DAIL: Beyond Task Ambiguity for Language-Conditioned Reinforcement Learning

arXiv 2510.19562 ↗First posted 22 Oct 2025cs.AIcited by 1

On the map: Deep reinforcement learning methods → RL training of language and agent models

Abstract (arXiv)

Comprehending natural language and following human instructions are critical capabilities for intelligent agents. However, the flexibility of linguistic instructions induces substantial ambiguity across language-conditioned tasks, severely degrading algorithmic performance. To address these limitations, we present a novel method named DAIL (Distributional Aligned Learning), featuring two key components: distributional policy and semantic alignment. Specifically, we provide theoretical results that the value distribution estimation mechanism enhances task differentiability. Meanwhile, the semantic alignment module captures the correspondence between trajectories and linguistic instructions. Extensive experimental results on both structured and visual observation benchmarks demonstrate that DAIL effectively resolves instruction ambiguities, achieving superior performance to baseline methods. Our implementation is available at https://github.com/RunpengXie/Distributional-Aligned-Learning.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in RL training of language and agent models, one of 15 directions in Deep reinforcement learning methods. Builds on 53 of its 64 references that are in the map. First posted on arXiv 22 Oct 2025.

Builds on

53 of 64 references.

References · 53

2017Attention Is All You Needcited by 60,361
2015Deep Residual Learning for Image Recognitioncited by 52,096
2014Adam: A Method for Stochastic Optimizationcited by 37,240
2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
2018Representation Learning with Contrastive Predictive Codingcited by 7,777
2013Efficient Estimation of Word Representations in Vector Spacecited by 4,375
2023RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Controlcited by 3,080
2017FiLM: Visual Reasoning with a General Conditioning Layercited by 2,333
2020Conservative Q-Learning for Offline Reinforcement Learningcited by 1,809
2021Offline Reinforcement Learning with Implicit Q-Learningcited by 1,293
2018Off-Policy Deep Reinforcement Learning without Explorationcited by 1,241
2017Hindsight Experience Replaycited by 1,136

38 more

2019ALFRED: A Benchmark for Interpreting Grounded Instructions for Everyday Taskscited by 783
2017A Distributional Perspective on Reinforcement Learningcited by 765
2022R3M: A Universal Visual Representation for Robot Manipulationcited by 694
2022BC-Z: Zero-Shot Task Generalization with Robotic Imitation Learningcited by 523
2023Minigrid & Miniworld: Modular & Customizable Reinforcement Learning Environments for Goal-Oriented Taskscited by 327
2018BabyAI: A Platform to Study the Sample Efficiency of Grounded Language Learningcited by 244
2022Pre-Trained Language Models for Interactive Decision-Makingcited by 221
2015Contextual Markov Decision Processescited by 220
2020Language Conditioned Imitation Learning over Unstructured Datacited by 209
2016Modular Multitask Reinforcement Learning with Policy Sketchescited by 203
2023Grounding Large Language Models in Interactive Environments with Online Reinforcement Learningcited by 195
2019Goal-conditioned Imitation Learningcited by 162
2019A Survey of Reinforcement Learning Informed by Natural Languagecited by 154
2023HIQL: Offline Goal-Conditioned RL with Latent States as Actionscited by 141
2022What Matters in Language Conditioned Robotic Imitation Learning over Unstructured Datacited by 138
2020Language-Conditioned Imitation Learning for Robot Manipulation Taskscited by 136
2021Goal-Conditioned Reinforcement Learning with Imagined Subgoalscited by 134
2021Learning Language-Conditioned Robot Behavior from Offline Data and Crowd-Sourced Annotationcited by 118
2019Using Natural Language for Reward Shaping in Reinforcement Learningcited by 114
2017Grounded Language Learning in a Simulated 3D Worldcited by 100
2023Drive as You Speak: Enabling Human-Like Interaction with Large Language Models in Autonomous Vehiclescited by 100
2022Rethinking Goal-conditioned Supervised Learning and Its Connection to Offline RLcited by 89
2017Mapping Instructions and Visual Observations to Actions with Reinforcement Learningcited by 81
2022Offline Q-Learning on Diverse Multi-Task Data Both Scales And Generalizescited by 73
2017Gated-Attention Architectures for Task-Oriented Language Groundingcited by 72
2019From Language to Goals: Inverse Reinforcement Learning for Vision-Based Instruction Followingcited by 67
2021Conservative Data Sharing for Multi-Task Offline Reinforcement Learningcited by 64
2020Generating Adjacency-Constrained Subgoals in Hierarchical Reinforcement Learningcited by 58
2020Generalized Hindsight for Reinforcement Learningcited by 51
2022Hierarchical Planning Through Goal-Conditioned Offline Reinforcement Learningcited by 50
2018Convergence and Concentration of Empirical Measures under Wasserstein Distance in Unbounded Functional Spacescited by 41
2018Guiding Policies with Language via Meta-Learningcited by 35
2023Goal Representations for Instruction Following: A Semi-Supervised Language Interface to Controlcited by 32
2020PixL2R: Guiding Reinforcement Learning Using Natural Language by Mapping Pixels to Rewardscited by 30
2017Beating Atari with Natural Language Guided Reinforcement Learningcited by 24
2022Meta-Reinforcement Learning via Language Instructionscited by 18
2019Conditional Driving from Natural Language Instructionscited by 17
2020Language-Conditioned Goal Generation: a New Approach to Language Grounding for RLcited by 12

50 most cited of 53 All 53 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

2026SpecRLBench: A Benchmark for Generalization in Specification-Guided Reinforcement Learningcited by 1

Similar, not cited

Close in wording, no citation either way