• Home
  • Search
  • ARM-Net: Adaptive Relation Modeling Network for Structured Data
  • Cite Icon35
  • https://doi.org/10.1145/3448016.3457321Copy DOI Icon

ARM-Net: Adaptive Relation Modeling Network for Structured Data

  • Jun 9, 2021
  • Shaofeng Cai +5 more
Show More
  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Relational databases are the de facto standard for storing and querying structured data, and extracting insights from structured data requires advanced analytics. Deep neural networks (DNNs) have achieved super-human prediction performance in particular data types, e.g., images. However, existing DNNs may not produce meaningful results when applied to structured data. The reason is that there are correlations and dependencies across combinations of attribute values in a table, and these do not follow simple additive patterns that can be easily mimicked by a DNN. The number of possible such cross features is combinatorial, making them computationally prohibitive to model. Furthermore, the deployment of learning models in real-world applications has also highlighted the need for interpretability, especially for high-stakes applications, which remains another issue of concern to DNNs. In this paper, we present ARM-Net, an adaptive relation modeling network tailored for structured data, and a lightweight framework ARMOR based on ARM-Net for relational data analytics. The key idea is to model feature interactions with cross features selectively and dynamically, by first transforming the input features into exponential space, and then determining the interaction order and interaction weights adaptively for each cross feature. We propose a novel sparse attention mechanism to dynamically generate the interaction weights given the input tuple, so that we can explicitly model cross features of arbitrary orders with noisy features filtered selectively. Then during model inference, ARM-Net can specify the cross features being used for each prediction for higher accuracy and better interpretability. Our extensive experiments on real-world datasets demonstrate that ARM-Net consistently outperforms existing models and provides more interpretable predictions for data-driven decision making.

Similar Papers
  • Research Article
  • Citations89

Multi-label learning based deep transfer neural network for facial attribute classification

  • Mar 21, 2018
  • Pattern Recognition
  • Ni Zhuang +4
  • Conference Article
  • Citations3

An Efficient Event-driven Neuromorphic Architecture for Deep Spiking Neural Networks

  • Sep 01, 2019
  • Duy-Anh Nguyen +3
  • PDF
  • Research Article
  • Citations35

Brain age prediction using combined deep convolutional neural network and multi-layer perceptron algorithms

  • Dec 16, 2023
  • Scientific Reports
  • Yoonji Joo +8
  • Research Article
  • Citations1

An Invigorating Perspective of a Practical Aspect of NoSQL Database

  • Sep 19, 2022
  • International Journal of Research Publication and Reviews
  • Shaida Begum
  • Conference Article
  • Citations39

Joint Optimization of DNN Partition and Scheduling for Mobile Cloud Computing

  • Aug 09, 2021
  • Yubin Duan +1
  • Research Article
  • Citations154

Feature extraction and genetic algorithm enhanced adaptive deep neural network for energy consumption prediction in buildings

  • Jun 29, 2020
  • Renewable and Sustainable Energy Reviews
  • X.J Luo +5
  • Conference Article
  • Citations484

Can FPGAs Beat GPUs in Accelerating Next-Generation Deep Neural Networks?

  • Feb 22, 2017
  • Eriko Nurvitadhi +10
  • Research Article
  • Citations16

Incremental Verification of Neural Networks

  • Jun 06, 2023
  • Proceedings of the ACM on Programming Languages
  • Shubham Ugare +3
  • PDF
  • Research Article
  • Citations15

A repository based on a dynamically extensible data model supporting multidisciplinary research in neuroscience

  • Oct 08, 2012
  • BMC Medical Informatics and Decision Making
  • Luca Corradi +8
  • PDF
  • Research Article
  • Citations74

Compact and Computationally Efficient Representation of Deep Neural Networks.

  • May 29, 2019
  • IEEE Transactions on Neural Networks and Learning Systems
  • Simon Wiedemann +2
  • PDF
  • Research Article
  • Citations171

Prediction of Students’ Academic Performance Based on Courses’ Grades Using Deep Neural Networks

  • Jan 01, 2021
  • IEEE Access
  • Aya Nabil +2
  • Conference Article
  • Citations6

Accelerating Spectral Normalization for Enhancing Robustness of Deep Neural Networks

  • Jul 01, 2021
  • Zhixin Pan +1
  • Conference Article
  • Citations43

Automated Runtime-Aware Scheduling for Multi-Tenant DNN Inference on GPU

  • Nov 01, 2021
  • Fuxun Yu +6
  • Dissertation

Adversarial multi-view action recognition

  • Jan 01, 2024
  • Deepak Kumar
  • PDF
  • Research Article
  • Citations29

Shallow Unorganized Neural Networks Using Smart Neuron Model for Visual Perception

  • Jan 01, 2019
  • IEEE Access
  • Richard Jiang +1
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.