Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Innovation Discovery System for Networking Research

Mengrui Zhang, Bang Huang, Yunxin Xu, Haiying Huang, Luxi Zhao, Mochun Long, Qingyu Song, Qiao Xiang, Xue Liu, Jiwu Shu

Abstract

As networking systems become increasingly complex, achieving disruptive innovation grows more challenging. At the same time, recent progress in Large Language Models (LLMs) has shown strong potential for scientific hypothesis formation and idea generation. Nevertheless, applying LLMs effectively to networking research remains difficult for two main reasons: standalone LLMs tend to generate ideas by recombining existing solutions, and current open-source networking resources do not provide the structured, idea-level knowledge necessary for data-driven scientific discovery. To bridge this gap, we present SciNet, a research idea generation system specifically designed for networking. SciNet is built upon three key components: (1) constructing a networking-oriented scientific discovery dataset from top-tier networking conferences, (2) simulating the human idea discovery workflow through problem setting, inspiration retrieval, and idea generation, and (3) developing an idea evaluation method that jointly measures novelty and practicality. Experimental results show that \system consistently produces practical and novel networking research ideas across multiple LLM backbones, and outperforms standalone LLM-based generation in overall idea quality.

Innovation Discovery System for Networking Research

Abstract

As networking systems become increasingly complex, achieving disruptive innovation grows more challenging. At the same time, recent progress in Large Language Models (LLMs) has shown strong potential for scientific hypothesis formation and idea generation. Nevertheless, applying LLMs effectively to networking research remains difficult for two main reasons: standalone LLMs tend to generate ideas by recombining existing solutions, and current open-source networking resources do not provide the structured, idea-level knowledge necessary for data-driven scientific discovery. To bridge this gap, we present SciNet, a research idea generation system specifically designed for networking. SciNet is built upon three key components: (1) constructing a networking-oriented scientific discovery dataset from top-tier networking conferences, (2) simulating the human idea discovery workflow through problem setting, inspiration retrieval, and idea generation, and (3) developing an idea evaluation method that jointly measures novelty and practicality. Experimental results show that \system consistently produces practical and novel networking research ideas across multiple LLM backbones, and outperforms standalone LLM-based generation in overall idea quality.

Paper Structure

This paper contains 17 sections, 3 equations, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Overview
  3. Dataset $\&$ Knowledge Graph Construction
  4. Paper Summary Dataset
  5. Paper Graph
  6. Citation Graph
  7. Idea Discovery
  8. Problem Setting
  9. Inspiration Retrieval
  10. Idea Generation
  11. Idea Evaluation Approach
  12. Evaluation
  13. Evaluation Setup
  14. Idea Quality Comparison
  15. Ablation Study
  16. ...and 2 more sections

Figures (4)

  • Figure 1: Framework: 1.1SciNet summarizes SIGCOMMsigcomm and NSDInsdi papers into a structured dataset, builds knowledge graphs, retrieves methods for a user-specified domain and problem, generates and iteratively refines ideas with an LLM, and evaluates their novelty and practicality against the dataset.
  • Figure 2: Workflow of Idea Discovery
  • Figure 3: Subsets of the 1.1SciNet Paper and Citation Graphs: (a) the Paper Graph links domains, problems, papers, and methods; (b) the Citation Graph links papers to cited papers and methods.
  • Figure 4: Ablation study of 1.1SciNet across different LLM backbones.