Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

NeuroGen: Neural Network Parameter Generation via Large Language Models

Jiaqi Wang, Yusen Zhang, Xi Li

TL;DR

Neural network parameters are traditionally obtained through gradient-based optimization. NeuroGen introduces a two-stage approach to generate NN weights with large language models, using Stage 1 for parameter-reference knowledge injection and Stage 2 for context-enhanced instruction tuning, guiding the LLM to produce usable weights conditioned on data, task, and architecture. Experimental results on both image and text classification show that LLM-generated parameters can be functional and task-adaptive, especially for simple architectures and in low-data regimes, indicating a promising new paradigm for data-efficient model construction. This work opens a direction for instruction-guided, multimodal, and edge-deployable neural parameter synthesis, bridging generative modeling and network design through language models.

Abstract

Acquiring the parameters of neural networks (NNs) has been one of the most important problems in machine learning since the inception of NNs. Traditional approaches, such as backpropagation and forward-only optimization, acquire parameters via iterative data fitting to gradually optimize them. This paper aims to explore the feasibility of a new direction: acquiring NN parameters via large language model generation. We propose NeuroGen, a generalized and easy-to-implement two-stage approach for NN parameter generation conditioned on descriptions of the data, task, and network architecture. Stage one is Parameter Reference Knowledge Injection, where LLMs are pretrained on NN checkpoints to build foundational understanding of parameter space, whereas stage two is Context-Enhanced Instruction Tuning, enabling LLMs to adapt to specific tasks through enriched, task-aware prompts. Experimental results demonstrate that NeuroGen effectively generates usable NN parameters. Our findings highlight the feasibility of LLM-based NN parameter generation and suggest a promising new paradigm where LLMs and lightweight NNs can coexist synergistically

NeuroGen: Neural Network Parameter Generation via Large Language Models

TL;DR

Neural network parameters are traditionally obtained through gradient-based optimization. NeuroGen introduces a two-stage approach to generate NN weights with large language models, using Stage 1 for parameter-reference knowledge injection and Stage 2 for context-enhanced instruction tuning, guiding the LLM to produce usable weights conditioned on data, task, and architecture. Experimental results on both image and text classification show that LLM-generated parameters can be functional and task-adaptive, especially for simple architectures and in low-data regimes, indicating a promising new paradigm for data-efficient model construction. This work opens a direction for instruction-guided, multimodal, and edge-deployable neural parameter synthesis, bridging generative modeling and network design through language models.

Abstract

Acquiring the parameters of neural networks (NNs) has been one of the most important problems in machine learning since the inception of NNs. Traditional approaches, such as backpropagation and forward-only optimization, acquire parameters via iterative data fitting to gradually optimize them. This paper aims to explore the feasibility of a new direction: acquiring NN parameters via large language model generation. We propose NeuroGen, a generalized and easy-to-implement two-stage approach for NN parameter generation conditioned on descriptions of the data, task, and network architecture. Stage one is Parameter Reference Knowledge Injection, where LLMs are pretrained on NN checkpoints to build foundational understanding of parameter space, whereas stage two is Context-Enhanced Instruction Tuning, enabling LLMs to adapt to specific tasks through enriched, task-aware prompts. Experimental results demonstrate that NeuroGen effectively generates usable NN parameters. Our findings highlight the feasibility of LLM-based NN parameter generation and suggest a promising new paradigm where LLMs and lightweight NNs can coexist synergistically
Paper Structure (22 sections, 2 equations, 3 figures, 1 table, 1 algorithm)

This paper contains 22 sections, 2 equations, 3 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Neural Network Parameter Acquisition
  4. Large Lauguage Model
  5. Key Differences from Existing Work
  6. Methodology
  7. Overview
  8. Notations
  9. Stage 1: Parameter Reference Knowledge Injection
  10. Input Preparation
  11. Neural Network Alignment Learning
  12. Stage 2: Context-enhanced Instruction Tuning
  13. Triplet Input Construction
  14. Context Understanding Training Procedure
  15. Experiments
  16. ...and 7 more sections

Figures (3)

  • Figure 1: Framework of NeuroGen.
  • Figure 2: Ablation study on MNIST with CNN generation and AG News with MLP generation. (a) and (b): Training loss of NeuroGen versus Phase-2-only implementation on MNIST and AG News, respectively. (c) and (d): Test accuracy of NeuroGen versus Phase-2-only implementation on MNIST and AG News, respectively.
  • Figure 3: Model generation study on SVHN and CIFAR-10 datasets using CNN generation. Training loss and accuracy on SVHN with (a–b) sufficient data and (c–d) limited data; and on CIFAR-10 with (e–f) sufficient data and (g–h) limited data.