Robotics · arXiv:2602.20119 v2 · submitted 23 Feb 2026
NovaPlan: Zero-Shot Long-Horizon Manipulation via Closed-Loop Video Language Planning
The paper says
“Across diverse multi-stage tasks, NovaPlan substantially outperforms prior zero-shot systems, achieving complex assembly and dexterous error recovery entirely without task-specific training or demonstrations.”
from the abstract · picked out as the main claim by a classifier
Reviews
Reviews by people
No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.
The paper
Abstract
Solving complex long-horizon robotic tasks requires joint reasoning over abstract task structure and low-level physical interaction. While combining Vision-Language Models (VLMs) and video generation models offers a promising path for zero-shot planning, their individual tendencies to hallucinate physics or violate geometric consistency often compound over time, preventing reliable real-world execution. We introduce NovaPlan, a hierarchical framework that enables robust, zero-shot long-horizon manipulation by systematically proposing, verifying, and repairing visual plans. At the high level, a VLM planner decomposes tasks and filters out dynamically inconsistent futures by verifying multiple candidate video rollouts. To translate these imagined futures into reliable physical actions, NovaPlan utilizes a hybrid geometric representation that adaptively switches between object-centric flow and human hand flow. Finally, NovaPlan closes the loop by continuously monitoring execution to verify outcomes and synthesize local, non-prehensile corrective behaviors, such as fingertip poking, when failures occur. Across diverse multi-stage tasks, NovaPlan substantially outperforms prior zero-shot systems, achieving complex assembly and dexterous error recovery entirely without task-specific training or demonstrations. Please visit our project website for additional results: https://nova-plan.github.io/
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. No sentence in this record was written by a model.
abstract compiled · 2026-10-08 · no language-model review
How the main claim is stated: asserted 1.00
What the abstract itself states, as probabilities: claims first 0.06 · states limitation 0.10 · human participants 0.07 · controlled comparison 0.53 · randomized assignment 0.01 · states quantitative result 0.05 · compares with prior results 0.94 · mentions code or data availability 0.17
The abstract, sentence by sentence
This table labels the abstract only (arXiv listing abstract v2). The number is the probability the classifier gave the label.
| Sentence | Role | The authors' own work | States the main conclusion | Strength |
|---|---|---|---|---|
| Solving complex long-horizon robotic tasks requires joint reasoning over abstract task structure and low-level physical interaction. | background 1.00 | 0.05 | 0.03 | asserted 1.00 |
| While combining Vision-Language Models (VLMs) and video generation models offers a promising path for zero-shot planning, their individual tendencies to hallucinate physics or violate geometric consistency often compound over time, preventing reliable real-world execution. | gap 0.52 | 0.06 | 0.03 | asserted 0.67 |
| We introduce NovaPlan, a hierarchical framework that enables robust, zero-shot long-horizon manipulation by systematically proposing, verifying, and repairing visual plans. | contribution 1.00 | 0.88 | 0.24 | asserted 1.00 |
| At the high level, a VLM planner decomposes tasks and filters out dynamically inconsistent futures by verifying multiple candidate video rollouts. | method 0.99 | 0.84 | 0.04 | asserted 1.00 |
| To translate these imagined futures into reliable physical actions, NovaPlan utilizes a hybrid geometric representation that adaptively switches between object-centric flow and human hand flow. | method 0.98 | 0.85 | 0.03 | asserted 1.00 |
| Finally, NovaPlan closes the loop by continuously monitoring execution to verify outcomes and synthesize local, non-prehensile corrective behaviors, such as fingertip poking, when failures occur. | method 0.98 | 0.88 | 0.06 | asserted 1.00 |
| Across diverse multi-stage tasks, NovaPlan substantially outperforms prior zero-shot systems, achieving complex assembly and dexterous error recovery entirely without task-specific training or demonstrations. | result 1.00 | 0.94 | 0.89 | asserted 1.00 |
| Please visit our project website for additional results: https://nova-plan.github.io/ | future work 0.93 | 0.04 | 0.02 | unclear 0.95 |
What kind of paper the classifier reports this is: method or tool 1.00 · evidence from observation or experiment 0.95
What was examined, and what was not
Sources read
- arXiv listing abstract v2
Not examined
- full text
- figures and tables
- references and citing sentences
- relations to other papers
Levels run
- P abstract
Not run
- S sections
- T body sentences
- C citations
- R cross-paper relations
How the labels were made
Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-10-08: 1 request, 6,273 input tokens.
Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. This record as JSON
Cited by 1 Pith paper
Reviewed papers on Pith that cite this one. The full citation record
How the literature uses it
6 arXiv papers cite this in 9 sentences
Papers citing it · 6 sentences without a confident label
Cites it as background 3 papers
-
From latent-imagination control (Dreamer V3, TD-MPC2) [ref] and text-conditioned video planning (UniPi, RoboDreamer, ManipDreamer, NovaPlan) [ref] to unified action-future models—coupled diffusion (UWM), interleaved autoregressive streams (WorldVLA, RynnVLA-002, UniVLA), JEPA-style and latent WAMs (VLA-JEPA, Being-H0.7), memory or chain-of-world reasoning (MemoryVLA, CoT-VLA, CoWVLA, AdaWorldPolicy, dreamzero, lingbotva), generation-action bridges (F1), foundation video-policy fine-tuning (GE-Act, Cosmos-Policy), and recent variants targeting test-time efficiency (Fast-WAM), condition-space world representation (World Guidance), compositional world models (RISE), and forward-inverse verification (WAV) [ref] —these models predict futures in pixel space, autoregressive token streams or global latents, providing extra temporal supervision but no decomposed, queryable view of the individual
OA-WAM: Object-Addressable World Action Model for Robust Robot Manipulation · Related Works · 0.95
-
A parallel line stacks a VLM planner above a VLA executor [ref] , sometimes paired with a critic for failure detection and replanning [ref] .
VoLo: A Physical Orchestrator for Open-Vocabulary Long-Horizon Manipulation · Related Work · 0.83
-
NovaPlan [ref] combines a closed-loop VLM planner with generated video references and geometrically grounded execution for zero-shot long-horizon manipulation.
WorldScape Policy 2.0: Empowering Steerable World Action Modeling with Reasoning-Augmented Memory and In-Context Learning · Long-Term Memory and Long-Horizon Planning · 0.84
Challenges and discussion
Know something about this paper? Leave a comment.
Come back by email
A short list of the week's newly reviewed papers. Nothing is sent until you confirm from your inbox, and every email has a link that stops it. No account.
No comments yet.
Sign in to comment