Foresight uses iterative VLM plan proposal and critique with RL from human feedback to raise navigation success 37% and cut interventions 52% in real-world tests.
hub Canonical reference
Grape: Generalizing robot policy via preference alignment
Canonical reference. 86% of citing Pith papers cite this work as background.
28
Pith papers citing it
Background
86% of classified citations
hub tools
citation-role summary
background 6
baseline 1
citation-polarity summary
representative citing papers
MoLA infers a mixture of latent actions from generated future videos via modality-aware inverse dynamics models to improve robot manipulation policies.
Freeform Preference Learning trains language-conditioned multi-axis reward models from human pairwise preferences to produce steerable and compositional robot policies that outperform sparse and binary-preference baselines by 38 percentage points.
FAT decomposes structured prediction into specialist hypothesis generation and foundation-model proxy reasoning, yielding consistent gains over baselines on detection, trajectory, and segmentation tasks.
T^2VLA is a test-time reinforcement learning framework for VLAs that uses internal confidence to define intrinsic rewards via similarity to high-confidence expert demonstrations and a dual-expert bootstrapping mechanism.
Success Visitation Matching uses a discriminator to turn sparse outcome rewards into dense process rewards by matching visitations of successful episodes, provably preserving the optimal policy and speeding up robotic RL finetuning.
SafeDojo is a new world model-based safe RL framework for VLA that outperforms baselines on SafeLIBERO and real robot tasks.
UniIntervene uses future-conditioned action-value estimation and a temporal value-risk critic to trigger memory-based recovery interventions, reporting 8.6% higher success rates and 57% fewer human interventions than prior HiL-RL methods on real manipulation tasks.
FlowPRO applies proximalized preference optimization to flow-matching VLAs with intervention-rollback data to reach higher success rates on long-horizon bimanual tasks without rewards or critics.
RePO-VLA raises average adversarial success rates in VLA manipulation from 20% to 75% by using recovery-aware initialization, a progress-aware semantic value function, and value-conditioned refinement on success and corrective trajectories.
LWD is a fleet-scale offline-to-online RL framework that continually improves pretrained VLA policies using autonomous rollouts and human interventions, reaching 95% average success on real-world manipulation tasks.
LaST-R1 introduces a RL post-training method called LAPO that optimizes latent Chain-of-Thought reasoning in vision-language-action models, yielding 99.9% success on LIBERO and up to 22.5% real-world gains.
TwinRL expands RL exploration via digital twin reconstruction and twin RL warm-up to guide real-world learning, reaching near-100% success with 20 minutes of on-robot time across four tasks.
RECAP enables a generalist VLA to self-improve via advantage-conditioned RL on mixed real-world data, more than doubling throughput and halving failure rates on hard manipulation tasks.
DeepThinkVLA shows CoT improves VLA models only under decoding and causal alignment, delivering 97% success on LIBERO and 21.7-point gains via hybrid attention and SFT-RL training.
SimpleVLA-RL applies tailored reinforcement learning to VLA models, reaching SoTA on LIBERO, outperforming π₀ on RoboTwin, and surpassing SFT in real-world tasks while reducing data needs and identifying a 'pushcut' phenomenon.
VLA-RL applies online RL to pretrained VLAs, yielding a 4.5% gain over strong baselines on 40 LIBERO manipulation tasks and matching commercial models like π₀-FAST.
UniVLA trains cross-embodiment vision-language-action policies from unlabeled videos via a latent action model in DINO space, beating OpenVLA on benchmarks with 1/20th pretraining compute and 1/10th downstream data.
AgiBot World supplies over 1 million trajectories enabling GO-1 to deliver 30% average gains over Open X-Embodiment and over 60% success on complex dexterous tasks while open-sourcing everything.
DexVLA combines a scaled diffusion action expert with embodiment curriculum learning to achieve better generalization and performance than prior VLA models on diverse robot hardware and long-horizon tasks.
PAPO-VLA identifies planning actions via variation and outcome, estimates their causal importance, and folds that importance into GRPO to emphasize key decisions while still using full-trajectory feedback.
DyGRO-VLA is a two-stage optimization framework for cross-task scaling of Vision-Language-Action models via dynamic grouped residual optimization in RL.
ProcVLM learns procedure-grounded dense progress rewards for robotic manipulation via a reasoning-before-estimation VLM trained on a 60M-frame synthesized corpus from 30 embodied datasets.
RESample uses exploratory sampling guided by a lightweight Coverage Function to expand VLA training data coverage, yielding 12% performance gains on LIBERO and real-world tasks with 10-20% added samples.
citing papers explorer
-
Foresight: Iterative Reasoning About Clues that Matter for Navigation
Foresight uses iterative VLM plan proposal and critique with RL from human feedback to raise navigation success 37% and cut interventions 52% in real-world tests.
-
From Imagined Futures to Executable Actions: Mixture of Latent Actions for Robot Manipulation
MoLA infers a mixture of latent actions from generated future videos via modality-aware inverse dynamics models to improve robot manipulation policies.
-
Freeform Preference Learning for Robotic Manipulation
Freeform Preference Learning trains language-conditioned multi-axis reward models from human pairwise preferences to produce steerable and compositional robot policies that outperform sparse and binary-preference baselines by 38 percentage points.
-
Rethinking Foundation Model Collaboration: Enhancing Specialized Models through Proxy Task Reasoning
FAT decomposes structured prediction into specialist hypothesis generation and foundation-model proxy reasoning, yielding consistent gains over baselines on detection, trajectory, and segmentation tasks.
-
Trust Your Instincts: Confidence-Driven Test-Time RL for Vision-Language-Action Models
T^2VLA is a test-time reinforcement learning framework for VLAs that uses internal confidence to define intrinsic rewards via similarity to high-confidence expert demonstrations and a dual-expert bootstrapping mechanism.
-
Learning Process Rewards via Success Visitation Matching for Efficient RL
Success Visitation Matching uses a discriminator to turn sparse outcome rewards into dense process rewards by matching visitations of successful episodes, provably preserving the optimal policy and speeding up robotic RL finetuning.
-
SafeDojo: Safe Reinforcement Learning for VLA via Interactive World Model
SafeDojo is a new world model-based safe RL framework for VLA that outperforms baselines on SafeLIBERO and real robot tasks.
-
UniIntervene: Agentic Intervention for Efficient Real-World Reinforcement Learning
UniIntervene uses future-conditioned action-value estimation and a temporal value-risk critic to trigger memory-based recovery interventions, reporting 8.6% higher success rates and 57% fewer human interventions than prior HiL-RL methods on real manipulation tasks.
-
FlowPRO: Reward-Free Reinforced Fine-Tuning of Flow-Matching VLAs via Proximalized Preference Optimization
FlowPRO applies proximalized preference optimization to flow-matching VLAs with intervention-rollback data to reach higher success rates on long-horizon bimanual tasks without rewards or critics.
-
RePO-VLA: Recovery-Driven Policy Optimization for Vision-Language-Action Models
RePO-VLA raises average adversarial success rates in VLA manipulation from 20% to 75% by using recovery-aware initialization, a progress-aware semantic value function, and value-conditioned refinement on success and corrective trajectories.
-
Learning While Deploying: Fleet-Scale Reinforcement Learning for Generalist Robot Policies
LWD is a fleet-scale offline-to-online RL framework that continually improves pretrained VLA policies using autonomous rollouts and human interventions, reaching 95% average success on real-world manipulation tasks.
-
LaST-R1: Reinforcing Robotic Manipulation via Adaptive Physical Latent Reasoning
LaST-R1 introduces a RL post-training method called LAPO that optimizes latent Chain-of-Thought reasoning in vision-language-action models, yielding 99.9% success on LIBERO and up to 22.5% real-world gains.
-
TwinRL: Digital Twin-Driven Reinforcement Learning for Real-World Robotic Manipulation
TwinRL expands RL exploration via digital twin reconstruction and twin RL warm-up to guide real-world learning, reaching near-100% success with 20 minutes of on-robot time across four tasks.
-
$\pi^{*}_{0.6}$: a VLA That Learns From Experience
RECAP enables a generalist VLA to self-improve via advantage-conditioned RL on mixed real-world data, more than doubling throughput and halving failure rates on hard manipulation tasks.
-
DeepThinkVLA: Enhancing Reasoning Capability of Vision-Language-Action Models
DeepThinkVLA shows CoT improves VLA models only under decoding and causal alignment, delivering 97% success on LIBERO and 21.7-point gains via hybrid attention and SFT-RL training.
-
UniVLA: Learning to Act Anywhere with Task-centric Latent Actions
UniVLA trains cross-embodiment vision-language-action policies from unlabeled videos via a latent action model in DINO space, beating OpenVLA on benchmarks with 1/20th pretraining compute and 1/10th downstream data.
-
AgiBot World Colosseo: A Large-scale Manipulation Platform for Scalable and Intelligent Embodied Systems
AgiBot World supplies over 1 million trajectories enabling GO-1 to deliver 30% average gains over Open X-Embodiment and over 60% success on complex dexterous tasks while open-sourcing everything.
-
DexVLA: Vision-Language Model with Plug-In Diffusion Expert for General Robot Control
DexVLA combines a scaled diffusion action expert with embodiment curriculum learning to achieve better generalization and performance than prior VLA models on diverse robot hardware and long-horizon tasks.
-
PAPO-VLA: Planning-Aware Policy Optimization for Vision-Language-Action Models
PAPO-VLA identifies planning actions via variation and outcome, estimates their causal importance, and folds that importance into GRPO to emphasize key decisions while still using full-trajectory feedback.
-
DyGRO-VLA: Cross-Task Scaling of Vision-Language-Action Models via Dynamic Grouped Residual Optimization
DyGRO-VLA is a two-stage optimization framework for cross-task scaling of Vision-Language-Action models via dynamic grouped residual optimization in RL.
-
ProcVLM: Learning Procedure-Grounded Progress Rewards for Robotic Manipulation
ProcVLM learns procedure-grounded dense progress rewards for robotic manipulation via a reasoning-before-estimation VLM trained on a 60M-frame synthesized corpus from 30 embodied datasets.
-
RESample: A Robust Data Augmentation Framework via Exploratory Sampling for Robotic Manipulation
RESample uses exploratory sampling guided by a lightweight Coverage Function to expand VLA training data coverage, yielding 12% performance gains on LIBERO and real-world tasks with 10-20% added samples.
-
Reflection-Based Task Adaptation for Self-Improving VLA
Reflective Self-Adaptation combines failure-reflective reinforcement learning with success-guided imitation learning to enable faster and more reliable task adaptation for pre-trained Vision-Language-Action models.
-
Large VLM-based Vision-Language-Action Models for Robotic Manipulation: A Survey
This survey organizes large VLM-based VLA models for robotic manipulation into monolithic and hierarchical paradigms, reviews their integrations and datasets, and outlines future directions.
-
SafeVLA: Towards Safety Alignment of Vision-Language-Action Model via Constrained Learning
SafeVLA applies constrained reinforcement learning via CMDP min-max optimization to VLAs, cutting safety violation costs by 83.58% while preserving task success on long-horizon mobile manipulation tasks.
-
Position: Good Embodied Reward Models Need Bad Behavior Data
Embodied reward models systematically over-reward unsafe, suboptimal, and shortcut robot behaviors due to training on successful data only, and modest inclusion of bad behavior data improves alignment with human preferences.