mlsanityData to 14 Oct 2026

Generalization to New Actions in Reinforcement Learning

arXiv 2011.01928 ↗First posted 3 Nov 2020cs.LG cs.AI cs.ROcited by 25

Not on the map. Why?

Abstract (arXiv)

A fundamental trait of intelligence is the ability to achieve goals in the face of novel circumstances, such as making decisions from new action choices. However, standard reinforcement learning assumes a fixed set of actions and requires expensive retraining when given a new action set. To make learning agents more adaptable, we introduce the problem of zero-shot generalization to new actions. We propose a two-stage framework where the agent first infers action representations from action information acquired separately from the task. A policy flexible to varying action sets is then trained with generalization objectives. We benchmark generalization on sequential tasks, such as selecting from an unseen tool-set to solve physical reasoning puzzles and stacking towers with novel 3D shapes. Videos and code are available at https://sites.google.com/view/action-generalization

Key facts. Cited by 25 papers in the map (data through 14 Oct 2026). First posted on arXiv 3 Nov 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 3; 2022: 5; 2023: 5; 2024: 6; 2025: 5; 2026: 1.

25 papers.

Sort:

By area:

  1. 2021A Survey of Zero-shot Generalisation in Deep Reinforcement Learningcited by 170
  2. 2021Augmenting Reinforcement Learning with Behavior Primitives for Diverse Manipulation Taskscited by 78
  3. 2021Learning to Synthesize Programs as Interpretable and Generalizable Policiescited by 59
  4. 2024Grounding Multimodal Large Language Models in Actionscited by 35
  5. 2024Active Evaluation Acquisition for Efficient LLM Benchmarkingcited by 35
  6. 2023In-Context Reinforcement Learning for Variable Action Spacescited by 24
  7. 2023I-PHYRE: Interactive Physical Reasoningcited by 15
  8. 2022Learn the Time to Learn: Replay Scheduling in Continual Learningcited by 11
  9. 2024SPARTAN: A Sparse Transformer World Model Attending to What Matterscited by 10
  10. 2024A Survey Analyzing Generalization in Deep Reinforcement Learningcited by 8
  11. 2022Creative Problem Solving in Artificially Intelligent Agents: A Survey and Frameworkcited by 4
  12. 2023Dynamic Neighborhood Construction for Structured Large Discrete Action Spacescited by 4
13 more
  1. 2022BusyBot: Learning to Interact, Reason, and Plan in a BusyBoard Environmentcited by 3
  2. 2023Action Pick-up in Dynamic Action Space Reinforcement Learningcited by 3
  3. 2024The Path To Autonomous Cyber Defensecited by 3
  4. 2024Mitigating Suboptimality of Deterministic Policy Gradients in Complex Q-functionscited by 3
  5. 2025Neural Force Field: Few-shot Learning of Generalized Physical Reasoningcited by 1
  6. 2025Attacks and Defenses Against LLM Fingerprintingcited by 1
  7. 2022Learning to reason about and to act on physical cascading eventscited by 0
  8. 2022AIGenC: An AI generalisation model via creativitycited by 0
  9. 2023DCT: Dual Channel Training of Action Embeddings for Reinforcement Learning with Large Discrete Action Spacescited by 0
  10. 2025Zero-Shot Action Generalization with Limited Observationscited by 0
  11. 2025A representational framework for learning and encoding structurally enriched trajectories in complex agent environmentscited by 0
  12. 2025Actor-Free Continuous Control via Structurally Maximizable Q-Functionscited by 0
  13. 2026Training Language Models to Cooperate with Inference-Time Controllerscited by 0