Learning Invariable Semantical Representation from Language for Extensible Policy Generalization

• URL: http://arxiv.org/abs/2202.00466v1
• Date: Wed, 26 Jan 2022 08:04:27 GMT
• Title: Learning Invariable Semantical Representation from Language for Extensible Policy Generalization
• Authors: Yihan Li, Jinsheng Ren, Tianrun Xu, Tianren Zhang, Haichuan Gao, and Feng Chen
• Abstract summary: We propose a method to learn semantically invariant representations called element randomization. We theoretically prove the feasibility of learning semantically invariant representations through randomization. Experiments on challenging long-horizon tasks show that our low-level policy reliably generalizes to tasks against environment changes.
• Score: 4.457682773596843
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recently, incorporating natural language instructions into reinforcement learning (RL) to learn semantically meaningful representations and foster generalization has caught many concerns. However, the semantical information in language instructions is usually entangled with task-specific state information, which hampers the learning of semantically invariant and reusable representations. In this paper, we propose a method to learn such representations called element randomization, which extracts task-relevant but environment-agnostic semantics from instructions using a set of environments with randomized elements, e.g., topological structures or textures, yet the same language instruction. We theoretically prove the feasibility of learning semantically invariant representations through randomization. In practice, we accordingly develop a hierarchy of policies, where a high-level policy is designed to modulate the behavior of a goal-conditioned low-level policy by proposing subgoals as semantically invariant representations. Experiments on challenging long-horizon tasks show that (1) our low-level policy reliably generalizes to tasks against environment changes; (2) our hierarchical policy exhibits extensible generalization in unseen new tasks that can be decomposed into several solvable sub-tasks; and (3) by storing and replaying language trajectories as succinct policy representations, the agent can complete tasks in a one-shot fashion, i.e., once one successful trajectory has been attained.

Related papers

• Learning with Language-Guided State Abstractions [58.199148890064826]
  Generalizable policy learning in high-dimensional observation spaces is facilitated by well-designed state representations. Our method, LGA, uses a combination of natural language supervision and background knowledge from language models to automatically build state representations tailored to unseen tasks. Experiments on simulated robotic tasks show that LGA yields state abstractions similar to those designed by humans, but in a fraction of the time.
  arXiv  Detail & Related papers  (2024-02-28T23:57:04Z)
• Invariant Causal Imitation Learning for Generalizable Policies [87.51882102248395]
  We propose Invariant Causal Learning (ICIL) to learn an imitation policy. ICIL learns a representation of causal features that is disentangled from the specific representations of noise variables. We show that ICIL is effective in learning imitation policies capable of generalizing to unseen environments.
  arXiv  Detail & Related papers  (2023-11-02T16:52:36Z)
• A Comprehensive Empirical Evaluation of Existing Word Embedding Approaches [5.065947993017158]
  We present the characteristics of existing word embedding approaches and analyze them with regard to many classification tasks. Traditional approaches mostly use matrix factorization to produce word representations, and they are not able to capture the semantic and syntactic regularities of the language very well. On the other hand, Neural-network-based approaches can capture sophisticated regularities of the language and preserve the word relationships in the generated word representations.
  arXiv  Detail & Related papers  (2023-03-13T15:34:19Z)
• Interpretable Reinforcement Learning with Multilevel Subgoal Discovery [77.34726150561087]
  We propose a novel Reinforcement Learning model for discrete environments. In the model, an agent learns information about environment in the form of probabilistic rules. No reward function is required for learning; an agent only needs to be given a primary goal to achieve.
  arXiv  Detail & Related papers  (2022-02-15T14:04:44Z)
• Skill Induction and Planning with Latent Language [94.55783888325165]
  We formulate a generative model of action sequences in which goals generate sequences of high-level subtask descriptions. We describe how to train this model using primarily unannotated demonstrations by parsing demonstrations into sequences of named high-level subtasks. In trained models, the space of natural language commands indexes a library of skills; agents can use these skills to plan by generating high-level instruction sequences tailored to novel goals.
  arXiv  Detail & Related papers  (2021-10-04T15:36:32Z)
• Neural Abstructions: Abstractions that Support Construction for Grounded Language Learning [69.1137074774244]
  Leveraging language interactions effectively requires addressing limitations in the two most common approaches to language grounding. We introduce the idea of neural abstructions: a set of constraints on the inference procedure of a label-conditioned generative model. We show that with this method a user population is able to build a semantic modification for an open-ended house task in Minecraft.
  arXiv  Detail & Related papers  (2021-07-20T07:01:15Z)
• Infusing Finetuning with Semantic Dependencies [62.37697048781823]
  We show that, unlike syntax, semantics is not brought to the surface by today's pretrained models. We then use convolutional graph encoders to explicitly incorporate semantic parses into task-specific finetuning.
  arXiv  Detail & Related papers  (2020-12-10T01:27:24Z)
• Distilling a Hierarchical Policy for Planning and Control via Representation and Reinforcement Learning [18.415568038071306]
  We present a hierarchical planning and control framework that enables an agent to perform various tasks and adapt to a new task flexibly. Rather than learning an individual policy for each task, the proposed framework, DISH, distills a hierarchical policy from a set of tasks by representation and reinforcement learning.
  arXiv  Detail & Related papers  (2020-11-16T23:58:49Z)
• Ask Your Humans: Using Human Instructions to Improve Generalization in Reinforcement Learning [32.82030512053361]
  We propose the use of step-by-step human demonstrations in the form of natural language instructions and action trajectories. We find that human demonstrations help solve the most complex tasks. We also find that incorporating natural language allows the model to generalize to unseen tasks in a zero-shot setting.
  arXiv  Detail & Related papers  (2020-11-01T14:39:46Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.