paper-with-me

홈 › Papers

Pruned Neural Networks are Surprisingly Modular

2020-03-10 · Daniel Filan, Shlomi Hod, Cody Wild, Andrew Critch, Stuart Russell

The learned weights of a neural network are often considered devoid of scrutable internal structure. To discern structure in these weights, we introduce a measurable notion of modularity for multi-layer perceptrons (MLPs), and investigate the modular structure of MLPs trained on datasets of small images. Our notion of modularity comes from the graph clustering literature: a "module" is a set of neurons with strong internal connectivity but weak external connectivity. We find that training and weight pruning produces MLPs that are more modular than randomly initialized ones, and often significantly more modular than random MLPs with the same (sparse) distribution of weights. Interestingly, they are much more modular when trained with dropout. We also present exploratory analyses of the importance of different modules for performance and how modules depend on each other. Understanding the modular structure of neural networks, when such structure exists, will hopefully render their inner workings more interpretable to engineers. Note that this paper has been superceded by "Clusterability in Neural Networks", arxiv:2103.03386 and "Quantifying Local Specialization in Deep Neural Networks", arxiv:2110.08058!

📄 PDF Abstract BibTeX arXiv:2003.04881

Code (1)

shlomihod/shlomihod.github.io

Tasks

ClusteringGraph Clustering

Methods 이 논문이 사용한 방법론

Pruning 설명 없음

Similar Papers 제목 키워드 기반

Scaling Submodular Maximization via Pruned Submodularity Graphs

2016-06-01 · Tianyi Zhou, Hua Ouyang, Yi Chang, Jeff Bilmes 외

We propose a new random pruning method (called "submodular sparsification (SS)") to reduce the cost of submodular maximization. The pruning is applied via a "submodularity graph" over the $n$ ground elements, where each …

Video Summarization

Scalable Submodular Policy Optimization via Pruned Submodularity Graph

2025-07-18 · Aditi Anand, Suman Banerjee, Dildar Ali arxiv

In Reinforcement Learning (abbreviated as RL), an agent interacts with the environment via a set of possible actions, and a reward is generated from some unknown distribution. The task here is to find an optimal set of a…

Reinforcement Learning

Data-Efficient Structured Pruning via Submodular Optimization

2022-03-09 · Marwa El Halabi, Suraj Srinivas, Simon Lacoste-Julien

Structured pruning is an effective approach for compressing large pre-trained neural networks without significantly affecting their performance. However, most current structured pruning methods do not provide any perform…

Investigating Hallucinations in Pruned Large Language Models for Abstractive Summarization

2023-11-15 · George Chrysostomou, Zhixue Zhao, Miles Williams, Nikolaos Aletras

Despite the remarkable performance of generative large language models (LLMs) on abstractive summarization, they face two significant challenges: their considerable size and tendency to hallucinate. Hallucinations are co…

Abstractive Text SummarizationHallucinationHallucination Evaluation

On the Interplay Between Sparsity, Naturalness, Intelligibility, and Prosody in Speech Synthesis

2021-10-04 · Cheng-I Jeff Lai, Erica Cooper, Yang Zhang, Shiyu Chang 외

Are end-to-end text-to-speech (TTS) models over-parametrized? To what extent can these models be pruned, and what happens to their synthesis capabilities? This work serves as a starting point to explore pruning both spec…

Knowledge DistillationSpeech Synthesistext-to-speechText to Speech