arXiv AI

arXiv:2610.24306v2 Announce Type: replace-cross Abstract: Location encoders learn compact representations of latitude--longitude coordinates for downstream geospatial prediction tasks. Existing approaches are pretrained using continuous or naturally occurring supervision, such as imagery, climate variables, or species distributions, and therefore do not explicitly capture administrative boundaries or place names, despite their strong influence on many real-world geospatial phenomena. We introduce, to our knowledge, the first method for learning administrative boundary-aware location encoders, using reverse-geocoded text as a self-supervised training signal. Our approach can be applied to existing location encoder architectures and consistently improves their downstream performance. We further introduce RevGeoRQ, a custom architecture tailored to reverse-geocoding supervision, which yields additional gains and exceeds contemporary location encoders on several downstream tasks. Finally, the use of human-interpretable reverse-geocoded strings and residual-quantized semantic IDs provides a more interpretable representation of geographic location than existing approaches. Code and model weights to be released soon.

arXiv:2610.23185v2 Announce Type: replace-cross Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has substantially improved Mixture-of-Experts (MoE) LLMs, but its outcome-level rewards provide only sparse supervision. On-Policy Distillation offers dense token-level feedback, yet requires a stronger compatible teacher, while on-policy self-distillation relies on hand-designed privileged information. We observe that the MoE architecture itself offers a way to resolve this dilemma, as increasing the activation budget per token exposes latent expert capacity and yields a potentially stronger teacher within the model. However, naive expansion degrades performance because newly co-activated experts are not optimized to work together. We therefore propose MoEvolve, a bidirectional distillation framework following a recovery-then-improvement strategy: we first restore the activation-expanded high-$k$ model and further strengthen it by exploiting its additional active capacity, then reverse the direction of distillation to transfer the enhanced capabilities back into the original low-$k$ configuration, retaining its inference-time activation budget. Across diverse MoE models and benchmarks, MoEvolve improves the initial models by up to $22.2$ domain-average points and outperforms same-budget RLVR baselines by up to $6.5$ points. Notably, the final low-$k$ models often surpass their activation-expanded teachers.

arXiv:2610.22887v2 Announce Type: replace-cross Abstract: Long rollout windows make reinforcement learning with verifiable rewards (RLVR) expensive, yet simply shortening them can remove the correctness signal: when every single-problem rollout truncates before an answer, group-relative accuracy advantages vanish. As a technical continuation of Batched Contextual Reinforcement (BCR), we study how concurrent rewards relax this dependence on long windows. Concurrent Constraint Reinforcement (CCR) places three independently verifiable problems in one capped training completion. Correct answers completed before truncation retain credit, while a shared budget encourages shorter useful reasoning without an explicit length penalty. We connect this mechanism to conditional reward- and update-variance analyses and study 4B and 8B mathematical reasoning models. Against single-problem GRPO with a 32k window, CCR with an 8,000-token window reduces tokens per training step from 975,464 to 174,957 on 4B and from 793,832 to 165,788 on 8B, saving 82.1% and 79.1%, respectively. An 8B training trace shows increasing accuracy reward alongside shorter completions. Single-problem evaluations provide supporting evidence of retained reasoning quality. These results identify reward-preserving training structure as a practical route to lower per-step cost in short-window RLVR.

arXiv:2610.22300v2 Announce Type: replace-cross Abstract: Not all facts a language model can recall support further learning equally well. We distinguish active facts, which support generalization beyond their own recall, from inert facts, which can be recalled but do not support such generalization in a given task. We study this distinction by constructing new relations from existing ones. Take the relation that maps a place to its continent (Paris -> Europe) and assign each of its answers an arbitrary new label (Europe -> crab, Asia -> zebra). The new relation then maps Paris -> crab. We fine-tune the model on the new relation for a subset of subjects, without telling it the original relation or the relabeling rule, and test whether it assigns the corresponding labels to held-out subjects that share an answer in the original relation. Having learned Paris -> crab, does it answer Munich -> crab? Across eleven language models and 70 relations, how linearly readable the original relation is in pretrained representations strongly predicts this generalization. Within each model, the two correlate across relations at Pearson r = 0.87 to 0.92. This association persists when controlling, separately and jointly, for the number of fine-tuning subjects and answers, the model's accuracy on the original relation, and how often the subjects occur in pretraining text. It is also robust to re-randomizing the label assignment and fine-tuning seed, and to replacing the animal labels with labels from thirteen other semantic categories. When we map a fine-tuned model's wrong label back to the original answer it was assigned to, Munich -> zebra becomes Munich -> Asia. Answers recovered this way are more often among the linear probe's top answers for the subject than among the pretrained model's own. Whether a new relation generalizes depends less on whether its facts can be recalled than on how readily their shared structure can be read out linearly.

arXiv:2610.22236v2 Announce Type: replace-cross Abstract: On-policy distillation consolidates specialist models into one student at the price of rollouts, teacher passes and gradient steps for every teacher. We show that the consolidation does not need training: with forward passes only, the capabilities of separately tuned specialists can be aggregated into one model that is, on average, more accurate than either specialist alone. When specialists share a base model, distillation has an exact layer-wise solution: each linear layer of the fused model must reproduce every specialist's output on the inputs that specialist receives while reading its own answers, a least-squares problem whose solution needs only the second moments of those inputs. One forward pass over a few hundred rollouts provides them. We call this training-free on-policy distillation. What the rollouts contain matters more than the solver: answers truncated in repetition loops are few among a mathematics specialist's rollouts but carry a large share of their tokens, and removing them lifts the fused model on AIME25 from task arithmetic's level to the specialist's at matched precision. On Olmo-3-7B, a few GPU-hours of forward computation per specialist yield a generalist that keeps most of each specialist's gain on each of seven mathematics and code benchmarks.

arXiv:2610.21337v3 Announce Type: replace-cross Abstract: On-policy distillation (OPD) transfers teacher knowledge to a student through feedback on student-generated rollouts, but generating complete rollouts and computing per-token teacher feedback are computationally expensive. Existing truncation methods save computation by stopping rollouts early. However, they do not compensate for truncated token gradients, which biases the gradient estimates and changes Vanilla OPD's original training objective. Furthermore, their truncation decisions rely on assumptions about supervision quality or criteria that may not generalize across settings. We introduce Stochastic Truncated On-Policy Distillation (STOPD), which accelerates OPD by randomly truncating student rollouts and correcting the resulting sampling bias with blockwise inverse-probability weighting (IPW). At each training step, we sample a stopping position for each rollout before generation and let the student generate until that position or natural termination. We obtain teacher feedback on every generated token and use blockwise IPW to compensate for the lower probability of retaining later tokens while controlling the weights. We prove that this weighting preserves Vanilla OPD's gradient sum in expectation and derive one-step update-error bounds for SGD, Adam, and Muon. We evaluate STOPD across multiple teacher-student model pairs, task domains, and optimizers. STOPD reduces generated tokens by 25-48% while maintaining accuracy comparable to Vanilla OPD, achieving 1.3-2.2x training speedups.

arXiv:2610.19388v2 Announce Type: replace-cross Abstract: Massively multilingual speech modeling has made remarkable progress in recent years, yet existing training resources and benchmarks still underrepresent much of the linguistic long tail. To address this gap, we introduce ELARSpeech, a large-scale multilingual speech resource comprising 30,202 hours of speech across 675 underrepresented languages, curated from the Endangered Languages Archive. Combining human-annotated and unlabeled recordings, the corpus spans diverse recording conditions, domains, and language families. Beyond corpus construction, our experiments show how ELARSpeech supports extending ASR and TTS capabilities to diverse underrepresented languages, with further benefits from pseudo-labeling unlabeled data. At the same time, our speech representation analysis reveals remaining limitations in transferring existing multilingual models to the linguistic long tail. Together, ELARSpeech provides both a resource and a testbed for extending multilingual speech technology further into the linguistic long tail. We publicly release the dataset, all trained ASR and TTS models, and the complete data construction pipeline to facilitate research on long-tail multilingual speech technologies.

arXiv:2610.18625v2 Announce Type: replace-cross Abstract: Relying on manipulable characteristics to target the allocation of resources is fraught with difficulties. Naively optimal rules lead to falsification and allocative inefficiency. Optimizing the rule for allocative efficiency while accounting for falsification typically induces falsification, at a heavy cost to the agents. Moreover, falsification generates negative externalities. We characterize optimal falsification-proof rules under a broad range of falsification technologies. We also provide an honesty condition under which they solve the unconstrained problem of a planner who maximizes a weighted sum of allocative surplus and agents' welfare.

arXiv:2610.18515v3 Announce Type: replace-cross Abstract: Large language models (LLMs) increasingly act as personal assistants; however, little is known about how they capture national-level preferences on a global scale. Given that "time" is a universal and representative concept, we examine LLMs' time preferences by varying only users' country and measuring the compensation LLMs deem appropriate for a 60-minute appointment delay. We then compare LLMs' responses with GDP per capita on Purchasing Power Parity (GDP_PPP), Mean Income, and a literature-based index (Patience Score). Our results across 193 countries and 5 LLMs demonstrate that people's time across the world is evaluated in noticeably different ways among LLMs, as well as varying levels of robustness under prompt variations. LLMs consistently recommend the highest compensation for Europe, but mostly the lowest for Africa. Based on GDP_PPP, LLMs underestimate the value of time for users living in Asia and South America. These findings indicate that geographical patterns are embedded in LLMs and may translate in different treatment in agentic settings.

arXiv:2610.17544v2 Announce Type: replace-cross Abstract: Thought templates, skill libraries, agent memories and caches turn many reasoning steps into one. We call this reasoning contraction. A contracted memory saves reasoning steps and errors until its facts change, then quietly answers wrong. Most systems trust memories by counting successes and failures. We prove that such counters are badly tuned change detectors: a symmetric counter notices a partial change only after a power of its false-alarm horizon, and an unclipped one may never notice it. On five LLMs answering multi-hop questions over edited facts, ExpeL- and ACE-style counters serve 7.0-13.5x more stale answers than certified memories given the same information, and on code tasks with an exact verifier 35-73x more. Safety also has a price. When facts change and each memory is judged only by its own verified outcomes, one number, the certification load, sets how much reasoning any safe memory can skip, and under Zipf demand it yields a scaling law. Our recipe weighs each memory's failures by its own likelihood ratio, certifies its utility against live reasoning, and restarts the certificate when facts change, with the likelihood-ratio detector as the safety net for unannounced changes; for large edits it also invalidates the memory.

arXiv:2610.17309v2 Announce Type: replace-cross Abstract: We introduce StartLux-Decision, decision models from 0.8B to 27B parameters that combine efficient single-pass prediction with adaptive, confidence-guided reasoning. The Instant models use bidirectional attention in full-attention layers to improve decision accuracy while keeping single-pass inference. The AutoThink models invoke explicit reasoning for uncertain decisions and answer confident cases directly. A three-stage training pipeline establishes shared decision capabilities, specializes the two variants, and equips AutoThink with reasoning through reasoning supervision and reinforcement learning. We introduce Reinforcement Learning for Calibrated Deliberative Decisions (RLCD$^2$), which uses a frozen model and proper scoring rules to reward improvements in decision probability quality and applies a soft penalty to overlong reasoning traces. RLCD$^2$ raises the share of reasoning traces that terminate within the generation limit by up to 48 percentage points and cuts average reasoning length by up to 65%, while largely preserving decision accuracy. On the public subset of Decision Index 0.3, the Instant models outperform the strongest evaluated comparable-scale open-weight decision models by 7.4-16.9 points, and the 27B AutoThink model reaches 73.4. We also introduce DecisionPivot, a programmatically verifiable benchmark of computational decisions built from contrast pairs that differ in a single answer-changing detail. On DecisionPivot, AutoThink outperforms its Instant counterparts by 11.5-18.4 points in pair accuracy and leads among evaluated systems. The models can be run using frameworks such as vLLM and MLX, with quantized versions available to reduce memory usage.Inference code and links to model weights are available at https://github.com/StartLuxLabs/Startlux-Decision .

arXiv:2610.11376v2 Announce Type: replace-cross Abstract: Latent LiDAR pipelines suffer from flying pixels: convolutional VAEs blur sharp radial depth discontinuities, yielding edge depths that back-project to points floating between surfaces. We identify this as a major, directly correctable decoder bottleneck and introduce CRISP: a pixel-space diffusion decoder with a backbone-agnostic latent adapter, DiT-based denoiser, and support mask predictor. CRISP replaces video-VAE and LiDAR-native decoders alike while keeping the encoder and latent generator fixed. Across KITTI-360, SemanticKITTI, and nuScenes, replacing only the decoder reduces FSVD/FPVD by 50.5% on average across frozen backbones; for generic video VAEs, the reductions reach 71%/74%. On the LiDAR-native LiDM backbone, FRID drops by 71%, with the largest gains at depth discontinuities. In a pretrained LiDM world model, the same zero-shot replacement improves FSVD by 15.5%, narrowing the sim-to-real gap. Code and checkpoints will be released at https://andrea25512.github.io/CRISP/.

arXiv:2610.10460v2 Announce Type: replace-cross Abstract: Multi-teacher on-policy distillation (MOPD) is used in two settings. In common-domain composition, several teachers score each student rollout from one prompt domain and their signals form a single target; in routed-domain distillation, prompts from different domains are assigned to the corresponding specialist. Both settings usually transfer each teacher's endpoint policy, which mixes what post-training changed with preferences inherited from the teacher's base. We introduce $\Delta$-MOPD, which transfers each teacher's teacher-minus-base logit shift re-anchored at the student's frozen initialization, and compare it with endpoint supervision in both settings while holding teacher selection fixed. We first expose the mechanism that impedes endpoint transfer: inherited base pull can exceed the post-training shift. Removing it reduces the teacher-term norm ratio and target--student KL. Across our experiments, the results suggest that shift targets are particularly useful when teacher signals are combined at a state. With three composed teachers, $\Delta$-MOPD exceeds endpoint composition by $4.11$ Math and $1.95$ five-benchmark points; with two, it matches endpoint accuracy. Under phased routing, it achieves higher mean performance in both phase orders and reduces the observed order gap from $10.50$ to $6.42$ points. Under interleaved routing, where each update involves one teacher, the two targets perform comparably. The phased results provide supporting evidence that the benefit may extend to signals accumulated across training phases. Target construction is thus an independent design axis in MOPD, complementary to teacher selection. The experiment repo could be find at https://github.com/HJSang/Delta-MOPD

arXiv:2610.06814v3 Announce Type: replace-cross Abstract: World models learn to predict how their environment will evolve, making them an important foundation for general-purpose robotic control. Yet world action models depend on camera inputs whose manipulation can corrupt the visual representations used across tasks and action policies. Existing attacks on these models optimize against the victim's actions or predicted futures and therefore require access to target-model outputs. In this paper, we propose an attack, TAPDreamer, against world action models that instead uses a public encoder alone to construct a fixed local perturbation that transfers across tasks and action architectures. TAPDreamer requires no target-policy queries. Our key insight is that interactions between patch-induced changes in attention weights and value vectors broadcast a nearly identical representation shift far beyond the patch footprint, and this shift remains stable across task observations. Guided by this insight, TAPDreamer uses six frames from one source task to maximize the global L1 distance between clean and patched encoder representations. In closed-loop evaluation, one frozen patch per benchmark, covering about 6.5% of the input, reduces FastWAM's success rate from 97.7% to 0.0% across 40 LIBERO tasks and from 90.86% to 0.0% across 50 RoboTwin tasks; matched random patches retain 81.5% and 79.2% success. The same patches reduce success to 1.45% and 1.00% on two DreamWAM configurations and to 10.60% on Motus. These results show that protecting downstream action generation alone is insufficient: defenses for world action models must also secure shared visual encoders against persistent local perturbations.

arXiv:2610.04953v2 Announce Type: replace-cross Abstract: Euclidean combinatorial optimization problems (ECOPs), such as the Traveling Salesman Problem (TSP) and Capacitated Vehicle Routing Problem (CVRP), possess inherent symmetries under the two-dimensional Euclidean group E(2), including rotations, reflections, and translations. Existing learning-based methods, including recent diffusion-based methods, rely on data augmentation or regularization to approximate E(2)-equivariance. This paper presents EDISCO, the first discrete diffusion model for ECOPs with exact E(2)-invariant generative distributions over node-index solutions. EDISCO introduces an E(2)-equivariant edge-score network coupled with a categorical continuous-time Markov chain over discrete edge variables, and exact posterior sampling provides efficient multi-step inference. This design gives EDISCO a local geometric inductive bias: edge neighborhoods with the same relative geometry and combinatorial context are represented consistently regardless of absolute position or orientation, making learning more efficient and inference more robust than non-equivariant methods. EDISCO outperforms previous learning-based state-of-the-art solvers on synthetic TSP from 100 to 10000 nodes and CVRP from 50 to 2000 customers, while using only 33-50% of the training instances. Trained only on uniform synthetic data, EDISCO also outperforms competing learning-based baselines under spatial distribution shift and CVRP constraint-tightness shift. Code is available at https://github.com/ValleyC/EDISCO.

arXiv:2610.03759v3 Announce Type: replace-cross Abstract: Ocean wave forecasting is essential for maritime safety, offshore operations, and coastal resilience, yet remains challenging due to systematic biases in physics-based models. Physical models, while widely used, rely on approximations and parameterizations that limit their accuracy under complex ocean-atmosphere conditions. To enhance ocean wave forecasting, we propose BridgeCast, within a physics-AI hybrid framework for bias correction. BridgeCast is a probabilistic model based on conditional flow matching (CFM) that learns to transform physical model forecasts into reanalysis-like fields. It treats physical forecasts as corrupted observations and employs a continuous-time generative process to bridge their distribution toward that of reanalysis data. BridgeCast is parameterized by a Transformer-based architecture that enables spatiotemporal modeling, incorporation of exogenous atmospheric variables, and flexible inference via both ordinary and stochastic differential equation formulations. Extensive experiments on real-world datasets demonstrate that BridgeCast consistently outperforms state-of-the-art baselines across regions and forecast lead times.

arXiv:2610.02198v2 Announce Type: replace-cross Abstract: Several state-of-the-art methods for online reinforcement learning in continuous control improve policies using action gradients of a learned critic. However, critics are typically trained to predict returns, and accurate value predictions do not necessarily yield accurate action derivatives, potentially leading to unreliable policy updates. We propose Forward Entropy-Regularized Policy Optimization (FERPO), an on-policy maximum entropy reinforcement learning algorithm that performs policy improvement using critic values without differentiating the critic with respect to actions. FERPO derives an optimal target action distribution from a policy-improvement objective regularized by entropy and Kullback-Leibler (KL) divergence. We then fit the actor to this target by minimizing a forward-KL objective, estimated using self-normalized importance sampling (SNIS) with actions drawn from the rollout policy. By limiting the target distribution's deviation from the rollout policy, the KL regularization helps keep these importance weights well behaved. In contrast to reverse-KL objectives, which can favor a subset of the target distribution's modes, the forward-KL objective encourages coverage of multiple high-value modes and thereby promotes exploration. Experiments and ablations on MuJoCo Playground and ManiSkill show competitive performance and sample-efficiency gains. Computational benchmarks also demonstrate faster actor updates than Relative Entropy Pathwise Policy Optimization (REPPO).

arXiv:2610.00661v3 Announce Type: replace-cross Abstract: Diffusion large language models (dLLMs) generate text by denoising a sequence or successive blocks, allowing several tokens to be revealed in parallel. Reinforcement learning with verifiable rewards (RLVR) reuses terminal feedback across these decisions, even as their conditioning context changes. We propose stepwise risk-sensitive GRPO (StepRS-GRPO), which varies the risk coefficient of the group-advantage transformation across denoising states while retaining the underlying trainer. For binary rewards, we show that this transformation is exactly a prompt- and state-dependent rescaling of centered outcome advantages. A capability-based calibration suggests a coefficient scale, while endpoint and interpolation ablations guide schedule selection. Across multiple dLLM backbones and mathematical reasoning benchmarks, StepRS-GRPO improves both pass@1 accuracy and pass@k coverage over centered GRPO, while increasing answer diversity. In our ablation studies, mass-matched controls support the contributions of state allocation and schedule direction, and the gains persist after matching the root mean square (RMS) of the advantages to that of centered GRPO. Reasoning-trace diagnostics further show that the diversity gains from StepRS-GRPO extend beyond final-answer strings.

arXiv:2609.38578v3 Announce Type: replace-cross Abstract: Cross-structural motion retargeting aims to transfer motion between different skeletal topologies. Despite recent progress, existing state-of-the-art models struggle with reliability in zero-shot settings, i.e. skeletons with different topologies which were unseen during training, and recent Transformer-based attempts have failed to outperform specialized geometric methods. We bridge this gap with a Transformer Autoencoder that learns a topology- and translation-invariant latent space. Our core contribution is a learnable flattening of skeletal graphs that captures both local dependencies and global structure. Unlike the standard transformer architecture, which adds positional information to token content, we integrate graph-based positional encodings multiplicatively, a design choice that follows directly from our flattening formulation. The resulting model handles diverse skeletal topologies within a single unified architecture and trains in a fully unsupervised manner, requiring no paired retargeting data. Ablation studies show, that the graph encodings, multiplicative formulation, and Transformer backbone is critical for the performance. In zero-shot evaluations, our method reduces global joint position error by $43-47\%$ over current benchmarks. A user study ($n = 37$), including expert animators, further ranks our approach highest in motion alignment and physical plausibility ($p < 0.05$). These results demonstrate that our model design is key to making transformer architectures effective for motion retargeting, outperforming existing approaches.

arXiv:2609.37976v2 Announce Type: replace-cross Abstract: LLMs trained with Chain-of-thought excel in reasoning capability, but often come with excessive token cost. We find that the core of reasoning capacity lies in the Thinking model's weight component within the null space of a projection defined by the corresponding Non-thinking model's dominant singular directions, and removing the subspace component can largely improve reasoning efficiency without hurting the accuracy gained during thinking-mode post-training. Unlike existing efforts that mostly operate within the dominant subspace, we are the first to unveil the critical role of the null space and harness it for model optimization. Motivated by this finding, we propose Spectral Null-Space Swap ($S^3$), a training-free composition of paired Non-thinking and Thinking checkpoints. Our method keeps the Non-thinking model inside its own dominant subspace and takes the Thinking checkpoint outside it, improving reasoning efficiency while maintaining accuracy. We extensively evaluate $S^3$ on 2B-30B dense and mixture-of-experts (MoE) architectures spanning 28 evaluation environments across mathematical, multimodal, and audio reasoning domains. $S^3$ establishes new empirical Pareto Frontiers among training-free model composition strategies: across all settings, it reduces inference token overhead by an average of 27.4% compared to full Thinking models while simultaneously improving overall task accuracy by 1.0 percentage point (e.g., yielding +8.3% accuracy on HMMT25 alongside a 33.0% token speedup). We further use attention entropy for explanation and find that the retained component produces more concentrated attention, and we use a simplified analytical model about optimization to demonstrate why null-space can effectively reduce attention entropy, thereby improving the efficiency of reasoning.