Action
-
MAGAR: Multimodal Agents for GUI-Based 3D Mechanical Assembly and Minimal Rework
2026-11-10 09:01:56Jingxuan Wei, Jiahao Fu, Fuqiang Wang, Xiaoming Wu, Yuwen Liu, Kaiwen Wei, Ming Yang, Junnan ZhuAbstract
A GUI command can be correct at the interface level yet still produce an incorrect assembly. An agent must identify the relevant components, respect their structural relationships, and determine which parts should move during recovery. We introduce MAGAR-Bench, a benchmark of 4,000 multimodal samples from 26 assembly families that separates Assembly Decision, Atomic Action, and Minimal Rework. For rework, the objective counts distinct moved parts within the validated recovery candidate set for each state, rather than the number of actions. We also introduce MAGAR, a selective correction framework built on a frozen policy obtained by supervised fine-tuning. Its key design matches correction granularity to output dependencies. Structured candidates revise assembly fields and rework scope, while complete action candidates keep native GUI primitives coupled with their parameters. A learned selector retains the original prediction unless a candidate passes a task-specific intervention threshold. On recorded states from held-out assembly families, MAGAR improves Overall, the task-balanced aggregate, from 37.50 to 41.39, with gains on all three tasks. Across 209 interventions, 70 correct an error and 22 introduce one. These results support task-specific correction while highlighting the remaining tradeoff between fixing errors and preserving reliable predictions.
Abstract (translated)
URL
https://arxiv.org/abs/2611.02472
PDF
-
Bridging the Evidence-to-Execution Gap:A Reflective Agent for Multi-Objective Peptide Design
2026-11-10 08:57:01Haosen Zhang, Yang YangAbstract
Large language models can reason over scientific literature, but textual plans do not reliably produce biological sequences with the intended properties. Sequence generators capture sequence patterns, yet they do not interpret evidence, choose interventions, or reflect on outcomes. This separation creates an evidence-to-execution gap between scientific reasoning and sequence-level control. We introduce EASER, a reflective agent that connects the two through a learned property interface. The interface consists of property-specific low-rank directions trained offline and fixed during search; an LLM controller (the controller) composes them to steer a frozen diffusion generator. From retrieved evidence, sequence context, and prior outcomes, it proposes intervention hypotheses that specify anchors, editable positions, and coefficients. Probe-and-Steer tests these hypotheses and allocates further samples from their joint predicted responses; reflection updates the agent's memory. On computational antimicrobial peptide design, EASER balances activity, non-hemolysis, and non-toxicity. Explicit hypothesis formulation yields larger multi-objective gains than direct action generation, and ablations support evidence retrieval, episodic history, reflection, and Probe-and-Steer. EASER achieves the highest mean hypervolume and lowest mean IGD+ among the evaluated search policies and generators on screened candidates. These results show that a learned property interface can make literature-grounded reasoning executable at the sequence level.
Abstract (translated)
URL
https://arxiv.org/abs/2610.06190
PDF
-
CIALLO: Learning Shared Intent Representations for Cross-Embodiment Manipulation
2026-11-10 08:35:27Choo Sin Wai, Dake Zhong, Jingye Zhang, Zhifang Liu, Siyao Xiao, Zirong Ding, Yuhong Zhang, Kangjie Chen, Dongbin Zhang, Guorui Song, Guocun Wang, Peng Jiao, Haoqian WangAbstract
Robot demonstrations contain reusable manipulation knowledge, but transferring this knowledge across embodiments requires representations that bridge differences in control interfaces. Action reconstruction alone does not ensure that similar behaviors across robots share a representation. To address this gap, we introduce CIALLO, an action codec that aligns actions, visual changes, and observation-grounded language in a shared space of short-horizon behavioral intent. An embodiment-conditioned decoder maps this intent back to robot-specific actions, allowing a shared representation to support different control interfaces. After pretraining on eight embodiments, CIALLO adapts to a new robot from one recording of at most 10 seconds by updating only its embodiment embeddings. The calibrated codec then remains frozen while a policy learns to predict intent from observations and instructions. We instantiate this policy with Qwen3.5-4B to form Qwen-CIALLO, whose backbone requires no large-scale robot-action pretraining before policy finetuning. With 20 demonstrations per task, Qwen-CIALLO achieves 76.7\% average success across real SO-101 and single-arm Piper robots, compared with 70.0\% for fully finetuned $\pi_0$. In simulation, we additionally finetune the action decoder, achieving 97.3\% success on LIBERO and 52.7\% on RoboTwin 2.0. The LIBERO policy also achieves 80.8\% on LIBERO-Plus without further finetuning. Together, these results support a shared action codec as a transferable control interface that separates embodiment calibration from task learning.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13732
PDF
-
Broad-Persistent Advising in Interactive Reinforcement Learning via Latent State Encoding
2026-11-10 08:06:06Zhuoran Liang, Francisco CruzAbstract
Deep Interactive Reinforcement Learning (Deep IntRL) accelerates reinforcement learning training by incorporating external advice. However, conventional Deep IntRL uses each piece of advice only in the current step and discards it after a single interaction. Persistent Advising and Broad-Persistent Advising (BPA) extend this process by storing advice in persistent memory and reusing it in similar states through state generalisation based on K-means clustering. This research proposes Encoder BPA, which employs a latent encoder to retrieve and reuse advice. In this framework, a trained encoder is used to map a state sequence into a latent representation at every time step. Advice can be retrieved through latent comparison, together with a top-(k) aggregation and a consistency condition to improve reuse stability. Experimental results show that Encoder BPA improves learning efficiency and advice utilisation at an early stage of training compared with different Deep IntRL approaches, especially in complex and partially observable environments.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13712
PDF
-
SeqFold: Timestamp-Addressable Compressed States for User Action Sequence Backbones
2026-11-10 07:55:16Mianwei Zhou, Duo Zhang, Henry Tsai, Yuqi Zheng, Jian Li, Juexin Lin, Sujay Khandagale, Vijay Rajaram, Dong Jiao, Tao ChengAbstract
Existing sequential user modeling methods mainly focus on improving the backbone encoder or learning task-specific user representations for a single prediction objective. We instead study how to export the multi-layer user states of a frozen long-history backbone as compact, timestamp-addressable representations for downstream training backfill and serving. In this setting, the user state at a given timestamp summarizes the user's actions before that time, so downstream systems need only the state aligned to each target timestamp rather than the full hidden-state timeline. To this end, we present \texttt{SeqFold}, a compressed-state interface that combines learned multi-channel fusion, channel-specific temporal decay, and Lookup-Free Quantization (LFQ) to produce compact discrete codes from frozen backbone states. Task-specific decoders then map these shared codes into flexible embedding spaces for downstream models. To make timestamp-addressable export practical at scale, we further develop a fused streaming implementation that updates the fusion recurrence after every action while materializing states only at target timestamps. On large-scale proprietary datasets, SeqFold reduces per-state representation storage by $3{,}400\times$ while preserving most of the downstream ranking performance of the uncompressed model, showing that SeqFold is an effective interface for reusing long-history sequence backbones in downstream models.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13704
PDF
-
On Private Knowledge Retrieval and Private Inference for secure multi-agent systems
2026-11-10 07:43:02Vicen\c{c} Torra, Maria Bras-Amor\'osAbstract
Multi-agent systems provide a convenient framework for the massive deployment of agents based on Large Language Models (LLMs). It provides a sound framework where agents can interact with existing software to perform actions in an autonomous way. Agentic AI is the new term to denote these AI-based autonomous systems. In real multi-agent systems, agents need to deal with malicious agents some of which can exploit security vulnerabilities to do harm. A memory poisoning attack is an example of exploiting a vulnerability to modify agents knowledge and, thus, affect its behavior. In an agent, there are different types of memory systems (semantic, episodic, and short-term) which have their own characteristics and for which different memory poisoning attacks may be developed. In this paper we focus on semantic memory, we discuss its attack, and mitigation strategies. We introduce private knowledge retrieval and private inference. We propose a signature system in the knowledge base to prevent from undesired manipulation. This should avoid memory poisoning attacks by introducing security by design in the construction of the semantic memory. In our architecture, an agent contains a local Prolog-like inference system (i.e., using Horn clauses to represent knowledge) which accesses an external knowledge base to get the required knowledge. This externally signed knowledge base is accessed privately. We provide different implementations in Python using different privacy levels including k-anonymity and information-theoretical privacy (IT-PIR).
Abstract (translated)
URL
https://arxiv.org/abs/2611.13696
PDF
-
Joint-State Inference for Autonomous Excavators Using Cab-View Vision-Control Fusion
2026-11-10 07:38:26Junwei Li, Jian Pan, Mengchi Cai, Shuai Zhao, Xizhuo Jiang, Zheng Wang, Qing Xu, Keqiang Li, Jianqiang WangAbstract
Reliable state feedback of excavator front-end mechanisms is essential for autonomous excavation, yet retrofitting existing machines with dedicated joint sensors, external cameras, or visual markers can increase deployment and maintenance burdens. This paper proposes a closed-loop autonomous excavation framework for continuous inference of excavator joint states using onboard visual observations and control-related temporal information. The framework encodes structural cues from the current cab-view image, models short-term motion trends from historical joystick commands, and introduces a rolling state buffer for the bucket joint to preserve temporal continuity under severe occlusion. Mechanical joint-range constraints are incorporated into the learning objective to suppress infeasible estimates. The current cab-view image and inferred states are supplied to an Action Chunking with Transformers (ACT)-based control policy for online action generation. Experiments in an AGX Dynamics for Unity simulation environment under diverse operating conditions evaluate joint-state inference and end-to-end task execution. Across 20 autonomous excavation trials, the framework achieved a 95% task-completion rate, with an average joint-state-inference-and-control latency of 10.11 ms per step. The bucket-joint inference achieved a mean absolute error (MAE) of 0.965 degrees. These results demonstrate a sensor-light and computationally efficient joint-state-inference approach that can support closed-loop autonomous excavation with reduced dependence on additional sensing infrastructure.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13690
PDF
-
Ego Guidance: Enhancing Robot Policy Learning with Action-Free Human Egocentric Videos
2026-11-10 07:33:35Lin Long, Tianren Ma, Zhong Chu, Di Wu, Maosong Cao, Qixiang Ye, Junbo Zhao, Shanghang ZhangAbstract
In this paper, we investigate how human egocentric videos can benefit robot policy learning without action supervision. Building on a pretrained VLM, we propose Shared Prediction on Action and Representation (SPAR), an MoT architecture in which future visual state prediction can provide independent and complementary learning signals alongside action prediction. We find that representations learned from egocentric videos improve downstream performance on out-of-distribution robot tasks, with further gains as video data scales. We also study the effects of objective choices and data curation, providing insights into how to leverage action-free videos for scalable robot learning.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13685
PDF
-
DexForesight: Grounding Dexterous Policies with Implicit Spatial Foresight
2026-11-10 07:28:24Zhongxi Chen, Yi Yang, Xinyi Chen, Mingzhe Ye, Yaping Li, Zhihong Liu, Congsheng Xu, Haonan Wang, Zhijie DengAbstract
Dexterous manipulation requires rich spatial foresight over both the surrounding scene and the robot hand. Existing world-action models often acquire such predictive knowledge through explicit future prediction. We introduce DexForesight, a framework that learns spatial foresight implicitly by injecting dynamics-aware world representations into the latent space of a dexterous policy. During training, DexForesight learns an action-conditioned spatial world model from offline scene depth, kinematic hand geometry, and demonstration trajectories. The world model predicts 3D hand-point motion and visual feature changes, capturing both articulated hand dynamics and scene changes under demonstrated actions, including the motion of fingers occluded from the cameras. We then freeze the world model and align policy queries with its predictive world queries, transferring the learned spatial dynamics into policy latents and using them to condition action generation. At deployment, only the policy is retained, operating on RGB observations, language, and robot state without depth input or explicit future prediction. On 11 DexJoCo simulation tasks, DexForesight improves the overall success rate of the base policy from 49.4% to 60.5%. On three real-robot dexterous manipulation tasks, DexForesight achieves 58.3% success, compared with 50.0% for the strongest baseline. Project page this https URL
Abstract (translated)
URL
https://arxiv.org/abs/2611.11166
PDF
-
A computational framework for understanding embodied plans
2026-11-10 07:15:38Aalap D. Shah, Basil Saeed, Joshua B. Tenenbaum, Tobias Gerstenberg, Hyowon Gweon, Ilker YildirimAbstract
Reasoning about people's plans and goals based on their actions is fundamental to human social interaction and essential for cooperative, embodied artificial intelligence. Formalizing this capacity in naturalistic, embodied settings has been a longstanding challenge. Here, we present GenEA, a novel agent representation capable of reasoning over multi-step, long-horizon embodied plans. At the heart of GenEA lies a multilevel state-space that couples the coarse biomechanics of an articulated body and its interactions with nearby objects by extending the task-and-motion planning framework from robotics with simulation-based intuitive physics. By embedding this state-space in a generative model, GenEA yields a joint probability distribution over goals and embodied plans. Across several experiments spanning the qualitatively distinct domains of object manipulation and navigation, featuring diverse tasks and multiple behavioral measures, we show that GenEA, through simple transformations of this joint distribution, recapitulates how humans reason about embodied plans. Crucially, GenEA consistently outperforms strong alternative models, including plausible heuristics, standard task-and-motion planning, and a state-of-the-art vision-language model. GenEA thereby offers a new perspective on the architecture of human reasoning and social cognition, suggesting that embodied artificial intelligence systems seeking to seamlessly interact with humans should adopt GenEA-like generative models of action planning.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13669
PDF
-
EgoHMP: Achieving Precise Human Motion Prediction in 3D Scenes via Egocentric Cues
2026-11-10 06:54:06Xin Zhao, Pengzhan Zhou, Ziyi Li, Riheng Jia, Zhida Qin, Yu LiuAbstract
The complex interactive dependencies between human behavior and 3D scenes make accurately predicting human intent a major challenge for human motion prediction (HMP). While recent studies use human gaze to infer intent, gaze signals suffer from ambiguity, drift, and a reliance on eye-tracking hardware. Research in cognitive science indicates that egocentric views inherently encode potential interactions, while head motion serves as a crucial precursor in motion planning. Motivated by this, we propose EgoHMP, a novel framework that uses egocentric views and head motions as robust carriers of interaction intent. By translating this intent into an envisioning of future interactions, EgoHMP achieves precise HMP in 3D scenes. EgoHMP first employs a vision-motion alignment encoder to align the visual and motion feature spaces. Based on these aligned features, an egocentric intent predictor with a modality-aware modulation mechanism adaptively suppresses invalid interaction cues to predict an accurate interaction probability map. Furthermore, to enhance the stability and realism of the generated motions, we propose a contact-aware motion decoder. Specifically, inspired by the human cognitive process of planning trajectories prior to action execution, we first predict a global trajectory to serve as a self-prompt, effectively mitigating the optimization conflict in diffusion models. Subsequently, we introduce a multi-stage denoiser, which relieves motion artifacts by incorporating human contact modeling through a multi-stage optimization strategy. Finally, a GCN-based motion decoder is employed to synthesize physically plausible and semantically consistent motions. Extensive experiments demonstrate that EgoHMP achieves state-of-the-art performance.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13654
PDF
-
picoEmergencyNet : Attention-Enhanced Ultra-Small Deep Learning Model for Aerial Image Classification in Emergency Monitoring
2026-11-10 06:28:44Christos KyrkouAbstract
Real-time disaster recognition from aerial imagery requires deep learning models that operate under strict memory, compute, and latency constraints. While recent approaches achieve high accuracy, their computational cost limits deployment on resource-constrained UAV platforms. This work introduces picoEmergencyNet, an ultra-lightweight convolutional neural network designed for aerial image disaster classification. The proposed architecture combines depthwise dilated convolutions with efficient channel attention and channel shuffling to enable multi-scale feature extraction at minimal computational cost. picoEmergencyNet has 23k parameters and 7M FLOPs, requiring only 0.1 MB of model storage. Evaluated on the AIDER and AIDERv2 disaster datasets, it achieves up to 96.1% accuracy, matching or exceeding state-of-the-art lightweight models that are orders of magnitude larger. Deployment experiments on the NVIDIA Jetson AGX Orin show that, without hardware-specific optimizations, picoEmergencyNet achieves 258 FPS on GPU and 46 FPS on CPU with a 7x lower activation memory footprint compared to competing models. These results demonstrate that careful architectural specialization, rather than model scaling, enables highly efficient and accurate TinyML-ready disaster recognition for real-time UAV-based emergency monitoring.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13638
PDF
-
Fade to... Red? Scheduling Coloured Noise Exploration Supports Efficient Robot Learning with Reduced Movement Jerk
2026-11-10 06:06:58Francisco M. L\'{o}pez, Francisco CruzAbstract
Exploration in deep reinforcement learning (RL) for continuous control is commonly driven by temporally uncorrelated Gaussian (white) action noise, which leads to jerky motions with locally constrained exploration. In contrast, temporally correlated coloured noise, particularly pink noise, has been shown to support more efficient exploration in RL. Motivated by these findings, we inquire whether scheduling changes to the noise colour over the course of training can outperform white noise for continuous control robotics. We perform a systematic study of 49 linear schedules between white and red noise, together with generalised state-dependent exploration (gSDE), on four sparse-reward manipulation tasks with a Fetch arm and a Shadow Dexterous Hand. We find that the noise colour at the end of training does not affect learning efficiency or end performance, whereas fading towards red noise halves the jerk of exploration in every task. Balancing learning efficiency, end performance and smoothness, red noise emerges as the preferred strategy: it learns as efficiently as pink noise with almost 2-fold less movement jerk. We therefore recommend exploration schedulers that fade towards red noise as the default for off-policy robot learning.
Abstract (translated)
URL
https://arxiv.org/abs/2611.11868
PDF
-
PUMA: Probability-Constrained Sequential Interaction for Forensic Localization
Abstract
Image manipulation localization must jointly exploit semantic structure and noise-sensitive forensic evidence. However, the effects of spatial correspondence, matching quality, and interaction order are often conflated. Existing semantic-forensic fusion methods provide complementary representations, but do not explicitly preserve quality-mismatch information during spatial correspondence or adaptively select the order in which information is propagated across branches. We propose PUMA (Probability-constrained Unbalanced Matching and Aggregation), which combines spatially constrained unbalanced optimal transport with position-wise convex selection between two sequential operators through a shared positional interface. Before normalization, PUMA retains signed transport-mass residuals, constructs row-stochastic cross-branch and within-noise-branch operators, and updates noise features through a gated residual path. PUMA achieves strong localization results on four test sets, with scores on COVERAGE and NIST16 exceeding the published results of the representative methods listed in our comparison. The experiments support the utility of local matching states and adaptive sequential interaction for cross-source localization and several post-processing conditions.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13621
PDF
-
Synapse-Control: Making Agent Experience Actionable with Lightweight Runtime Control
2026-11-10 05:50:12Zhiyuan Huang, Shengan Zheng, Chenghu Yang, Song Zhang, Lingyun Li, Chen Yu, Weiquan Huang, Linpeng HuangAbstract
Large language model agents reuse execution experience to improve reliability on complex tool-based tasks. Systems such as Sentry call large models to generate failure diagnoses before retrieving advice, adding inference cost to experience reuse. We present Synapse-Control, an agent runtime whose lightweight control plane selects advice and determines when to send it to the task-executing model. Rules shortlist relevant advice, and a lightweight 1.7B classifier fine-tuned on teacher-labeled state-suggestion pairs scores those candidates instead of generating a diagnosis. Micro-contracts encode reusable interface checks, such as required arguments or valid options, without prescribing complete task programs. A Prompt Value Table records earlier guidance and feedback to limit repeated suggestions through cooldown and token-budget checks. Across 1,308 attempts per method in AppWorld and WebShop, Synapse-Control achieves 270 successes (29.8% more than Base), replaces 4,987 generative diagnosis calls, and reduces auxiliary control time relative to Sentry by 50.2% and 51.0%, respectively.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13615
PDF
-
From Egocentric Human Video to Robot Manipulation in the Era of Embodied Intelligence
2026-11-10 05:36:17Yi Wang, Jingqiu Wu, Lei Yao, Zhaoshuo Li, Haoyu Zhu, Yirui Zhou, Huaiyuan Xu, Ying Zheng, Chuang Yang, Lap-Pui ChauAbstract
Egocentric human video (EHV) records close-range hand-object interaction (HOI) and the temporal structure of everyday activities, offering a potentially scalable source of experience for robot manipulation. Using this experience for robot control requires recovering useful supervision without recorded robot actions and accounting for differences in morphology, viewpoint, and action space. We define Ego2Robot learning as the process of learning or adapting robot manipulation policies from EHV, organized into four learning paradigms: explicit embodiment transfer, explicit action and task supervision, egocentric video representation pretraining, and latent action and dynamics modeling. The four paradigms differ in what they derive from human video: robot-format demonstrations, interpretable supervision, visual representations, or predictive models of actions and dynamics. For each paradigm, we trace what EHV information is extracted, how robot policies use it, and when the additional data is required. Besides, we review the benchmark datasets that cover EHV, robot demonstrations, and synthetic and simulated resources. This paper links EHV data's contribution to the task information preserved and the robot-specific grounding used during training or deployment. Finally, we identify challenges in physical feasibility, agent coordination, in-context adaptation, controlled evaluation, and predictive control.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13610
PDF
-
Compute the Interactions, Not the Attributes: What a Fact Layer Must Supply in the Chinese Sexagenary System
Abstract
Language-model systems in rule-governed domains often pair the model with a deterministic fact layer, but which facts that layer should supply is rarely measured. We measure it in the Chinese sexagenary system underlying BaZi chart analysis, where every fact has one table-checkable value: 682 items over twelve fact categories and 160 transit--natal collision items, posed closed-book, with the reference table, and with computed facts at increasing levels. In eleven open-weight models answering directly, errors concentrate where attributes must be combined: none gets more than 4\% of sexagenary days or 1\% of collisions right. Static attributes have no consistent effect, and the complete relation list leaves most direct-answering models near zero although it contains the answer. A query-filtered layer (exploratory) raises Qwen3-8B and Qwen3-32B to 78--92\% and four further models to 96--100\%. Three proprietary reasoning models compute 68--95\% of collisions unaided. On chess positions, under predictions locked in advance, attributes again do not substitute for relations and listed distractors are misattributed. The fact layer must supply computed interactions; how they are presented depends on the model and the domain.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13597
PDF
-
Towards Reliable VLM Judges: State-Conditional Invariance and Presentation-Aware Diagnostics
2026-11-10 05:15:50Qinan Zhang, Xinyu Wang, Qihang Jin, Tianyuan Wang, Hongbing Luo, Pengfei Li, Wenjun Lv, Yu Kang, Jian DiAbstract
Vision-language models (VLMs) are increasingly used as automated judges for multimodal and embodied tasks, yet aggregate accuracy or consistency alone does not reveal whether a judge changes its verdict for the right reason. We propose State-Conditional Invariance (SCI), a reliability principle requiring a judge to remain stable under semantics-preserving changes in presentation while responding to causal interventions that change the task outcome. We instantiate SCI in GridWM-Judge, a simulator-grounded diagnostic benchmark built from deterministic MiniGrid trajectories with aligned Full, evidence-removal, and counterfactual variants, plus presentation and representation probes. Across five diagnostic tasks, GridWM-Judge tests local perception, action-conditioned transitions, presentation stability, representation exchangeability, and Full-CF causal discrimination. Our diagnostic evaluations reveal heterogeneous failure modes, including sensitivity to evidence presentation and encoding, low activation on successful trajectories, and weak paired counterfactual discrimination. Perception-controlled analyses further indicate that visual grounding and objective recovery are substantial, model-dependent bottlenecks. These results motivate jointly reporting correctness, presentation stability, and causal sensitivity instead of relying on a single score.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13596
PDF
-
Transformers Provably Learn Graph Search: Training Dynamics and the Exponential Power of Depth
2026-11-10 05:09:36Xutao Ma, Somayeh SojoudiAbstract
Large Language Models (LLMs) have achieved remarkable progress on complex reasoning tasks, yet the mechanisms by which they acquire reasoning capabilities and the efficiency of their reasoning remain poorly understood. In this paper, we investigate the path-finding problem over directed graphs, a symbolic abstraction of multi-step reasoning. We first characterize the architectural requirements for this task, proving that one-layer transformers require $\Omega(N^2)$ size on dense graphs to implement the key DFS child-selection primitive. While a two-layer transformer implements full DFS with $O(N\log(N))$ size. Thus, a second layer is necessary and sufficient for near-linear DFS reasoning. We then extend this expressivity analysis to deeper models: for every $k\geq3$, a $k$-layer transformer implements a bounded-reachability DFS policy with look-ahead horizon $1+2^{k-3}$, establishing an exponential increase in reachability horizon with depth. Finally, going beyond expressivity, given appropriately curated training data, we show that two-layer transformers can provably learn to execute DFS via gradient flow and generalize to unseen graphs. Together, our results provide a mechanistic account of how transformers implement graph search, how depth expands bounded look-ahead, and how such reasoning algorithms can emerge through training.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13594
PDF
-
Few-Shot Residual Adaptation of Pretrained Latent World Models
2026-11-10 05:04:41Jakub Suliga, Pankhuri Vanjani, Nisarga Nilavadi, Wolfram Burgard, Rudolf LioutikovAbstract
Pretrained latent world models can support visual planning and control, but their predictions may fail when physical dynamics change at deployment. We study whether such models can adapt from only a few interactions without modifying their pretrained weights or identifying the changed physical parameters. We introduce a few-shot residual adaptation approach that learns self-supervised corrections from latent prediction errors. The corrections are applied both inside the latent predictor and directly to its predicted latent state, while the pretrained world model remains frozen. We evaluate the approach in simulation and on a real robot under center-of-mass and friction shifts. These changes reduce the success of the frozen model to near zero. Using three interaction episodes for adaptation, our dual-site approach recovers 90\% success under the center-of-mass shift and 100\% under the modified friction on the real robot. Ablations show that correcting both locations outperforms single-site adaptation. Our diagnostics further suggest that much of the learned correction reduces a general prediction mismatch, while successful control also depends on shift-specific information captured during adaptation.
Abstract (translated)
URL
https://arxiv.org/abs/2611.13592
PDF
