# What's the Move? Hybrid Imitation Learning via Salient Points

arXiv: https://arxiv.org/abs/2412.05426 · HTML: https://mlsanity.com/paper/2412.05426/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Placed in Robot manipulation imitation learning, one of 13 directions in Offline RL and imitation learning from data.
- Builds on 14 of its 37 references that are in the map.
- First posted on arXiv 6 Dec 2024.

## Placed in

- Offline RL and imitation learning from data / [Robot manipulation imitation learning](https://mlsanity.com/offlinerl/robot-manipulation-imitation-learning/)

## Builds on (most cited first, 14 of 14)

- [OpenVLA: An Open-Source Vision-Language-Action Model](https://mlsanity.com/paper/2406.09246/) (2024)
- [Diffusion Policy: Visuomotor Policy Learning via Action Diffusion](https://mlsanity.com/paper/2303.04137/) (2023)
- [Do As I Can, Not As I Say: Grounding Language in Robotic Affordances](https://mlsanity.com/paper/2204.01691/) (2022)
- [Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware](https://mlsanity.com/paper/2304.13705/) (2023)
- [Universal Manipulation Interface: In-The-Wild Robot Teaching Without In-The-Wild Robots](https://mlsanity.com/paper/2402.10329/) (2024)
- [Scaling Robot Learning with Semantically Imagined Experience](https://mlsanity.com/paper/2302.11550/) (2023)
- [Act3D: 3D Feature Field Transformers for Multi-Task Robotic Manipulation](https://mlsanity.com/paper/2306.17817/) (2023)
- [3D Diffusion Policy: Generalizable Visuomotor Policy Learning via Simple 3D Representations](https://mlsanity.com/paper/2403.03954/) (2024)
- [EquiBot: SIM(3)-Equivariant Diffusion Policy for Generalizable and Data Efficient Learning](https://mlsanity.com/paper/2407.01479/) (2024)
- [Equivariant Diffusion Policy](https://mlsanity.com/paper/2407.01812/) (2024)
- [Waypoint-Based Imitation Learning for Robotic Manipulation](https://mlsanity.com/paper/2307.14326/) (2023)
- [KITE: Keypoint-Conditioned Policies for Semantic Manipulation](https://mlsanity.com/paper/2306.16605/) (2023)
- [STAP: Sequencing Task-Agnostic Policies](https://mlsanity.com/paper/2210.12250/) (2022)
- [Leveraging Locality to Boost Sample Efficiency in Robotic Manipulation](https://mlsanity.com/paper/2406.10615/) (2024)

## Built on by (most cited first, 9 of 9)

- [3D CAVLA: Leveraging Depth and 3D Context to Generalize Vision Language Action Models for Unseen Tasks](https://mlsanity.com/paper/2505.05800/) (2025)
- [SAIL: Faster-than-Demonstration Execution of Imitation Learning Policies](https://mlsanity.com/paper/2506.11948/) (2025)
- [VIA: Visual Interface Agent for Robot Control](https://mlsanity.com/paper/2607.11119/) (2026)
- [LodeStar: Long-horizon Dexterity via Synthetic Data Augmentation from Human Demonstrations](https://mlsanity.com/paper/2508.17547/) (2025)
- [Enhancing Reusability of Learned Skills for Robot Manipulation via Gaze Information and Motion Bottlenecks](https://mlsanity.com/paper/2502.18121/) (2025)
- [Make Code as Policy Great Again: Frontier Agents Write, Call, and Evolve Robot Tools](https://mlsanity.com/paper/2609.39018/) (2026)
- [Hybrid Imitation Learning: Teleoperation Augmentation Primitives that Policies Learn to Trigger](https://mlsanity.com/paper/2512.04960/) (2025)
- [Triplet2Track: A Hierarchical System with Object-Centric Representations for Reliable Long-Horizon Manipulation](https://mlsanity.com/paper/2608.22800/) (2026)
- [Fast Plans, Faithful Actions: Closing the Planning-Execution Gap in Hierarchical Vision-Language-Action Models](https://mlsanity.com/paper/2609.30833/) (2026)

## Abstract (arXiv)

While imitation learning (IL) offers a promising framework for teaching robots various behaviors, learning complex tasks remains challenging. Existing IL policies struggle to generalize effectively across visual and spatial variations even for simple tasks. In this work, we introduce SPHINX: Salient Point-based Hybrid ImitatioN and eXecution, a flexible IL policy that leverages multimodal observations (point clouds and wrist images), along with a hybrid action space of low-frequency, sparse waypoints and high-frequency, dense end effector movements. Given 3D point cloud observations, SPHINX learns to infer task-relevant points within a point cloud, or salient points, which support spatial generalization by focusing on semantically meaningful features. These salient points serve as anchor points to predict waypoints for long-range movement, such as reaching target poses in free-space. Once near a salient point, SPHINX learns to switch to predicting dense end-effector movements given close-up wrist images for precise phases of a task. By exploiting the strengths of different input modalities and action representations for different manipulation phases, SPHINX tackles complex tasks in a sample-efficient, generalizable manner. Our method achieves 86.7% success across 4 real-world and 2 simulated tasks, outperforming the next best state-of-the-art IL baseline by 41.1% on average across 440 real world trials. SPHINX additionally generalizes to novel viewpoints, visual distractors, spatial arrangements, and execution speeds with a 1.7x speedup over the most competitive baseline. Our website (http://sphinx-manip.github.io) provides open-sourced code for data collection, training, and evaluation, along with supplementary videos.