Set-to-Sequence Methods in Machine Learning: a Review
arXiv:2103.09656 · doi:10.1613/jair.1.12839
Abstract
Machine learning on sets towards sequential output is an important and ubiquitous task, with applications ranging from language modeling and meta-learning to multi-agent strategy games and power grid optimization. Combining elements of representation learning and structured prediction, its two primary challenges include obtaining a meaningful, permutation invariant set representation and subsequently utilizing this representation to output a complex target permutation. This paper provides a comprehensive introduction to the field as well as an overview of important machine learning methods tackling both of these key challenges, with a detailed qualitative comparison of selected model architectures.
arXiv comment46 pages of text, with 10 pages of references. Contains 2 tables and 4 figures. Updated version includes expanded notes on method comparison
References in corpus (49)
- Neural Machine Translation by Jointly Learning to Align and Translate
- Sequence to Sequence Learning with Neural Networks
- Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
- Inductive Representation Learning on Large Graphs
- Language Models are Few-Shot Learners
- PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation
- Learning Combinatorial Optimization Algorithms over Graphs
- Learning Convolutional Neural Networks for Graphs
- Seq2SQL: Generating Structured Queries from Natural Language using Reinforcement Learning
- Reasoning about Entailment with Neural Attention
- Learning Representations and Generative Models for 3D Point Clouds
- MolGAN: An implicit generative model for small molecular graphs
- Universal Transformers
- Adaptive Computation Time for Recurrent Neural Networks
- Actor-Attention-Critic for Multi-Agent Reinforcement Learning
- TransferTransfo: A Transfer Learning Approach for Neural Network Based Conversational Agents
- Exploiting Unlabeled Data in CNNs by Self-supervised Learning to Rank
- Attention, Learn to Solve Routing Problems!
- Reinforcement Learning for Solving the Vehicle Routing Problem
- Set Transformer: A Framework for Attention-based Permutation-Invariant Neural Networks
- Robust Attentional Aggregation of Deep Feature Sets for Multi-view 3D Reconstruction
- Learning Latent Permutations with Gumbel-Sinkhorn Networks
- Theoretical Limitations of Self-Attention in Neural Sequence Models
- Sequence-to-Sequence Learning as Beam-Search Optimization
- CondenseNet: An Efficient DenseNet using Learned Group Convolutions
- Optimizing Rank-based Metrics with Blackbox Differentiation
- Deep Set Prediction Networks
- Sentence Ordering and Coherence Modeling using Recurrent Neural Networks
- Janossy Pooling: Learning Deep Permutation-Invariant Functions for Variable-Size Inputs
- Dynamical Mass Measurements of Contaminated Galaxy Clusters Using Machine Learning
- FSPool: Learning Set Representations with Featurewise Sort Pooling
- Seq2Slate: Re-ranking and Slate Optimization with RNNs
- Dynamic Neural Networks: A Survey
- An Introductory Survey on Attention Mechanisms in NLP Problems
- Learning Permutations with Sinkhorn Policy Gradient
- Stochastic Optimization of Sorting Networks via Continuous Relaxations
- Rep the Set: Neural Networks for Learning Set Representations
- On the Limitations of Representing Functions on Sets
- Conditional Set Generation with Transformers
- Unsupervised Adversarial Invariance
- Natural Language to Structured Query Generation via Meta-Learning
- Set2Graph: Learning Graphs From Sets
- SetRank: Learning a Permutation-Invariant Ranking Model for Information Retrieval
- Learning Representations of Sets through Optimized Permutations
- Reparameterizing the Birkhoff Polytope for Variational Permutation Inference
- Dynamic Prediction Length for Time Series with Sequence to Sequence Networks
- Deep Architectures for Learning Context-dependent Ranking Functions
- Deep Sets for Generalization in RL
- Meta-Learning for One-Class Classification with Few Examples using Order-Equivariant Network