Appropriate Penalty Term

Appropriate penalty terms in machine learning aim to improve model performance and address specific limitations, such as enforcing constraints, promoting fairness, or enhancing energy efficiency. Current research focuses on developing and refining penalty methods within various model architectures, including deep neural networks, reinforcement learning agents, and large language models, often employing techniques like L1 regularization, gradient penalties, and reward shaping. These advancements have implications for diverse applications, from improving the fairness and robustness of AI systems to optimizing resource consumption in spiking neural networks and enhancing the efficiency of training processes.

Papers

October 24, 2024

Revisiting Differentiable Structure Learning: Inconsistency of $\ell_1$ Penalty and Beyond
Kaifeng Jin, Ignavier Ng, Kun Zhang, Biwei Huang
Structure Learning Hard to Easy Inconsistency Markov Equivalence Class Appropriate Penalty Term

September 29, 2024

Large Language Models Overcome the Machine Penalty When Acting Fairly but Not When Acting Selfishly or Altruistically
Zhen Wang, Ruiqi Song, Chen Shen, Shiya Yin, Zhao Song, Balaraju Battu, Lei Shi, Danyang Jia, Talal Rahwan, Shuyue Hu
Long Term Cooperation Social Dilemma Selfish Agent Appropriate Penalty Term

August 19, 2024

Minor DPO reject penalty to increase training robustness
Shiming Xie, Hong Chen, Fred Yu, Zeye Sun, Xiuyu Wu, Yingfan Hu
Reinforcement Learning Direct Preference Optimization Preference Optimization Preference Pair Implicit Reward Appropriate Penalty Term Training Robustness

May 31, 2024

Shape Constraints in Symbolic Regression using Penalized Least Squares
Viktor Martinek, Julia Reuter, Ophelia Frotscher, Sanaz Mostaghim, Markus Richter, Roland Herzog
Symbolic Regression Least Square Parameter Estimation Shape Constraint Appropriate Penalty Term

March 22, 2024

An Agent-Centric Perspective on Norm Enforcement and Sanctions
Elena Yan, Luis G. Nardin, Jomi F. Hübner, Olivier Boissier
Multi Agent System Normative System Appropriate Penalty Term Norm Violation Agent Centric

September 23, 2023

Penalties and Rewards for Fair Learning in Paired Kidney Exchange Programs
Margarida Carvalho, Alison Caulfield, Yi Lin, Adrian Vetta
Reward Report Fair Learning Appropriate Penalty Term Egalitarian Fairness

August 30, 2023

Penalization Framework For Autonomous Agents Using Answer Set Programming
Vineel S. K. Tummala
Autonomous Agent Answer Set Programming Intelligent Agent Prior Policy Appropriate Penalty Term

August 3, 2023

Baby Llama: knowledge distillation from an ensemble of teachers trained on a small dataset with no performance penalty
Inar Timiryasov, Jean-Loup Tastet
Language Model Knowledge Distillation Diverse Ensemble Good Teacher Small Datasets LLaMA Model Appropriate Penalty Term BabyLM Challenge

May 10, 2023

Penalized deep neural networks estimator with general loss functions under weak dependence
William Kengne, Modou Wade
Loss Function Appropriate Penalty Term Deep Neural Network Estimator

February 3, 2023

Spiking Synaptic Penalty: Appropriate Penalty Term for Energy-Efficient Spiking Neural Networks
Kazuma Suetake, Takuya Ushimaru, Ryuji Saiin, Yoshihide Sawada
SNN Model Accuracy Trade Energy Efficient Neural Network Appropriate Penalty Term

January 27, 2023

Solving Richly Constrained Reinforcement Learning through State Augmentation and Reward Penalties
Hao Jiang, Tien Mai, Pradeep Varakantham, Minh Huy Hoang
Constrained Reinforcement Learning State Augmentation Cost Constraint Appropriate Penalty Term

October 3, 2022

spred: Solving $L_1$ Penalty with SGD
Liu Ziyin, Zihao Wang
Stochastic Gradient Descent Differentiable Solver Appropriate Penalty Term Differentiable Objective Norm Penalty Flow Based Reparametrization

August 8, 2022

Controlled Sparsity via Constrained Optimization or: How I Learned to Stop Tuning Penalties and Love Constraints
Jose Gallego-Posada, Juan Ramirez, Akram Erraqabi, Yoshua Bengio, Simon Lacoste-Julien
Constrained Optimization Sparse Model Physical Constraint Appropriate Penalty Term Model Sparsity Sparsity Penalty Target Sparsity

June 24, 2022

ML-Based Approach for NFL Defensive Pass Interference Prediction Using GPS Tracking Data
Arian Skoki, Jonatan Lerga, Ivan Štajduhar
Next Generation GPS Trajectory Appropriate Penalty Term

May 17, 2022

Monotonicity Regularization: Improved Penalties and Novel Applications to Disentangled Representation Learning and Robust Classification
Joao Monteiro, Mohamed Osama Ahmed, Hossein Hajimirsadeghi, Greg Mori
Generative Modeling Disentangled Representation Robust Classification Risk Minimization Partial Monotonicity Penalty Method Novel Application Monotonic Neural Network Appropriate Penalty Term

April 5, 2022

Comment on "Black Box Prediction Methods in Sports Medicine Deserve a Red Card for Reckless Practice: A Change of Tactics is Needed to Advance Athlete Care"
Jakim Berndsen, Derek McHugh
Machine Learning Model Black Box Appropriate Penalty Term Running Related Injury

Appropriate Penalty Term

Papers

Revisiting Differentiable Structure Learning: Inconsistency of $\ell_1$ Penalty and Beyond

Large Language Models Overcome the Machine Penalty When Acting Fairly but Not When Acting Selfishly or Altruistically

Minor DPO reject penalty to increase training robustness

Shape Constraints in Symbolic Regression using Penalized Least Squares

An Agent-Centric Perspective on Norm Enforcement and Sanctions

Penalties and Rewards for Fair Learning in Paired Kidney Exchange Programs

Penalization Framework For Autonomous Agents Using Answer Set Programming

Baby Llama: knowledge distillation from an ensemble of teachers trained on a small dataset with no performance penalty

Penalized deep neural networks estimator with general loss functions under weak dependence

Spiking Synaptic Penalty: Appropriate Penalty Term for Energy-Efficient Spiking Neural Networks

Solving Richly Constrained Reinforcement Learning through State Augmentation and Reward Penalties

spred: Solving $L_1$ Penalty with SGD

Controlled Sparsity via Constrained Optimization or: How I Learned to Stop Tuning Penalties and Love Constraints

ML-Based Approach for NFL Defensive Pass Interference Prediction Using GPS Tracking Data

Monotonicity Regularization: Improved Penalties and Novel Applications to Disentangled Representation Learning and Robust Classification

Comment on "Black Box Prediction Methods in Sports Medicine Deserve a Red Card for Reckless Practice: A Change of Tactics is Needed to Advance Athlete Care"