction Loss

Prediction loss, the discrepancy between a model's predictions and actual outcomes, is a central challenge across diverse machine learning applications. Current research focuses on minimizing this loss through various strategies, including developing novel algorithms for efficient optimization (e.g., in inverse optimization problems and high-dimensional settings), learning representations that are robust to adversarial prediction tasks, and adapting loss functions to better align with downstream task objectives. Reducing prediction loss is crucial for improving model accuracy, efficiency, and reliability in applications ranging from resource allocation and weather forecasting to large language model pre-training and active learning in perception systems.

Papers

October 21, 2024

LLM-TS Integrator: Integrating LLM for Enhanced Time Series Modeling
Can Chen, Gabriel Oliveira, Hossein Sharifi Noghabi, Tristan Sylvain
Medical LLM Time Series Long Term Forecasting LLM Integrated Application Prediction Loss Mutual Information Loss

May 23, 2024

A fast algorithm to minimize prediction loss of the optimal solution in inverse optimization problem of MILP
Akira Kitaoka
Fast Algorithm Optimal Solution Prediction Loss

March 11, 2024

A representation-learning game for classes of prediction tasks
Neria Uzan, Nir Weinberger
Loss Function Low Dimensional Representation Feature Vector Prediction Task Visual Based Class Optimal Representation Prediction Loss Game Representation

December 8, 2023

TaskMet: Task-Driven Metric Learning for Model Learning
Dishank Bansal, Ricky T. Q. Chen, Mustafa Mukadam, Brandon Amos
Deep Learning Model Loss Function Metric Learning Downstream Task Model Learning Task Loss Prediction Loss

September 11, 2023

Stream-based Active Learning by Exploiting Temporal Properties in Perception with Temporal Predicted Loss
Sebastian Schmidt, Stephan Günnemann
Deep Learning Autonomous Vehicle Active Learning Perception Aware Temporal Feature Pool Based Active Learning Prediction Loss

April 14, 2023

nanoLM: an Affordable LLM Pre-training Benchmark via Accurate Loss Prediction across Scales
Yiqun Yao, Siqi fan, Xiusheng Huang, Xuezhi Fang, Xiang Li, Ziyi Ni, Xin Jiang, Xuying Meng, Peng Han, Shuo Shang, Kang Liu, Aixin Sun, Yequan Wang
Language Model New Benchmark Large Model Ghanaian Musical Scale Scaling Behavior Hyperparameter Space Prediction Loss Resource Intensive LLM

January 11, 2023

Loss-Controlling Calibration for Predictive Models
Di Wang, Junzhi Shi, Pingping Wang, Shuo Zhuang, Hongyue Li
Loss Function Predictive Model Calibration Performance Prediction Loss

January 6, 2023

Conformal Loss-Controlling Prediction
Di Wang, Ping Wang, Zhong Ji, Xiaojun Yang, Hongyue Li
Loss Function Conformal Prediction Coverage Criterion Conformal Risk Control Prediction Loss

November 9, 2022

A Note on Task-Aware Loss via Reweighing Prediction Loss by Decision-Regret
Connor Lawless, Angela Zhou
Short Note Predict Then Optimize Task Specific Loss Prediction Loss Decision Aware

October 26, 2022

Deep Subspace Encoders for Nonlinear System Identification
Gerben I. Beintema, Maarten Schoukens, Roland Tóth
State Estimation Deep Subspace Nonlinear System Identification Prediction Loss

April 16, 2022

The Tree Loss: Improving Generalization with Many Classes
Yujie Wang, Mike Izbicki
Strong Generalization Stochastic Gradient Descent Supervised ImageNet Cross Entropy Loss Multi Class Classification Class Specific Prediction Loss

January 8, 2022

Forecasting Loss of Signal in Optical Networks with Machine Learning
Wenjie Du, David Cote, Chris Barber, Yan Liu
Machine Learning Machine Learning Model Abnormal Signal Optical Network Multiple Network Prediction Loss