On Computational Limits of Modern Hopfield Models: A Fine-Grained Complexity Analysis

Authors: Jerry Yao-Chieh Hu, Thomas Lin, Zhao Song, Han Liu

ICML 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We investigate the computational limits of the memory retrieval dynamics of modern Hopfield models from the fine-grained complexity analysis. Our key contribution is the characterization of a phase transition behavior... By the formal nature of this work, our results do not lead to practical implementations.
Researcher Affiliation Collaboration 1Department of Computer Science, Northwestern University, Evanston, IL, USA 2Department of Physics, National Taiwan University, Taipei, Taiwan 3Adobe Research, Seattle, WA, USA 4Department of Statistics and Data Science, Northwestern University, Evanston, IL, USA.
Pseudocode Yes Algorithm 1 The algorithm to solve AHop
Open Source Code No The paper is theoretical and explicitly states in its 'Limitation' section: 'By the formal nature of this work, our results do not lead to practical implementations.' Thus, no source code is provided.
Open Datasets No The paper is a theoretical work focusing on complexity analysis and does not involve experiments with publicly available datasets for training, validation, or testing.
Dataset Splits No The paper is a theoretical work and does not involve empirical validation with dataset splits for training, validation, or testing.
Hardware Specification No The paper is purely theoretical and does not describe any experiments that would require hardware specifications.
Software Dependencies No The paper is theoretical and does not detail any software dependencies or versions required for replicating experiments.
Experiment Setup No The paper is a theoretical work focusing on complexity analysis and does not describe any experimental setup details such as hyperparameters or training configurations.