Conditional Neural Processes for Self-Supervised Goal Discovery in Action Imitation

Marco Gabriele Fedozzi; Yukie Nagai; Francesco Rea; Alessandra Sciutti

cs.ROOct 23, 2026

Conditional Neural Processes for Self-Supervised Goal Discovery in Action Imitation

Authors: Marco Gabriele Fedozzi, Yukie Nagai, Francesco Rea, Alessandra Sciutti

Organizations: DIBRIS Department, University of Genoa · CONTACT Unit, Italian Institute of Technology · International Research Center for Neurointelligence, The University of Tokyo

Abstract

Inspired by infants' ability to learn novel tasks by imitating others, in this work we evaluate to what extent Conditional Neural Processes can autonomously learn representations of actions that are informative about their underlying goal. We show that the addition of an attention mechanism in the hidden space fosters the emergence of robust goal representations early on in the observation of an action sequence, making it a promising model for a robotic self-supervised general learner.

Figures & tables

Fig. 1: AttnDMBN architecture. The same color convention as in [ 9 ] has been adopted. Networks with the same color share weights.

Panel of Figure 2 — Fig. 2: Absolute distance of the mean classification from the correct label as function of the number of observations, either form the start or the end of the sequence. Solid lines and shaded area represent the mean and standard deviation of the behavior across the test set. Dash-dotted lines show the output for the individual test sequences. The frame and joint lines are the result of heads appended to the respective backbone encoder, whilst the hidden head was placed after the CNP layer. 2a ) Autoencoder; 2b ) original DMBN; 2c ) DMBN with Positional Time Encodings; 2d ) DMBN with AttnCNP hidden layer.

Action Representation Learning

See all 16 papers

Consistent Zero-Shot Imitation with Contrastive Goal Inference
Oct 20, 2025Kathryn Wantlin, Chongyi Zheng, Benjamin EysenbachContrastive Learning Reinforcement Learning
Department of Computer Science, Princeton University, Princeton, NJ, USA.
SCAR: Self-Supervised Continuous Action Representation Learning
May 13, 2026Hongjia Liu, Fan Feng, Minghao Fu +3Latent Action Learning Representation Learning
University of California, San Diego · KTH Royal Institute of Technology
Exploring Temporal Representation in Neural Processes for Multimodal Action Prediction
Apr 9, 2026Marco Gabriele Fedozzi, Yukie Nagai, Francesco Rea +1Temporal Representation Learning Robotics
DIBRIS Department, University of Genoa, Via All’Opera Pia 13 16145 Genoa, Italy · CONTACT Unit, Italian Institute of Technology, Via Enrico Melen 83, 16152 Genoa, Italy · International Research Center for Neurointelligence, The University of Tokyo 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan

Oct 20, 2025J/K move · Enter open · S save

Consistent Zero-Shot Imitation with Contrastive Goal Inference

Kathryn Wantlin, Chongyi Zheng, Benjamin Eysenbach

Department of Computer Science, Princeton University, Princeton, NJ, USA.

Contrastive Learning Reinforcement Learning Imitation Learning Goal-Conditioned RL

Open paper AskPDF arXiv

Zero-shot imitation learning requires an agent to reproduce expert behavior from a single demonstration without additional environment interaction or gradient updates at test time. We introduce Contrastive Inverse Reinforcement Learning (CIRL), a self-supervised framework for pre-training zero-shot imitation agents. Our methods rests on a key observation that many useful tasks can be summarized by a single goal state. We can thus convert the multi-task inverse RL problem into a more tractable goal-inference problem, and utilize state-of-the-art goal-conditioned RL methods to recover a policy that reaches the goal. During pre-training, CIRL jointly employs three components to learn without any rewards or demonstrations: (1) a variant of contrastive RL designed to learn maximum-entropy goal-conditioned policies, (2) an automatic goal proposal mechanism (GoalKDE) that drives exploration, and (3) a mean-field variational model that performs amortized goal inference from trajectories. We prove that this procedure consistently recovers the demonstrator's intent by accounting for the relative difficulty of reaching different states and show how structurally similar prior work may otherwise fail to infer the correct reward. Experiments on goal-conditioned and standard reward-maximizing control tasks show that CIRL outperforms prior zero-shot imitation methods, supporting the expressiveness of goals as a compact summary of behavior.

Abstract

Action Representation Learning

Explore similar work

Consistent Zero-Shot Imitation with Contrastive Goal Inference

SCAR: Self-Supervised Continuous Action Representation Learning

Exploring Temporal Representation in Neural Processes for Multimodal Action Prediction