arXiv subjects

An T. Le

Publications and source records attributed to An T. Le.

At least 19 recordsLinked to original sources

arXiv2026

Radiative steering of warp shells

We construct exact timelike junctions between compact flat cavities and a Kinnersley photon-rocket exterior. The shells satisfy the strict surface dominant energy condition and steer subluminally under nonnegative radiation. Burns joined at static spheres give $m_f/m_i=e^{-3L}$ for exterior velocity-space path length $L$; the supported cavity center obeys a separate sharp bound. Self-similar shells require anisotropic stress and have growing normal modes with the ambient geometry and coefficients held fixed. A separate elastic model admits local self-gravitating recoil control with outward emission. At zero gravitational coupling, exact unstrained turns yield sharp fuel bounds and travel limits at fixed duration. Limits on angular concentration and inward emission reduce efficiency. For smooth emission and positive inner-face density, stress-free recoil requires photons to enter the cavity. Self-gravitating settling after a turn remains open.

gr-qc↗

arXiv2026

Observer-robust energy condition verification for warp drive spacetimes

Whether a warp drive metric requires exotic matter is decided by energy conditions quantified over all observers, not only the Eulerian. Each of the null, weak, strong and dominant conditions is equivalent, at a point, to feasibility of a $4\times4$ linear matrix inequality $A_{ab}+σg_{ab}\succeq0$, by the S-lemma, with $A_{ab}$ the stress-energy tensor or its trace reverse and the dominant condition a conjunction of two such tests. It forms no eigendecomposition of $T^a{}_b$, imposes no rapidity cap and assumes no Hawking-Ellis type, so it decides all four alike, Types I and IV not being exhaustive; its multiplier margin is exactly half the null-cone minimum, so the same test returns the severity. Composed with an interval enclosure of the curvature chain it decides a point from the metric itself, not from a floating-point copy of its stress-energy. At Type I each condition reduces instead to an eigenvalue inequality holding for all observers at once. The type label is numerical and tolerance-bound; the reported severities are rapidity-capped diagnostics, not certificates. Everything decided uses only boost-invariant data and stays well posed through $v_s=1$. On a flat slice the Eulerian momentum that opens the Type-IV wall vanishes only for a gradient shift, so among four matched drives the irrotational Rodal geometry is Type I identically, its shift curl-free by an exact profile identity, while Alcubierre and Natário are Type-IV dominated at every sampled speed and Van den Broeck above its transition. A single-frame reading of Rodal misses about 73% of its wall weak-energy violations. All four violate the pointwise null energy condition at every sampled speed, consistent with the Santiago-Schuster-Visser no-go, whose null step is conditional. Both are realized in warpax, a JAX toolkit building $T^a{}_b$ by automatic differentiation.

gr-qc↗

arXiv2026

Anytime Global Tensor Motion Planning

Global Tensor Motion Planning (GTMP) solves motion planning with batched tensor operations over a layered multipartite graph. We generalize GTMP so that adjacent-layer edges are realized by any black-box local planner (e.g., linear interpolation, splines, sampling-based planning, trajectory optimization, or generative sampling). We provide two anytime policies on top of this generalization: Anytime GTMP with random restarts at a fixed budget, which covers every homotopy class almost surely, and AO-GTMP with informed expansion with growing budgets, which converges to the optimal cost. We prove that a single sampled graph covers every endpoint-fixed homotopy class admitting a $δ$-clear representative of bounded length. We also prove that additional samples per layer reduce the per-layer miss probability exponentially, whereas stronger local planners reduce the required layer count only sublinearly. On manipulation benchmarks the method matches state-of-the-art performance, and on 2D navigation it returns batches of topologically diverse solutions, while the informed baselines concentrate on one or two classes.

cs.RO↗

arXiv2026

Training Non-Differentiable Networks via Optimal Transport

We optimize losses that jump: spiking thresholds, quantized layers, and discrete routing put jumps in the forward pass, where backpropagation does not apply. Finite differences fail: at a derivative-estimating radius, 99.5% of probe pairs on a quantized network leave the loss bit-identical, against 1.6% on a smooth control. At a jump, Clarke and conservative stationarity are undefined, and any radius-$h$ smoothing steepens like $1/h$, so no vanishing radius defines a derivative. Fixed-resolution stationarity survives. PolyStep attains it from forward passes alone, ranking probe points on a rotated polytope and stepping along a softmax-weighted average of their directions, the $λ=0$ endpoint of a KL-penalized transport program. A second impossibility fixes the frame: on a plateau wider than the probes reach, every rule that reweights one cost row freezes, softmax included, and only the column constraint escapes. In expectation, the cost-weighted average of the probed directions is exactly a gradient step on a smoothed loss, for every bounded measurable objective, with no probe asked to clear the jump set. Over $T$ steps it reaches subspace stationarity at rate $O(T^{-(1/2-γ)})$ up to a bias floor with an interior optimal probe radius, upgrading to Goldstein stationarity where the loss is Lipschitz. At matched optimizer steps, PolyStep leads all 36 comparisons of six architectures against six gradient-free baselines. It reaches 93.0% on hard LIF spiking networks against 79.6% for the best tuned baseline, and satisfies 92.6% of clauses on million-variable MAX-SAT against an 87.5% random floor. Matching total evaluations instead reverses argmax attention and one MAX-SAT size, and levels hard MoE. A step costs one forward pass per subspace dimension, leaving PolyStep at chance at 4.2M parameters from scratch; where gradients exist Adam is faster and more accurate.

cs.LG↗

arXiv2026

StructSAM: Structure- and Spectrum-Preserving Token Merging for Segment Anything Models

Recent token merging techniques for Vision Transformers (ViTs) provide substantial speedups by reducing the number of tokens processed by self-attention, often without retraining. However, their direct application to the Segment Anything Model (SAM) family is nontrivial: SAM's image encoder mixes windowed and global attention, and its mask decoder relies on dense, prompt-conditioned features for precise boundary prediction. We systematically evaluate representative token-merging methods on SAM and Medical SAM in a strict off-the-shelf setting, and find that existing destination-selection heuristics can erode boundaries and leak prompt information as merge rates increase. We propose \textbf{StructSAM}, a resolution-preserving merge-unmerge framework tailored to SAM. StructSAM computes a lightweight token-energy score from first-order feature gradients, uses grid-based flatness screening to protect boundary and prompt regions, and merges tokens within flat areas toward low-energy destinations with explicit token recovery. We further provide a spectral graph coarsening view showing that score-guided merging yields bounded Laplacian spectral distortion compared to random or window-restricted baselines. Across eight natural and medical benchmarks, StructSAM reduces encoder FLOPs by 25-30\% (up to 40\%+ with prompt-aware merging) with minor drops in mIoU/Dice, consistently outperforming ToMe, PiToMe, ToMeSD, VidToMe, and ALGM at the same compute.

cs.CV↗

arXiv2026

vla.cpp: A Unified Inference Runtime for Vision-Language-Action Models

Vision-Language-Action (VLA) policies are typically shipped as Python/PyTorch stacks that assume a workstation-class GPU, a mismatch for the hardware on which robots actually run. We present vla.cpp, a portable C++ inference runtime built on llama.cpp. To our knowledge, it is the first ggml-class engine to natively serve the flow-matching and diffusion VLA inference pattern, in which a cached vision-language prefix is consumed by a cross-attending action expert integrated over several solver steps. A single runtime serves seven architectures spanning five backbone and four action-head families behind one request/response protocol, with each model packaged as a self-contained bundle. On LIBERO-Object, the engine matches a state-of-the-art checkpoint to within one episode out of 200, and runs BitVLA at 100% success in 1.3 GiB of memory. The same bundle runs unchanged across three hardware tiers, from a consumer GPU down to an 8 GB embedded module. A cross-hardware roofline analysis shows that batch-1 VLA inference is compute-bound, so utilization rather than bandwidth is the deployment lever; an IMMA ladder GEMM derived from this analysis cuts BitVLA per-step latency by 4.5x. We then frame an on-robot stress test on an ALOHA arm that isolates the latency constraint under which a learned VLA must replan against a moving target on the hardware it was trained for. Code, demo videos, and the reproducible benchmark scaffold are available at https://fai-modelopt-tech.github.io/vla-cpp.github.io/.

cs.RO↗

arXiv2026

TACT-ful: Multi-Channel Terrain Affordance and Compliance Training for Payload-Robust Perceptive Humanoid Locomotion

Foothold selection on structured terrain requires explicit reasoning about contact planarity, surface steepness, and kinematic reachability, properties not captured by a single height-based terrain signal. We propose a multi-channel terrain cost combining flatness, steepness, and velocity-aware height feasibility, plus a forward climb reward, that simultaneously drives a GPU-parallel divergent component of motion (DCM) foothold planner and shapes a dense per-step affordance reward for an asymmetric actor-critic policy trained with proximal policy optimization (PPO) from depth images. A Bézier swing trajectory with adaptive apex bias extends foothold tracking to joint position-and-orientation, using the arc tangent to guide sole orientation through riser crossings and tread landings. To support payload tasks, we introduce a lower-body compliance training procedure in which a virtual wrench is injected at a sampled load attachment point, generating physically consistent force and moment; wrench-aware compliance targets replace rigid pose penalties, and the policy learns to yield to load-induced perturbations without force sensing. The full system trains end-to-end with standard PPO, no distillation, and no teacher-student staging, and is deployed on a humanoid directly from simulation with configuration changes only. In simulation, the policy reaches $1.0~\mathrm{m/s}$ on stairs with risers up to $0.20~\mathrm{m}$ and improves payload robustness up to ${\sim}15~\mathrm{kg}$ centered load and for moment-dominated wrist loads without fine-tuning. We also provide a qualitative hardware demonstration on structured terrain. Project website: https://fai-rl-tech.github.io/tact-locomotion.github.io/

cs.RO↗

arXiv2026

On the boundary cost of source-consistent warp shells

We study classical energy-condition admissibility for subluminal, positive-energy warp shells. For the constructions examined here, the energy-condition failures are localized at the smooth source--vacuum transition rather than in the bulk interior. We introduce two \emph{source-first} shell ans"atze whose metric potentials are obtained from the Einstein constraints for a prescribed matter model: a shift-free S-shell and a T-shell whose shift is derived from the momentum constraint. We assess them with a five-criterion standard comprising regularity, constraint satisfaction, an explicit matter model, frame-independent energy-condition margins, and global diagnostics; the standard responds to the source-consistency critique of Barzegar, Buchert, and Vigneron. Applied to eight constructions spanning the canonical warp-drive classes, none passes the full standard. An independent frame-independent verification of the Fuchs constant-velocity shell confirms interior energy-condition compliance (0 of 13 interior probes violate) but reveals Hawking--Ellis Type~IV violations in the smoothing tail beyond the nominal shell. A frame-independent scan over shell compactness and thickness (600 configurations) yields no admissible configuration in either source-first class. The same boundary deficit appears in the shift-free S-shell and persists in the static $v_0=0$ limit, which ties it to the transition geometry rather than to the shift. Along a representative off-axis null ray the null-energy line integral is nevertheless positive for every source-prescribed shell; this is an exploratory diagnostic rather than a proof of the averaged null energy condition, but it shows that the pointwise boundary failures need not appear in that integral.

gr-qc↗

arXiv2026

Rarity of rocket-driven Penrose extraction in Kerr spacetime

We study rocket-driven Penrose extraction in the test-particle limit on a fixed Kerr background for equatorial prograde flybys under explicit steering prescriptions. A spacecraft ejects exhaust inside the ergosphere; when the exhaust attains negative Killing energy, the remaining spacecraft gains energy by 4-momentum conservation. Across 320{,}000 simulated trajectories spanning black-hole spin, exhaust velocity, and orbital parameters, extraction with escape is rare in broad parameter scans (at most ${\sim}1\%$) and requires high spin ($a/M\gtrsim 0.89$), highly relativistic exhaust ($v_e\gtrsim 0.91c$), and finely tuned initial conditions. Under optimal tuning the success rate reaches ${\sim}70\%$ at $a/M = 0.95$. For representative escape trajectories, a single periapsis impulse is more propellant-efficient than the continuous-thrust controllers studied here. All quoted thresholds are empirical and specific to the orbit family, prior, and steering protocol studied.

astro-ph.HE↗

arXiv2026

AAC: Admissible-by-Architecture Differentiable Landmark Compression for ALT

We introduce \textbf{AAC} (Architecturally Admissible Compressor), a differentiable landmark-selection module for ALT (A*, Landmarks, and Triangle inequality) shortest-path heuristics whose outputs are admissible by construction: each forward pass is a row-stochastic mixture of triangle-inequality lower bounds, so the heuristic is admissible for \emph{every} parameter setting without requiring convergence, calibration, or projection. At deployment, the module reduces to classical ALT on a learned subset, composing end-to-end with neural encoders while preserving the classical toolchain. The construction is the first differentiable instance of the compress-while-preserving-admissibility tradition in classical heuristic search. Under a matched per-vertex memory protocol, we establish that ALT with farthest-point-sampling landmarks (FPS-ALT) has provably near-optimal coverage on metric graphs, leaving at most a few percentage points of headroom for \emph{any} selector. AAC operates near this ceiling: the gap is $0.9$--$3.9$ percentage points on 9 road networks and ${\leq}1.3$ percentage points on synthetic graphs, with zero admissibility violations across $1{,}500+$ queries and all logged runs. At matched memory, AAC is also $1.2$--$1.5{\times}$ faster than FPS-ALT at the median query on DIMACS road networks, amortizing its offline cost within $170$--$1{,}924$ queries. A controlled ablation isolates the binding constraint: training-objective drift under default initialization, not architectural capacity; identity-on-first-$m$ initialization closes the expansion-count gap entirely. We release the module, a reusable matched-memory benchmarking protocol with paired two-one-sided-test (TOST) equivalence and pre-registration, and a reference compressed-differential-heuristics baseline.

cs.AI↗

arXiv2026

ReFineVLA: Multimodal Reasoning-Aware Generalist Robotic Policies via Teacher-Guided Fine-Tuning

Vision-Language-Action (VLA) models have gained much attention from the research community thanks to their strength in translating multimodal observations with linguistic instructions into desired robotic actions. Despite their advancements, VLAs often overlook explicit reasoning and learn the functional input-action mappings, omitting crucial logical steps, which are especially pronounced in interpretability and generalization for complex, long-horizon manipulation tasks. In this work, we propose ReFineVLA, a multimodal reasoning-aware framework that fine-tunes VLAs with teacher-guided reasons. We first augment robotic datasets with reasoning rationales generated by an expert teacher model, guiding VLA models to learn to reason about their actions. Then, we fine-tune pre-trained VLAs with the reasoning-enriched datasets with ReFineVLA, while maintaining the underlying generalization abilities and boosting reasoning capabilities. We also conduct attention map visualization to analyze the alignment among visual observation, linguistic prompts, and to-be-executed actions of ReFineVLA, reflecting the model is ability to focus on relevant tasks and actions. Through this additional step, we explore that ReFineVLA-trained models exhibit a meaningful agreement between vision-language and action domains, highlighting the enhanced multimodal understanding and generalization. Evaluated across a suite of simulated manipulation benchmarks on SimplerEnv with both WidowX and Google Robot tasks, ReFineVLA achieves state-of-the-art performance, in success rate over the second-best method on the both the WidowX benchmark and Google Robot Tasks.

cs.RO↗

arXiv2025

Model Tensor Planning

Sampling-based model predictive control (MPC) offers strong performance in nonlinear and contact-rich robotic tasks, yet often suffers from poor exploration due to locally greedy sampling schemes. We propose \emph{Model Tensor Planning} (MTP), a novel sampling-based MPC framework that introduces high-entropy control trajectory generation through structured tensor sampling. By sampling over randomized multipartite graphs and interpolating control trajectories with B-splines and Akima splines, MTP ensures smooth and globally diverse control candidates. We further propose a simple $β$-mixing strategy that blends local exploitative and global exploratory samples within the modified Cross-Entropy Method (CEM) update, balancing control refinement and exploration. Theoretically, we show that MTP achieves asymptotic path coverage and maximum entropy in the control trajectory space in the limit of infinite tensor depth and width. Our implementation is fully vectorized using JAX and compatible with MuJoCo XLA, supporting \emph{Just-in-time} (JIT) compilation and batched rollouts for real-time control with online domain randomization. Through experiments on various challenging robotic tasks, ranging from dexterous in-hand manipulation to humanoid locomotion, we demonstrate that MTP outperforms standard MPC and evolutionary strategy baselines in task success and control robustness. Design and sensitivity ablations confirm the effectiveness of MTP tensor sampling structure, spline interpolation choices, and mixing strategy. Altogether, MTP offers a scalable framework for robust exploration in model-based planning and control.

cs.RO↗

arXiv2025

DoublyAware: Dual Planning and Policy Awareness for Temporal Difference Learning in Humanoid Locomotion

Achieving robust robot learning for humanoid locomotion is a fundamental challenge in model-based reinforcement learning (MBRL), where environmental stochasticity and randomness can hinder efficient exploration and learning stability. The environmental, so-called aleatoric, uncertainty can be amplified in high-dimensional action spaces with complex contact dynamics, and further entangled with epistemic uncertainty in the models during learning phases. In this work, we propose DoublyAware, an uncertainty-aware extension of Temporal Difference Model Predictive Control (TD-MPC) that explicitly decomposes uncertainty into two disjoint interpretable components, i.e., planning and policy uncertainties. To handle the planning uncertainty, DoublyAware employs conformal prediction to filter candidate trajectories using quantile-calibrated risk bounds, ensuring statistical consistency and robustness against stochastic dynamics. Meanwhile, policy rollouts are leveraged as structured informative priors to support the learning phase with Group-Relative Policy Constraint (GRPC) optimizers that impose a group-based adaptive trust-region in the latent action space. This principled combination enables the robot agent to prioritize high-confidence, high-reward behavior while maintaining effective, targeted exploration under uncertainty. Evaluated on the HumanoidBench locomotion suite with the Unitree 26-DoF H1-2 humanoid, DoublyAware demonstrates improved sample efficiency, accelerated convergence, and enhanced motion feasibility compared to RL baselines. Our simulation results emphasize the significance of structured uncertainty modeling for data-efficient and reliable decision-making in TD-MPC-based humanoid locomotion learning.

cs.RO↗

arXiv2025

Global Tensor Motion Planning

Batch planning is increasingly necessary to quickly produce diverse and quality motion plans for downstream learning applications, such as distillation and imitation learning. This paper presents Global Tensor Motion Planning (GTMP) -- a sampling-based motion planning algorithm comprising only tensor operations. We introduce a novel discretization structure represented as a random multipartite graph, enabling efficient vectorized sampling, collision checking, and search. We provide a theoretical investigation showing that GTMP exhibits probabilistic completeness while supporting modern GPU/TPU. Additionally, by incorporating smooth structures into the multipartite graph, GTMP directly plans smooth splines without requiring gradient-based optimization. Experiments on lidar-scanned occupancy maps and the MotionBenchMarker dataset demonstrate GTMP's computation efficiency in batch planning compared to baselines, underscoring GTMP's potential as a robust, scalable planner for diverse applications and large-scale robot learning tasks.

cs.RO↗

arXiv2025

TD-GRPC: Temporal Difference Learning with Group Relative Policy Constraint for Humanoid Locomotion

Robot learning in high-dimensional control settings, such as humanoid locomotion, presents persistent challenges for reinforcement learning (RL) algorithms due to unstable dynamics, complex contact interactions, and sensitivity to distributional shifts during training. Model-based methods, \textit{e.g.}, Temporal-Difference Model Predictive Control (TD-MPC), have demonstrated promising results by combining short-horizon planning with value-based learning, enabling efficient solutions for basic locomotion tasks. However, these approaches remain ineffective in addressing policy mismatch and instability introduced by off-policy updates. Thus, in this work, we introduce Temporal-Difference Group Relative Policy Constraint (TD-GRPC), an extension of the TD-MPC framework that unifies Group Relative Policy Optimization (GRPO) with explicit Policy Constraints (PC). TD-GRPC applies a trust-region constraint in the latent policy space to maintain consistency between the planning priors and learned rollouts, while leveraging group-relative ranking to assess and preserve the physical feasibility of candidate trajectories. Unlike prior methods, TD-GRPC achieves robust motions without modifying the underlying planner, enabling flexible planning and policy learning. We validate our method across a locomotion task suite ranging from basic walking to highly dynamic movements on the 26-DoF Unitree H1-2 humanoid robot. Through simulation results, TD-GRPC demonstrates its improvements in stability and policy robustness with sampling efficiency while training for complex humanoid control tasks.

cs.RO↗

arXiv2025

Machine Learning with Physics Knowledge for Prediction: A Survey

This survey examines the broad suite of methods and models for combining machine learning with physics knowledge for prediction and forecast, with a focus on partial differential equations. These methods have attracted significant interest due to their potential impact on advancing scientific research and industrial practices by improving predictive models with small- or large-scale datasets and expressive predictive models with useful inductive biases. The survey has two parts. The first considers incorporating physics knowledge on an architectural level through objective functions, structured predictive models, and data augmentation. The second considers data as physics knowledge, which motivates looking at multi-task, meta, and contextual learning as an alternative approach to incorporating physics knowledge in a data-driven fashion. Finally, we also provide an industrial perspective on the application of these methods and a survey of the open-source ecosystem for physics-informed machine learning.

cs.LG↗

arXiv2025

FlowMP: Learning Motion Fields for Robot Planning with Conditional Flow Matching

Prior flow matching methods in robotics have primarily learned velocity fields to morph one distribution of trajectories into another. In this work, we extend flow matching to capture second-order trajectory dynamics, incorporating acceleration effects either explicitly in the model or implicitly through the learning objective. Unlike diffusion models, which rely on a noisy forward process and iterative denoising steps, flow matching trains a continuous transformation (flow) that directly maps a simple prior distribution to the target trajectory distribution without any denoising procedure. By modeling trajectories with second-order dynamics, our approach ensures that generated robot motions are smooth and physically executable, avoiding the jerky or dynamically infeasible trajectories that first-order models might produce. We empirically demonstrate that this second-order conditional flow matching yields superior performance on motion planning benchmarks, achieving smoother trajectories and higher success rates than baseline planners. These findings highlight the advantage of learning acceleration-aware motion fields, as our method outperforms existing motion planning methods in terms of trajectory quality and planning success.

cs.RO↗

arXiv2024

Grasp Diffusion Network: Learning Grasp Generators from Partial Point Clouds with Diffusion Models in SO(3)xR3

Grasping objects successfully from a single-view camera is crucial in many robot manipulation tasks. An approach to solve this problem is to leverage simulation to create large datasets of pairs of objects and grasp poses, and then learn a conditional generative model that can be prompted quickly during deployment. However, the grasp pose data is highly multimodal since there are several ways to grasp an object. Hence, in this work, we learn a grasp generative model with diffusion models to sample candidate grasp poses given a partial point cloud of an object. A novel aspect of our method is to consider diffusion in the manifold space of rotations and to propose a collision-avoidance cost guidance to improve the grasp success rate during inference. To accelerate grasp sampling we use recent techniques from the diffusion literature to achieve faster inference times. We show in simulation and real-world experiments that our approach can grasp several objects from raw depth images with $90\%$ success rate and benchmark it against several baselines.

cs.RO↗