Vision-language-action models are highly vulnerable to membership inference attacks, including practical black-box versions that exploit generated actions and motion trajectories.
super hub Canonical reference
RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control
Canonical reference. 77% of citing Pith papers cite this work as background.
213
Pith papers citing it
Background
77% of classified citations
abstract
We study how vision-language models trained on Internet-scale data can be incorporated directly into end-to-end robotic control to boost generalization and enable emergent semantic reasoning. Our goal is to enable a single end-to-end trained model to both learn to map robot observations to actions and enjoy the benefits of large-scale pretraining on language and vision-language data from the web. To this end, we propose to co-fine-tune state-of-the-art vision-language models on both robotic trajectory data and Internet-scale vision-language tasks, such as visual question answering. In contrast to other approaches, we propose a simple, general recipe to achieve this goal: in order to fit both natural language responses and robotic actions into the same format, we express the actions as text tokens and incorporate them directly into the training set of the model in the same way as natural language tokens. We refer to such category of models as vision-language-action models (VLA) and instantiate an example of such a model, which we call RT-2. Our extensive evaluation (6k evaluation trials) shows that our approach leads to performant robotic policies and enables RT-2 to obtain a range of emergent capabilities from Internet-scale training. This includes significantly improved generalization to novel objects, the ability to interpret commands not present in the robot training data (such as placing an object onto a particular number or icon), and the ability to perform rudimentary reasoning in response to user commands (such as picking up the smallest or largest object, or the one closest to another object). We further show that incorporating chain of thought reasoning allows RT-2 to perform multi-stage semantic reasoning, for example figuring out which object to pick up for use as an improvised hammer (a rock), or which type of drink is best suited for someone who is tired (an energy drink).
hub tools
citation-role summary
background 47
method 6
baseline 4
other 2
extension 1
citation-polarity summary
claims ledger
- abstract We study how vision-language models trained on Internet-scale data can be incorporated directly into end-to-end robotic control to boost generalization and enable emergent semantic reasoning. Our goal is to enable a single end-to-end trained model to both learn to map robot observations to actions and enjoy the benefits of large-scale pretraining on language and vision-language data from the web. To this end, we propose to co-fine-tune state-of-the-art vision-language models on both robotic trajectory data and Internet-scale vision-language tasks, such as visual question answering. In contrast
authors
co-cited works
representative citing papers
OSWorld provides the first unified real-computer benchmark for open-ended multimodal agent tasks, exposing large performance gaps between humans and state-of-the-art LLM/VLM agents.
Labimus is the first benchmark for humanoid dexterous manipulation in organic chemistry laboratories, exposing a gap between task completion and required experimental precision.
LIBERO-Safety supplies a scalable benchmark, data-generation pipeline, and 19,664-demonstration dataset that exposes a generalization-safety tension in current VLA models where diverse training improves collision avoidance but task success stays limited by trajectory quality and semantic understandi
RoboGaze presents a structured multi-agent VLM pipeline and robotics-specific error taxonomy that improves video evaluation metrics by up to 43 F1 points over zero-shot baselines on a 382-clip dataset.
World models introduce a stealthy poisoning vector into robot learning pipelines where malicious prompts or dynamics in teleoperated data activate only during synthetic trajectory generation, enabling backdoors in downstream policies.
BOKBO is the first conformal abstention method for K-sample VLA policies that supplies finite-sample distribution-free guarantees on executed violation rates, with global and Mondrian per-task variants.
PhAIL provides an open benchmark and distributional evaluation method for real-robot VLA policies using time-to-success CDF, HRT scoring, and KS significance tests.
VLA architectures exhibit architecture-specific failure signatures at the motor-command level, with direction reversal as a universal predictor and velocity monitoring ineffective for continuous models.
Introduces Colosseum V2 benchmark for evaluating VLA model generalization in robotic manipulation with 28 tasks, revealing limitations in current methods and sim-real correlations.
Introduces MM-CreativityBench for affordance-grounded creative tool use and shows that DPO-based alignment with an affordance knowledge base improves entity and part selection while cutting hallucination errors in LMMs.
Visual CoT agents exhibit tool-use collapse where tool usage declines but task accuracy rises, and adding entropy regularization for rollout diversity produces the strongest performance.
CrossVLA develops a surrogate log-probability estimator for DPO on flow-matching VLAs, shows DoRA outperforming LoRA by +10.4 pp mean on LIBERO, and identifies inference bottlenecks with limited caching gains.
Demo-JEPA enables one-shot cross-embodiment imitation by mapping visual demonstrations to shared latent future trajectories that serve as subgoals for the target agent's own forward dynamics planning.
DSSP is a history-conditioned diffusion state space policy that uses SSMs to encode full observation streams with an auxiliary dynamics objective and hierarchical fusion, achieving SOTA results with reduced model size in robot manipulation.
A new VLA model called SI uses a four-step chain-of-thought to derive driving intent and applies it via classifier-free guidance to a flow-matching trajectory generator, showing competitive Waymo scores and intent-controllable plans.
A liveness-based Bellman operator enables conservative offline policy evaluation for manipulation tasks by encoding task progression and reducing truncation bias from finite horizons.
Pace-and-Path Correction decomposes a quadratic cost minimization into orthogonal pace and path channels to correct chunked actions in VLA models, raising success rates by up to 28.8% in dynamic settings.
SABER provides 44.8K multi-representation action samples from unscripted retail environments that raise a VLA model's mean success rate on ten manipulation tasks from 13.4% to 29.3%.
OA-WAM uses persistent address vectors and dynamic content vectors in object slots to enable addressable world-action prediction, improving robustness on manipulation benchmarks under scene changes.
A cross-version swap protocol reveals dominant skills that swing composition success by up to 50 percentage points, and an atomic probe with selective revalidation governs updates at lower cost than always re-testing full compositions.
CF-VLA uses a coarse initialization over endpoint velocity followed by single-step refinement to achieve strong performance with low inference steps on CALVIN, LIBERO, and real-robot tasks.
Mask World Model predicts semantic mask dynamics with video diffusion and integrates it with a diffusion policy head, outperforming RGB world models on LIBERO and RLBench while showing better real-world generalization and texture robustness.
AeroBridge-TTA achieves +22 pt average gains on out-of-distribution UAV dynamics mismatches by updating a latent state online from observed transitions in a language-conditioned policy.
citing papers explorer
-
Diversity Over Frequency: Rethinking Tool Use in Visual Chain-of-Thought Agents
Visual CoT agents exhibit tool-use collapse where tool usage declines but task accuracy rises, and adding entropy regularization for rollout diversity produces the strongest performance.
-
CrossVLA: Cross-Paradigm Post-Training and Inference Optimization for Vision-Language-Action Models
CrossVLA develops a surrogate log-probability estimator for DPO on flow-matching VLAs, shows DoRA outperforming LoRA by +10.4 pp mean on LIBERO, and identifies inference bottlenecks with limited caching gains.
-
CF-VLA: Efficient Coarse-to-Fine Action Generation for Vision-Language-Action Policies
CF-VLA uses a coarse initialization over endpoint velocity followed by single-step refinement to achieve strong performance with low inference steps on CALVIN, LIBERO, and real-robot tasks.
-
AlphaDrive: Unleashing the Power of VLMs in Autonomous Driving via Reinforcement Learning and Reasoning
AlphaDrive uses GRPO-based RL rewards and two-stage SFT+RL training on VLMs to improve autonomous driving planning performance and efficiency while producing emergent multimodal capabilities.
-
3D-VLA: A 3D Vision-Language-Action Generative World Model
3D-VLA is a new embodied foundation model that uses a 3D LLM plus aligned diffusion models to generate future images and point clouds for improved reasoning and action planning in 3D environments.
-
S-Agent: Spatial Tool-Use Elicits Reasoning for Spatial Intelligence
S-Agent augments VLMs with spatial tools, scene and agent memory for evidence accumulation on multi-view and video tasks, and produces an 8B model via SFT on its own trajectories that beats same-scale baselines.
-
LARA: Latent Action Representation Alignment for Vision-Language-Action Models
LARA jointly optimizes LAM and VLA models via representation alignment to improve robotic manipulation performance using human videos.
-
VLA-Hijack: A Transferable Patch Attack against Vision-Language-Action Models via Visual Proprioception Hijacking
VLA-Hijack is a new adversarial patch attack on Vision-Language-Action models that suppresses real arm features and injects the patch as surrogate embodiment to achieve high cross-architecture transferability.
-
Offline Semantic Guidance for Efficient Vision-Language-Action Policy Distillation
VLA-AD distills 7B VLA teachers into 158M students using offline VLM semantic guidance on task phases and directions, matching teacher performance on LIBERO with 44x size reduction and 3.28x speedup.
-
UAM: A Dual-Stream Perspective on Forgetting in VLA Training
UAM adds a Dorsal Expert initialized from a generative model and trained on visual dynamics prediction to preserve over 95% of VLM multimodal ability in VLA training while achieving top success rates on manipulation tasks including OOD cases.
-
SceneParser: Hierarchical Scene Parsing for Visual Semantics Understanding
SceneParser introduces hierarchical scene parsing as object-part-affordance chains, a VLM trained with pseudo labels and curriculum learning, and SceneParser-Bench with 1.74M affordance annotations, showing better structure-aware results than existing MLLMs.
-
DenseStep2M: A Scalable, Training-Free Pipeline for Dense Instructional Video Annotation
A scalable training-free pipeline using video segmentation, filtering, and off-the-shelf multimodal models creates DenseStep2M, a dataset of 100K videos and 2M detailed instructional steps that improves dense captioning, step grounding, and cross-modal retrieval.
-
SimScale: Learning to Drive via Real-World Simulation at Scale
SimScale synthesizes unseen driving states from real logs via neural rendering and reactive environments, generates pseudo-expert trajectories, and shows that co-training on real plus simulated data improves planning robustness and generalization on real benchmarks, with gains scaling by simulation
-
HybridVLA: Collaborative Diffusion and Autoregression in a Unified Vision-Language-Action Model
HybridVLA unifies diffusion and autoregression in a single VLA model via collaborative training and ensemble to raise robot manipulation success rates by 14% in simulation and 19% in real-world tasks.
-
Video Prediction Policy: A Generalist Robot Policy with Predictive Visual Representations
Video Prediction Policy conditions robot action learning on future-frame predictions inside fine-tuned video diffusion models, yielding 18.6% relative gains on Calvin ABC-D and 31.6% higher real-world success rates.
-
LongVILA: Scaling Long-Context Visual Language Models for Long Videos
LongVILA scales visual-language models from 8 to 2048 video frames with 99.8% needle-in-a-haystack accuracy using long-context extension, supervised fine-tuning, and multi-modal sequence parallelism on up to 256 GPUs.
-
NaVid: Video-based VLM Plans the Next Step for Vision-and-Language Navigation
NaVid, a video-based VLM trained on 510k navigation and 763k web samples, achieves SOTA VLN performance using only monocular RGB video for next-step action planning in sim and real environments.
-
DriveVLM: The Convergence of Autonomous Driving and Large Vision-Language Models
DriveVLM adds vision-language models with scene description, analysis, and hierarchical planning modules to autonomous driving, paired with a hybrid DriveVLM-Dual system tested on nuScenes and SUP-AD datasets and deployed on a production vehicle.
-
GPT-Driver: Learning to Drive with GPT
GPT-3.5 is turned into an autonomous-vehicle motion planner by representing driving scenes and trajectories as language tokens and applying a prompting-reasoning-finetuning pipeline, with results shown on nuScenes.
-
SAFE-Pruner: Semantic Attention-Guided Future-Aware Token Pruning for Efficient Vision-Language-Action Manipulation
SAFE-Pruner forecasts deep-layer token saliency in VLA models via semantic attention consistency and adaptive subtask detection to achieve up to 1.89x speedup with under 1.7% success rate loss.
-
Rethinking VLM Representation for VLA Initialization
Experiments indicate original VLM representations are crucial for VLA performance, LoRA outperforms full finetuning, and staged robot-data pretraining yields the strongest initialization.
-
HiVLA: A Visual-Grounded-Centric Hierarchical Embodied Manipulation System
HiVLA decouples VLM-based semantic planning with visual grounding from a cascaded cross-attention DiT action expert, outperforming end-to-end VLAs on long-horizon and fine-grained manipulation.
-
ThinkAct: Vision-Language-Action Reasoning via Reinforced Visual Latent Planning
ThinkAct introduces reinforced visual latent planning in a dual VLA system to enable better long-horizon reasoning and adaptation for embodied tasks.
-
Growing a Multi-head Twig via Distillation and Reinforcement Learning to Accelerate Large Vision-Language Models
TwigVLM adds a twig module to VLMs for twig-guided token pruning and self-speculative decoding, retaining 96% performance after pruning 88.9% visual tokens and delivering 154% speedup on long responses for LLaVA-1.5-7B.
-
AppAgent: Multimodal Agents as Smartphone Users
AppAgent lets large language models operate diverse smartphone apps via visual interactions and learns app usage from exploration or demonstrations.
-
Mags-RL: Wearing Multimodal LLMs a Magnifying Glass via Agentic Reinforcement Learning For Complex Scene Reasoning
Mags-RL uses agentic RL and a super-resolution agent for two-round reasoning in MLLMs, claiming gains on VSR, TallyQA, and GQA with a curriculum needing only 40 samples.
-
General Covariant Action Modeling: Constructing Generalized Manifolds via Spatio-Temporal Decoupling
GAM framework uses arc-length parameterization for temporal invariance and schema-affine factorization for geometric invariance to build a covariant action manifold integrated into VLA models for improved generalization from sparse data.