Equivalence Class

Equivalence classes represent sets of objects deemed equivalent based on a specific criterion, a concept crucial across diverse fields. Current research focuses on identifying and characterizing these classes within various contexts, including code functionality in large language models, reward functions in reinforcement learning, and graph structures in neural networks. This work employs diverse methods, ranging from analyzing second-order statistics in causal inference to developing novel algorithms for characterizing functional equivalence in neural networks. Understanding and manipulating equivalence classes offers significant potential for improving the efficiency and robustness of machine learning models, enhancing the interpretability of complex systems, and advancing our understanding of fundamental concepts in computer science and artificial intelligence.

Papers

August 20, 2024

What can Large Language Models Capture about Code Functional Equivalence?
Nickil Maveli, Antonio Vergari, Shay B. Cohen
Large Language Model Code LLM Code Semantics Code Corpus Code Transformation Equivalence Class

May 6, 2024

Functional Equivalence with NARS
Robert Johansson, Patrick Hammer, Tony Lofthouse
Human Like Human Level Non Axiomatic Reasoning Perceptual Similarity Equivalence Class

March 9, 2024

A Generalized Acquisition Function for Preference-based Reward Learning
Evan Ellis, Gaurav R. Ghosal, Stuart J. Russell, Anca Dragan, Erdem Bıyık
Reward Function Reward Learning Information Gain Equivalence Class Reward Function Parameter

November 8, 2023

Learning Linear Gaussian Polytree Models with Interventions
D. Tramontano, L. Waldmann, M. Drton, E. Duarte
Causal Structure Health Intervention Interventional Data Equivalence Class

May 23, 2023

Beyond Shared Vocabulary: Increasing Representational Word Similarities across Languages for Multilingual Machine Translation
Di Wu, Christof Monz
Jina Embeddings Word Embeddings Unknown Language Multilingual Machine Translation Multilingual Neural Machine Translation Representational Similarity Word Level Equivalence Class

May 19, 2023

Exploring the Complexity of Deep Neural Networks through Functional Equivalence
Guohao Shen
Deep Learning Deep Neural Network Complexity Matter Overparametrization Bound Equivalence Class Effective Parameter

May 8, 2023

Functional Equivalence and Path Connectivity of Reducible Hyperbolic Tangent Networks
Matthew Farrugia-Roberts
Hyperbolic Neural Network Equivalence Class

August 20, 2022

Calculus on MDPs: Potential Shaping as a Gradient
Erik Jenner, Herke van Hoof, Adam Gleave
Reinforcement Learning Full Potential Optimal Policy Natural Gradient First Integral Equivalence Class

June 23, 2022

A Topological characterisation of Weisfeiler-Leman equivalence classes
Jacob Bamberger
Graph Neural Network Topological Feature GNN Architecture Discriminative Graph Equivalence Class

March 1, 2022

A Transformational Characterization of Unconditionally Equivalent Bayesian Networks
Alex Markham, Danai Deligeorgaki, Pratik Misra, Liam Solus
Bayesian Network Markov Equivalence Class Equivalence Class Strong Equivalence

December 17, 2021

A singular Riemannian geometry approach to Deep Neural Networks II. Reconstruction of 1-D equivalence classes
Alessandro Benfenati, Alessio Marta
Deep Neural Network Full State Reconstruction Statistical Manifold Riemannian Geometry Approach Neural Mapping Geometric Framework Equivalence Class