mlsanityData to 7 Oct 2026

Learning Beam Search Policies via Imitation Learning

arXiv 1811.00512 ↗First posted 1 Nov 2018stat.ML cs.AI cs.LGcited by 9

On the map: Offline RL and imitation learning from data → Imitation learning methods

Abstract (arXiv)

Beam search is widely used for approximate decoding in structured prediction problems. Models often use a beam at test time but ignore its existence at train time, and therefore do not explicitly learn how to use the beam. We develop an unifying meta-algorithm for learning beam search policies using imitation learning. In our setting, the beam is part of the model, and not just an artifact of approximate decoding. Our meta-algorithm captures existing learning algorithms and suggests new ones. It also lets us show novel no-regret guarantees for learning beam search policies.

Key facts. Cited by 9 papers in the map (data through 7 Oct 2026). Placed in Imitation learning methods, one of 13 directions in Offline RL and imitation learning from data. Builds on 11 of its 25 references that are in the map. First posted on arXiv 1 Nov 2018.

Builds on

11 of 25 references.

References · 11

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2014Sequence to Sequence Learning with Neural Networkscited by 2,974
  3. 2014Show and Tell: A Neural Image Caption Generatorcited by 877
  4. 2015Scheduled Sampling for Sequence Prediction with Recurrent Neural Networkscited by 779
  5. 2013Speech Recognition with Deep Recurrent Neural Networkscited by 666
  6. 2016Sequence-to-Sequence Learning as Beam-Search Optimizationcited by 126
  7. 2014Reinforcement and Imitation Learning via Interactive No-Regret Learningcited by 116
  8. 2015Learning to Search Better Than Your Teachercited by 50
  9. 2016Globally Normalized Transition-Based Neural Networkscited by 26
  10. 2015Structured Training for Neural Network Transition-Based Parsingcited by 11
  11. 2017A Continuous Relaxation of Beam Search for End-to-end Training of Neural Sequence Modelscited by 10

Built on by

Cited by year

Citing papers per year: 2021: 1; 2022: 3; 2023: 3; 2024: 2.

9 papers.

  1. 2022Simulation-guided Beam Search for Neural Combinatorial Optimizationcited by 84
  2. 2022CrossBeam: Learning to Search in Bottom-Up Program Synthesiscited by 18
  3. 2023Learning Deductive Reasoning from Synthetic Corpus based on Formal Logiccited by 18
  4. 2023Transformers in Reinforcement Learning: A Surveycited by 11
  5. 2022Generative Cooperative Networks for Natural Language Generationcited by 8
  6. 2021Incremental Beam Manipulation for Natural Language Generationcited by 6
  7. 2023LambdaBeam: Neural Program Search with Higher-Order Functions and Lambdascited by 6
  8. 2024GraphKD: Exploring Knowledge Distillation Towards Document Object Detection with Structured Graph Creationcited by 2
  9. 2024Improving Autoregressive Training with Dynamic Oraclescited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024Amortized nonmyopic active search via deep imitation learning
  2. 2025Diverse Transformer Decoding for Offline Reinforcement Learning Using Financial Algorithmic Approaches
  3. 2026Search or Accelerate: Confidence-Switched Position Beam Search for Diffusion Language Models
  4. 2020Active Imitation Learning with Noisy Guidance
  5. 2026Beam Search as Test-Time Self-Distillation via Counterfactual Contexts
  6. 2026Provable Test-Time Scaling for Beam Search in LLM Reasoning
  7. 2018Learning to Search via Retrospective Imitation
  8. 2018Learning Neural Parsers with Deterministic Differentiable Imitation Learning