Probability
Equilibrium Joining Strategies for Queues in Two-Phase Random Environment
Konstantin Avrachenkov and Uri Yechiali
We study
equilibrium joining strategies in an M/M/1-type queueing system with strategic customers operating in a two-phase random environment described as a continuous-time Markov process. Strategic customers, upon arrival, choose whether to join or to balk based on available information and anticipated utility, considering the trade-off between reward from service and waiting cost. Four observation scenarios are analysed: fully observable (both queue length and environment phase are disclosed to a customer upon arrival), queue-only observable, environment-only observable, and fully unobservable. In each case,
equilibrium joining strategies are analysed. In the environment-only observable and fully unobservable cases, explicit solutions and
equilibrium conditions are derived under rapid oscillations and under very slow transitions between environment phases.
Read the full abstract ▾
Artificial Intelligence
ALLOT: Budgeted Hybrid-Memory Routing for Knowledge Updates in LLMs
Shanfeng Huang, Zhou Fang, Song Xiao et al. (+1 more)
For large
language models (LLMs), parametric adaptation is costly when
retrieval already suffices. We introduce ALLOT, a hybrid-
memory routing framework that separates learned write priority from a hard parametric budget. A
memory-aware router combines frozen text representations,
retrieval confidence, and relation metadata; a single ranking supports multiple write budgets while preserving all facts in external
memory. On CounterFact with Qwen3-4B, ALLOT reaches 0.760 accuracy at a 20% parametric-write budget and recovers 78.4% of the budget-matched oracle gain, with 80% fewer parametric writes than dual-writing every fact. At this budget, jointly adding
retrieval and relation features to text improves normalized oracle gain by 6.2 percentage points. Complementary Qwen3-0.6B shared-store results achieve dual-write-level accuracy with 6-14.5% parametric writes, and cross-
benchmark transfer retains approximately 88% of in-domain gain. These results support allocating adaptation capacity according to its incremental value rather than treating every factual update as an equally valuable training target.
Read the full abstract ▾
All authors Shanfeng Huang, Zhou Fang, Song Xiao, Hai Du
Computer Vision
OpenMASC: An Open-Source Pipeline for Cross-Trajectory Metal-Aware Sampling and Correction in Accelerated MRI
Zhengyi Lu, Ming Lu, Chongyu Qu et al. (+11 more)
Metal implants corrupt MRI measurements throughout $k$-space, yet existing accelerated MRI methods assume clean data and most metal artifact reduction approaches assume fully sampled acquisitions. No public
dataset provides paired $k$-space and images with and without metal for the same anatomy, and no framework jointly addresses artifact-aware acquisition and reconstruction across sampling trajectories. We present OpenMASC, an open-source pipeline covering the full workflow from data generation to deployment. A physics-based data generation module converts public CT volumes into paired clean and metal-corrupted MRI data in both Cartesian and radial formats. MA-VarNet, an unrolled reconstruction network with a per-cascade DC Rectifier, corrects artifacts that data-consistency steps reintroduce from corrupted measurements. A reinforcement learning
agent actively selects $k$-space readouts and co-trains with the reconstruction network through a decoupled three-stage procedure. The framework is trajectory-agnostic except for the data-consistency operator, supporting both Cartesian and radial acquisition without architectural changes. Experiments on two
datasets at $4\times$ and $8\times$ acceleration demonstrate consistent improvements over conventional and learned baselines on both trajectories.
Read the full abstract ▾
All authors Zhengyi Lu, Ming Lu, Chongyu Qu, Junchao Zhu, Junlin Guo, Marilyn Lionts, Yanfan Zhu, Yuechen Yang, Tianyuan Yao, Jayasai Rajagopal, Bennett Allan Landman, Xiao Wang, Xinqiang Yan, Yuankai Huo
Quantum Gases
Pancake-shaped vortex droplets in dipolar molecular BECs
Zibin Zhao, Xinyi Tang, Tianmiao Zhang et al. (+6 more)
Anisotropic interactions profoundly affect topological excitations in
quantum fluids. Motivated by recent advances in the studies of microwave-shielded polar molecules, we introduce self-trapped modes in the form of pancake-shaped
quantum droplets (QDs) with embedded vorticity, which are maintained by strongly anisotropic dipole-dipole interactions. The stability region of singly charged ($S=1$)
vortices nearly coincides with that of the ground-state QDs ($S=0$), demonstrating the robustness of the
vortex states. The strong anisotropy of the system splits the core (pivot) of
vortex QDs with $S=2$ into separated
unitary ones. The angular momentum and stability of the state with $S=2$ are affected by the separation between the
unitary cores. Higher-charge
vortex QDs with $S>2$ are stable too, for sufficiently large particle numbers and inter-core separations. On the other hand, bound
vortex-antivortex pairs with $S=\pm 1$ are unstable. Head-on collisions between the
vortex QDs exhibit distinct regimes, including rebound, merger, and fragmentation. Tuning the cylindrically symmetric component of the dipolar interaction reveals a pronounced sign-dependent response: positive tuning preserves the self-bound
vortex, whereas negative tuning drives expansion and fragmentation. The results demonstrate that the microwave-dressed molecular QDs offer a robust platform for the realization of self-trapped
vortex states, demonstrating how the strong anisotropy reshapes their structure, stability, and dynamics.
Read the full abstract ▾
All authors Zibin Zhao, Xinyi Tang, Tianmiao Zhang, Zhaopin Chen, Huanbo Luo, Guihua Chen, Bin Liu, Boris A. Malomed, Yongyao Li
Machine Learning
A Comparative Analysis of Attention versus State-Space Models for In-Context Learning
Enes Arda, Semih Cayci and Atilla Eryilmaz
Transformers and state-space models (SSMs) are two prominent sequential learning architectures, yet their comparison remains largely empirical and existing theoretical analyses are typically task-specific or architecturally restricted. In this paper, we develop belief geometry, a unified analytical framework for comparing the representational capabilities of broad classes of
attention and SSMs. Starting from a generalized formulation of in-context linear regression and using cumulative Bayes regret as our measure, we abstract three capabilities required by many sequential learning problems in our belief geometry: evidence assembly, belief maintenance, and addressing. We then study three cases of our formulation that isolate these capabilities and yield sharp architectural lessons: For belief maintenance, SSMs attain the optimal regret over stationary aggregation kernels; for positional assembly, SSMs have a
memory advantage; and for content addressing, softmax
attention has an exponential width advantage over sigmoid-selective SSMs. Experiments with LLaMA-type
Transformers and Mamba-2 show that these architectural insights extend beyond our analytically tractable classes and linear-regression testbed.
Read the full abstract ▾
Computer Vision
SGA-Flow-GRPO: Spatial Gradient-Guided Credit Assignment for Flow-GRPO
Yunkai Yang, Yudong Zhang, Xinying Chen et al. (+5 more)
Reinforcement Learning (RL) has proven effective in aligning flow-based generative models with human preferences. Recently, Flow-GRPO has emerged as an efficient critic-free paradigm by calculating advantages over sampled candidate trajectories. However, standard Flow-GRPO applies a uniform scalar advantage across both temporal denoising steps and spatial latent dimensions, without explicitly accounting for the spatial structure of generated images, which may lead to sub-optimal policy updates. To address this, we propose a novel gradient-guided spatial credit assignment framework tailored for
Diffusion Transformers (DiTs). We first reformulate the transition-level log-likelihood in Flow-GRPO into a
token-wise representation natively aligned with DiT patch architectures, constructing spatially fine-grained importance sampling ratios. To allocate localized credit without rigid, boundary-sensitive
segmentation heuristics, we introduce a continuous spatial credit map derived from reward gradients. Crucially, we employ an outlier-robust normalization scheme based on Median Absolute Deviation (MAD) coupled with temperature scaling, effectively eliminating gradient noise while highlighting functional
prompt-aligned regions. Extensive evaluations on GenEval show that our approach delivers SOTA
alignment quality, achieving a convergence rate comparable to top-tier methods like DiffusionNFT while substantially improving upon Flow-GRPO-based methods in
alignment performance.
Read the full abstract ▾
All authors Yunkai Yang, Yudong Zhang, Xinying Chen, Bin Luo, Jienan Lyu, Kunquan Zhang, Weitao Wan, Runmin Dong
Artificial Intelligence
Enabling Timely Guidance before Skill Retrieval: Retaining Helpful Warm Tips in Agent Context
Feng Liang, Yupeng Li, Runhao Zeng et al. (+2 more)
Reusable skills help LLM-based
agents solve complex tasks, but the
agent must receive guidance before it commits to an ineffective approach. Existing skill mechanisms often expose only metadata and load full content on demand, leaving useful guidance unavailable until the
agent decides to retrieve it. General
memory methods can incur substantial maintenance overhead, while keeping guidance in conversation context risks repeatedly exposing the
agent to irrelevant or harmful advice. We propose TipsWarm, a mechanism that complements existing skill mechanisms by maintaining a budgeted pool of skill-derived keypoints, or warm tips, for selective injection into the context of every message turn. By separating event-triggered LLM assessment from inexpensive per-turn screening, it makes transferable skill guidance readily available while controlling maintenance costs. In three coding and iterative task-execution
benchmarks, TipsWarm achieves the highest task success rate while remaining time-efficient, compared to recent skill and general
memory baselines.
Read the full abstract ▾
All authors Feng Liang, Yupeng Li, Runhao Zeng, Francis C. M. Lau, Xiping Hu
Logic
The effects of forcing on exacting and ultraexacting cardinals
Philipp Lücke and Alejandro Poveda
Motivated by recent work of Aguilera–Bagaria–Lücke and Aguilera–Bagaria–Goldberg–Lücke (ABGL), we analyze various effects of set-theoretic forcing upon the classes of exacting and ultraexacting cardinals. Using Magidor support products of Prikry forcings, we prove that an I2-
embedding yields a transitive, set-sized model of ZFC with a proper class of exacting cardinals. We show that under Woodin's HOD
Conjecture, no exacting cardinal can be a limit of exacting cardinals. In contrast, we prove that models of ZF containing large cardinals beyond choice have forcing extensions that are models of ZFC in which a regular cardinal is a stationary limit of ultraexacting cardinals. We also show that, assuming the HOD
Conjecture, an analogue of the classical Levy–Solovay
theorem holds for exacting and ultraexacting cardinals. Finally, we analyze the large cardinal properties possessed by exacting cardinals in HOD. We prove that if $\lambda$ is exacting and $V_\lambda$ satisfies the HOD Hypothesis, then $\lambda$ has strong large cardinal properties in HOD. Carrying forward forcing constructions from ABGL, we start with a model of ZFC with an I2-
embedding and produce a model of ZF in which the successor of an exacting cardinal $\lambda$ is extendible in all models of the form $\mathrm{HOD}_x$ for $x\subseteq \lambda$.
Read the full abstract ▾
Game Theory
Top Trading Cycles with Indifferences under Matroid Constraints
Yasushi Kawase
We study the reallocation of indivisible objects among
agents who may hold initial endowments and may be indifferent between objects. Each
agent has her own set of admissible objects, and a matroid constraint specifies which combinations of assigned objects are feasible. The model contains the housing
market, house allocation with newcomers who hold no endowment, and the assignment of students to schools under matroidal constraints at each school. For this model, we introduce the top-class trading cycles mechanism (TCTC), which extends the top trading cycles mechanism (TTC) to matroid constraints and weak preferences. When a trading cycle is selected, TCTC fixes each
agent's top indifference class rather than a specific object and postpones the choice within the class to the end. TCTC is individually rational, Pareto efficient, and weakly group-strategy-proof on the full domain of weak preferences, and it runs in
polynomial time. When preferences and reports are strict, TCTC is strongly group-strategy-proof. Together with a known impossibility result, our
theorem establishes a maximal domain of school-by-school constraints for individual rationality, Pareto efficiency, and strategy-proofness.
Read the full abstract ▾
Robotics
WSM-Aware HRI: An IoT-Enhanced Framework for Early Detection and Norm-Guided Repair of Failures with LLM Guidance
Hanlin Zhang, Yuquan Wang, Tianwei Zhang et al. (+1 more)
Human-
robot interaction (HRI) failures remain a major barrier to deploying
robots in real-world environments. Prior work often treats failures as isolated technical faults or focuses on post-hoc recovery behaviors. In practice, many breakdowns arise because humans and
robots operate under inconsistent assumptions about the current world state. We propose WSM-Aware HRI, an IoT-enhanced modular framework that unifies diverse HRI breakdowns as World-State Mismatches (WSMs) between a human's instruction-implied assumptions and a
robot's grounded world model built from
multimodal perception and digital augmentation. A Large
Language Model (LLM) is used to make implicit assumptions explicit, map them to a small set of mismatch types, and specify the evidence needed for verification against the
robot's world state. WSM-Aware HRI shifts failure handling from execution-time recovery to proactive mismatch detection during intention formation, enabling interventions guided by safety, norm compliance, and multi-user coordination with transparent explanations. We evaluate mismatch identification in ten everyday cases spanning both visual and latent-state mismatches. The system can accurately produce the expected output results, and ablations show that reliable identification depends on appropriate grounding representations and verification-oriented refinement. These results indicate that treating interaction breakdowns as explicit world-state mismatches enables earlier detection of impending failures and offers a principled mechanism for integrating external evidence and social constraints into human-
robot interaction.
Read the full abstract ▾
All authors Hanlin Zhang, Yuquan Wang, Tianwei Zhang, Zhenglong Sun
Differential Geometry
A decomposition theorem for elliptic structures of minimal rank on compact Lie groups and applications
Guido Drei and Max Reinhold Jahnke
We show that every left-invariant
elliptic structure of minimal rank on an odd-dimensional compact Lie group admits an algebraic decomposition analogous to those obtained by Pittie in the complex case and by Charbonnel and Khalgui for left-invariant CR structures of hypersurface type. We then use this decomposition to prove that the
cohomology associated with such an
elliptic structure can be reduced to a computation on a suitable maximal
torus.
Read the full abstract ▾
Functional Analysis
Duals of separable Banach Spaces as Calkin Algebras and Universal Ideal Quotients
Rui Liu and Jie Shen
For $\mathbb{K}=\mathbb{R}$ or $\mathbb{C}$ and every separable
Banach space $V$, we use an oracle-relative two-sorted finite-extension construction, whose oracle records the local rational structure of $V$, to construct a separable
Banach space $X_V$ such that $$ \operatorname{Cal}(X_V)=\mathcal{B}(X_V)/\mathcal{K}(X_V)\simeq \begin{pmatrix} \mathbb{K}&0\\ V^*&\mathbb{K} \end{pmatrix} $$ as
Banach algebras. The identification of $V^*$ with the Jacobson radical is isometric. Consequently, every nonzero dual
Banach space with a separable predual admits, after an equivalent renorming, a unital
Banach-
algebra structure isomorphic to the Calkin
algebra of a separable
Banach space. Moreover, there is a separable
Banach space $X$ such that every separable
Banach space is isometric to $\mathcal{J}/\mathcal{K}(X)$ for a closed two-sided ideal $\mathcal{J}$ of $\mathcal{B}(X)$, and the subspace–ideal correspondence preserves the canonical order.
Read the full abstract ▾
Computer Vision
Progressive-View On-Policy Distillation for Regional-to-Global Transfer in Multimodal LLMs
Shanfeng Huang, Zhou Fang, Song Xiao et al. (+1 more)
Regional-to-global
distillation uses crop-conditioned guidance to improve full-image understanding. The challenge is to effectively transfer the teacher's crop-based advantage to the student's full-image inference. We propose progressive-view on-policy
distillation (PVD), which shifts the student's view distribution from the crop toward the full image through an intermediate aspect-preserving padded crop. The padded crop preserves regional content while matching the full image's visual-
token grid. Across stages, the view mixture assigns increasing probability to the full image. A lightweight regional-advantage weighting reallocates
token-level supervision using the crop-conditioned teacher-student log-probability gap. Evaluated under each sampled input, it applies mild reweighting when the gap is small and emphasizes higher-gap
tokens when the gap widens. A Jensen-Shannon metric decomposition interprets this schedule as a shift from matched-input imitation toward the deployment objective. Across
benchmarks spanning perception, visual mathematics and general
multimodal question answering, PVD-full reaches an average accuracy of 77.51 over three seeds, improving on the reward-free
distillation baseline by 2.01 points and on its reward-matched variant by 1.00 point. In the reward-free setting, PVD-distill still gains 1.16 points.
Read the full abstract ▾
All authors Shanfeng Huang, Zhou Fang, Song Xiao, Hai Du
Machine Learning
When the Merge Coefficient Stops Mattering: Proximity Regularized Merging for Continual LoRA Adaptation
Yixuan Liu, Yuhao Sun, Sen Song et al. (+1 more)
Rehearsal-free continual learning with parameter-efficient adapters can be cast as a sequence of task-vector write-in operations: for each new task, a low-rank adapter is learned and merged into a running model. We propose Proximity Regularized Merging (PRM), a minimal modification to sequential LoRA merging that adds a proximal penalty during task-vector training without changing the subsequent write-in rule. PRM acts as a robust task-vector regularizer: in the reported Base->+Prox diagnostics, it improves AAA across multiple write-in rules, backbones, and class-incremental settings, while its fixed-coefficient variant remains competitive with strong coefficient-based baselines. Mechanistically, matched-prefix norm controls and proximal-strength sweeps show that proximal training shrinks the task-vector radius, lowers Fisher-weighted interference, broadens the coefficient plateau, and exposes a stability-plasticity trade-off. Together, these results suggest that the effectiveness of sequential LoRA merging depends not only on how much of a task vector is written in, but also on whether the task vector itself has been trained to be mergeable.
Read the full abstract ▾
All authors Yixuan Liu, Yuhao Sun, Sen Song, Jin Li
Computation and Language
Language Distances are Practical for Equitable Cross-Lingual Transfer
York Hay Ng, Razan Ahsan Rifandi, Aditya Khan et al. (+1 more)
Cross-lingual transfer is strongly conditional on how the source
language is chosen, but it is impractical to determine the best candidate source for every target
language, especially for low-resource target
languages.
Language distances are widely used to rank candidate sources due to their correlation with transfer efficacy and applicability in resource-sparse settings. However, the reliability of distance-based rankers across tasks and resource levels remains underexplored. We therefore present the first equity-focused evaluation of paradigms for ranking source
languages, studying resource-level inequality and task inequality across ten cross-lingual tasks and two
multilingual models. While both inequalities are most pronounced for individual
language distances and an English-always baseline, they are substantially reduced by training-free composite distances, and nearly eliminated by trained rankers. We further demonstrate the reliability of rankers using
language distances compared to rankers using
language model internals. Overall, we find that
language distances provide a practical basis for equitable and performant transfer
language selection. We recommend using trained rankers when task-specific transfer evaluations are available, and composite distances otherwise.
Read the full abstract ▾
All authors York Hay Ng, Razan Ahsan Rifandi, Aditya Khan, En-Shiun Annie Lee
Functional Analysis
Inverse-Limit Detection Envelopes as Calkin Algebras
Rui Liu and Jie Shen
For a unital
Banach algebra $A$ with $\|1_A\|=1$ and a decreasing sequence $(I_r)$ of proper closed two-sided ideals of finite codimension, we introduce the bounded detection envelope $$\widehat A:=\operatorname{Env}^{b}_{(I_r)}(A)=\varprojlim^{b}(A/I_r,\pi_r^s),\qquad \|(a_r)_r\|=\sup_r\|a_r\|.$$ Combining vector-valued Bourgain–Delbaen extensions, oracle Turing-machine recursion, and Argyros–Haydon operator analysis, we construct a separable
Banach space $X$ such that, isometrically as unital
Banach algebras, $$\operatorname{Cal}(X)=\mathcal{B}(X)/\mathcal{K}(X)\simeq\widehat A. $$ We also characterize these envelopes as the unital
Banach algebras with norm-one units admitting an isometric predual and countably many jointly faithful weak-
star-to-norm continuous finite-dimensional unital representations. Applications include arbitrary finite-dimensional unital
Banach algebras with norm-one units and their countable $\ell_\infty$-sums, as well as unitized coordinatewise
algebras over normalized $1$-unconditional boundedly complete Schauder bases. These include incidence
algebras, $\ell_\infty$, and the unitization of Tsirelson's original space. Further applications concern analytic operator
algebras,
Fourier–Stieltjes and
Fourier multiplier
algebras of countable discrete groups, and classical and
quantum convolution
algebras. In particular, we realize $H^\infty(\mathbb D)$, $\operatorname{Mult}(H_2^2)$, $\mathcal L_2$, $M_{\mathrm{cb}}A(\mathbb F_2)$, and $C^u(O_3^+)^*$ as Calkin
algebras, with their canonical norms and products.
Read the full abstract ▾
Quantum Physics
Exact High-Temperature Quantum Area Law
Ahmad Yousefi and Ali T. Rezakhani
We prove a $\beta^{2}$ area law for the
quantum mutual information of thermal states of local
lattice Hamiltonians: for any bipartitioning $A|B$ and all inverse temperatures $\beta$ below a critical value $\beta^{*}$, we show $\mathcal{I}_{\beta}(A,B) \leqslant \tilde{f}(\beta)\, \beta^{2}\,|\partial_{AB}|$, where $|\partial_{AB}|$ denotes the size of the boundary region and $\tilde{f}(\beta)$ remains finite as $\beta \to 0$. This quadratic scaling improves the relationship between high-temperature correlations and macroscopic thermodynamics. While a linear area law implicitly permits a loose, temperature-independent upper bound on extractable work and binding energy, the quadratic scaling ensures that the nonequilibrium free energy available after decoupling regions and the total energy of the interface both are proportional to the area of the interface of $A|B$ and vanish as $\beta$ at high temperatures. These results show that the high-temperature mutual information and correlation are suppressed, respectively, quadratically and linearly in inverse temperature, which are consistent with the expected thermal decoupling in macroscopic thermodynamic quantities.
Read the full abstract ▾
Machine Learning
Continual Data Unlearning in Diffusion Models via Transition-based Regularization
Sunbeom Jeong, Sehwan Kim, Sangwoo Hong et al. (+1 more)
Data unlearning in
diffusion models aims to remove the influence of specific training examples without suppressing the broader concepts they represent. However, when deletion requests arrive sequentially, updates for new requests can degrade generative utility and undermine earlier deletions. We propose a continual data unlearning framework that uses completed deletion transitions as directional references to regularize future updates. For each request, we record changes in denoiser responses on the same fixed noisy inputs before and after unlearning. Rather than matching full post-deletion responses, we apply a one-sided penalty that discourages reversal along the recorded directions relative to the post-deletion references, while leaving orthogonal response changes and progress beyond these references unpenalized. To keep storage independent of the number of requests, we maintain a fixed-capacity bank of representative transition records. Records are selected based on the local sensitivity of progress along their recorded directions to parameter updates, allowing them to be retained even when their penalties are inactive. Empirical evaluations show that the proposed framework achieves a better balance between deletion persistence and generative utility than existing unlearning baselines as requests accumulate, using only a small transition
memory.
Read the full abstract ▾
All authors Sunbeom Jeong, Sehwan Kim, Sangwoo Hong, Jungwoo Lee
Artificial Intelligence
HyperReCo: Retrieving and Connecting Evidence with Hypergraph Neural Networks for LLM Multi-hop Reasoning
Zicheng Zhao, Linhao Luo, Junnan Dong et al. (+4 more)
Large
language models (LLMs) have shown strong capabilities, with
retrieval-augmented generation (RAG) supporting complex multi-hop
reasoning by retrieving evidence distributed across documents. Graph-based approaches exploit connections among evidence, and hypergraph-based
retrieval further preserves higher-order entity associations within documents and connects documents through shared entities. However, existing hypergraph retrievers often rely on predefined structural expansion or
diffusion, which may miss query-dependent interactions needed to identify relevant evidence. They also leave connections among retrieved evidence implicit, requiring LLMs to reconstruct these connections before
reasoning. Therefore, we propose HyperReCo, a framework for retrieving and connecting evidence with a hypergraph neural network (HyperGNN). We represent each document as a hyperedge over its extracted entities, with shared entities connecting the hyperedges. Through hypergraph message passing with joint supervision over documents and entities, the HyperGNN learns query-dependent interactions to retrieve complementary evidence. We further introduce Gradient-Guided Hyper-Path Decoding (GGHD), which uses gradient attribution to interpret the learned interactions and translate them into explicit hyper-paths that help LLMs combine complementary facts for multi-hop
reasoning. Experiments on six
benchmarks show that HyperReCo achieves the best
retrieval performance among the compared methods on all three multi-hop QA
datasets, together with strong downstream QA performance. Case studies and further analyses demonstrate the utility of decoded hyper-paths for connecting retrieved evidence.
Read the full abstract ▾
All authors Zicheng Zhao, Linhao Luo, Junnan Dong, Haoran Luo, Xiaoli Li, Shirui Pan, Chen Gong
Artificial Intelligence
RLHarness: Co-evolving Procedural Skills with Reinforcement Learning for Long-horizon Multimodal Reasoning
Ziqiao Shang, Zian Xu, Ji-Chen Yan et al. (+6 more)
Multimodal reasoning requires models to preserve visual evidence through long decision chains while selecting appropriate procedures across diverse scenarios and rules. When learning is guided only by terminal verifiers, reinforcement learning (RL) reveals whether a final answer is correct but not how it should be produced. The policy must therefore discover reusable
reasoning procedures while learning to execute them, creating a program cold-start problem. Skills can externalize successful procedures, reduce repeated exploration, and provide inspectable guidance. However, a fixed Skill Bank assumes that this guidance remains compatible with an evolving policy, while updating Skills alone can leave their triggers, execution protocols, and demonstrations stale or mutually inconsistent. We introduce RLHARNESS, which organizes Skills, selection and execution protocols, few-shot demonstrations, and task contracts into a unified, versioned Harness and alternates Harness
evolution with policy learning. An Exploration-
Distillation Harness builds the initial Harness and version-aligned verified traces for SFT and DAPO I. After the first RL block, a Post-RL Reconstruction Harness rebuilds Skills, protocols, and demonstrations from fresh success-failure rollouts, and DAPO II adapts the policy to the reconstructed program. RLHARNESS improves Accuracy from 16.25%/27.50% to 62.00%/50.00% on MetroMap/TravelMap and raises F1 score from 37.13%/45.50% to 65.81%/65.51% on Fee-VL/Cancel-VL. All four tasks achieve their best results only after reconstruction and DAPO II, showing that an evolving Harness complements RL by continually updating the external program that the policy learns to execute.
Read the full abstract ▾
All authors Ziqiao Shang, Zian Xu, Ji-Chen Yan, Weiming Wu, Ziyi Jia, Jie Meng, Tao Huang, Shan Huang, Lan-Zhe Guo
Machine Learning
Active Feature Acquisition With Incomplete Training Data
Reza Rezvan, Valter Schütz, Han Wu et al. (+2 more)
In many prediction tasks, acquiring all features can be a prohibitively expensive or outright impossible task. Further, in many cases a static subset of features may not be enough to solve the problem sufficiently across various instances. Active Feature Acquisition (AFA) addresses these problems by formalizing the trade-off between feature cost and predictive performance during sequential feature selection. However, prior AFA work largely assumes access to complete training data, an assumption that is often violated in practice. Here we study AFA with Incomplete Training Data (AFA-ITD), showing that under missing completely at random (MCAR) data, one-step acquisition values remain unchanged, whereas multi-step values can decrease. We analyze three approaches to learning from incomplete data: aliasing, filtering, and generative restoration. We show that filtering can require a number of training instances scaling exponentially with the dimension, whereas generative restoration scales exponentially with the acquisition budget. We empirically test our theory on a controlled experiment and across common AFA
datasets and find that missingness mainly damages methods that exploit multi-step acquisitions and that generative restoration is able to recover lost performance in many experiments. Code is available at https://github.com/Linusaronsson/AFA-
Benchmark/tree/missing-data.
Read the full abstract ▾
All authors Reza Rezvan, Valter Schütz, Han Wu, Linus Aronsson, Morteza Haghir Chehreghani
High Energy Astrophysical Phenomena
15-Year X-ray Study Links Variable Mass Accretion to Changing Look Behavior in AGN NGC 7582
Santanu Mondal, V. Jithesh, Neeraj Kumari et al. (+1 more)
Changing Look Active
Galactic Nuclei (CLAGN) are a subclass of AGN that exhibit spectral state transitions in both optical/UV and X-ray bands on timescales of days to years. In this work, we studied fifteen years of X-ray data of the CLAGN NGC 7582 taken from Suzaku, XMM-Newton, and NuSTAR to understand the origin of the CL behavior. Our study reveals that the obscuration (Hydrogen column density, $N_H$) along the line of sight is associated with the failed wind scenario. The variation in mass
accretion is the key that can solely explain the origin of both the failed wind and CL behavior. The mass
accretion behavior in NGC 7582 resembles the hard spectral state of black hole binaries. Furthermore, the observed CL timescales in NGC 7582 can be explained either by the dynamical timescale at the BLR or by viscous and thermal timescales at the boundary layer of the corona for a black hole mass of $5.3^{+0.5}_{-0.3} \times 10^7 M_\odot$. Our study thus highlights a direct physical link between the corona and the BLR, which may be potential drivers of the CL behavior.
Read the full abstract ▾
All authors Santanu Mondal, V. Jithesh, Neeraj Kumari, Tek P. Adhikari
Computer Vision
FoundDSR: A Generalizable Foundation Model with Guided 2D Gaussian Splatting for Depth Super-Resolution
Zhengxue Wang, Zhiqiang Yan, Yuan Wu et al. (+3 more)
We introduce FoundDSR, a generalizable foundation model for robust depth reconstruction across unseen data distributions using RGB-D pairs. FoundDSR begins with a guided 2D Gaussian Splatting strategy to model depth representations with Gaussian primitives. This strategy employs high-resolution RGB as
prompts to optimize the Gaussian parameters, thereby encouraging each Gaussian primitive to anisotropically deform along high-frequency structural directions. The resulting Gaussian-upsampled representations are then mapped to high-resolution depth through an effective depth reconstruction branch. Furthermore, to mitigate training instability and bias toward dominant sources caused by distribution gaps in large-scale heterogeneous data, we introduce heterogeneous
federated learning that allocates each data source to an independent client for local optimization and global aggregation. This design effectively endows FoundDSR with stable scalability to diverse and large-scale training data. Extensive zero-shot evaluations on synthetic, real-world, arbitrary-scale, and noisy conditions demonstrate that FoundDSR consistently outperforms existing state-of-the-art approaches, confirming its strong robustness and generalization to unknown scenes.
Read the full abstract ▾
All authors Zhengxue Wang, Zhiqiang Yan, Yuan Wu, Guangwei Gao, Xiang Li, Jian Yang
Machine Learning
Not All Errors Matter: Decision-Relevant Prediction Error Predicts Planning Quality
Linhao Wang, Yiyan Fan and Dongjin Huang
World models are typically trained and evaluated by prediction error, assuming that more accurate predictions lead to better decisions. We show that this assumption can fail because models with similar total error can differ substantially in planning performance when their errors occur on different state dimensions. We introduce Decision-Relevant Prediction Error (DRPE), which measures prediction error on the state dimensions that affect decisions. We also develop an iso-error evaluation protocol that varies error allocation while keeping total error fixed. In a factored gridworld with known state relevance and a standardized planner, we evaluate 55 controlled and learned models across different error levels and allocations. Total prediction error is weakly related to planning success (Spearman $\rho=-0.25$), whereas DRPE is strongly predictive ($\rho=-0.84$; $-0.98$ within the controlled family). Models with only a 1% difference in total error can differ by 60 percentage points in planning success (97% vs 37%). The relevant error also depends on the task, with model rankings reversing across tasks at the same total error. Deeper imagination further amplifies decision-relevant errors, while learned models exhibit systematic bias on rare but decision-critical events. We formalize sufficient conditions under which DRPE correctly ranks models and total prediction error cannot.
Read the full abstract ▾
Computational Engineering
Stabilizing Autoregressive PDE Foundation Model Rollouts with Event-Triggered Context Healing
Jaewan Park, Jay Phil Yoo, Kazuma Kobayashi et al. (+5 more)
Autoregressive PDE foundation models enable fast full-field
forecasting but accumulate errors as predicted states are recursively reused. We introduce Event-Triggered Context Healing (ETCH), which compares forecasts with measurements from 12 cylinder-surface pressure taps. When their discrepancy exceeds a threshold, pressure-conditioned hourglass
diffusion transformers reconstruct the velocity–pressure context for subsequent predictions. On 43 RealPDEBench Cylinder trajectories, ETCH reduces velocity and pressure errors from 48.05 and 63.71% to 6.29 and 7.25% for DPOT, and from 74.97 and 76.91% to 5.16 and 6.84% for Poseidon-T, using the same reconstructors and threshold. For DPOT, correcting 14.56% of windows maintains errors close to reconstructing every window and supports average real-time
forecasting and correction on an A100 GPU. These results show that sparse physical feedback stabilizes long autoregressive rollouts across PDE foundation models while reducing reconstruction cost.
Read the full abstract ▾
All authors Jaewan Park, Jay Phil Yoo, Kazuma Kobayashi, Seid Koric, Syed Bahauddin Alam, Iwona Jasiuk, Souvik Chakraborty, Diab W. Abueidda
Machine Learning
Superposed Inference for Hyperdimensional Computing
Quanling Zhao, Nilesh Prasad Pandey, Ye Tian et al. (+1 more)
Hyperdimensional computing (HDC) is attractive for efficient and robust learning, but conventional inference still encodes every query independently, repeatedly paying the cost of high-dimensional projection. We introduce SupHDC, a new inference paradigm that processes multiple queries through a shared encoding computation. SupHDC assigns lightweight random slot keys, superposes the keyed queries before encoding, and uses slot-specific classifiers to recover their individual predictions. A random-feature kernel view explains why exact recovery of each hypervector is unnecessary: inference only needs to preserve the class evidence that determines the prediction. Across ten
datasets, SupHDC achieves 1.39x analytical speedup with no average accuracy loss, and up to 2.08x speedup with only a 2.67 percentage-point mean accuracy loss. On a Raspberry Pi~5, it delivers 2.01x measured wall-clock speedup with a 2.26 percentage-point loss in mean prediction accuracy. SupHDC shows that high-dimensional redundancy can be used not only for robustness, but also as capacity for shared inference.
Read the full abstract ▾
All authors Quanling Zhao, Nilesh Prasad Pandey, Ye Tian, Tajana Rosing
Machine Learning
Phase Space Attention:A Hairer Lift Circumvents the Single-Layer Induction Obstruction
Kingsuk Maitra, Shagun Sood Morteza Hosseini, Suman Gunnala et al. (+1 more)
We circumvent the Sanford-Hsu-Telgarsky (SHT) single-layer induction obstruction within a linear, one-step, causal, bilinear, symplectically consistent design class on the post-RoPE substrate, by lifting
attention onto a
symplectic phase space, mirroring Hairer's lift of Stormer-Verlet. The lift exits the premise of the SHT counting argument rather than the bound itself. The reframing exhibits the obstruction as a filter-order gap: a one-layer bilinear score realises a $z$-transform of joint order $(0,0)$, whereas the induction discriminator requires key-side order $\geq 1$. Applying the
symplectic upper shear $M_\gamma:(q,p)\mapsto(q+\gamma p,p)$ to the post-RoPE query and key streams closes it. We prove this lift is unique within the factorised subclass, exactly
symplectic at operator level, and requires post-RoPE placement; and in an explicit $T_4$-only Gaussian reduction we derive a closed-form two-branch induction phase transition, held out at $r=0.9876$ with zero fitted parameters. That law is an analytically solvable limit, not a robust prediction: restoring the $T_3$ channel moves $\gamma_c$ at $d_k=64$ from 1.030 to 0.569 and removes the crossover. Deployability follows by exact derivation: the KV cache is unchanged, prefix reuse and speculative decoding are preserved, overhead is $6d$ FLOPs per
token per layer, INT8 headroom grows by at most $\log_2(1+2\gamma)$ bits, fused kernels are unmodified, and no parameters are added. At 91.3M parameters a supercritical sweep locates an emergence band: induction forms 3/3 seeds at $\gamma=0.80$ in a mean of 717 steps, against 2/3 seeds and 2700 steps at $\gamma=0$. Adverse results are reported as directly: a key-only half-lift reaches 0.949 against 0.811 for the symmetric operator, so if induction accuracy is the objective, the half-lift is the better construction. Forty-one notebooks and result files ship as ancillary material.
Read the full abstract ▾
All authors Kingsuk Maitra, Shagun Sood Morteza Hosseini, Suman Gunnala, Vikram Gupta
Machine Learning
What Can a Leaderboard Certify? Compositional Controllability for Fair Evaluation and Training of Biomedical Literature-Review Agents
Zhaowei Han, Xiang Zhang, Lingxiao Guan et al. (+4 more)
Leaderboards rank long-horizon
agents by their final outputs. Yet a higher score alone does not establish whether two systems are comparable or which stage accounts for the difference. Unequal evidence, inputs, or budgets can affect scores, and statistical corrections do not remove this mismatch. We introduce compositional controllability to address these questions. A comparison window covers one stage, several stages, or the whole
agent. Our central result bounds the gap between observed and controlled score differences using only nuisance outside the window. This yields an admissibility test applied before scores are inspected. Inadmissible comparisons are refused. For admissible pairs, an ordering is certified only when the score gap exceeds the combined sampling and nuisance radii; otherwise, it remains undecided. These decisions give each system a rank interval. We introduce BioLitBench, a
benchmark of 2,042 biomedical articles represented as structured claim graphs. Among seven published pipelines, a conventional statistical analysis declares a winner in 14 of 21 pairwise comparisons. Yet the top-ranked system alone received the target review's bibliography. To isolate pipeline performance, our test requires matched inputs and a fixed backbone model. It refuses 11 of the 21 comparisons, including every comparison involving the top-ranked system. Seven of the 14 conventional conclusions fall within these refused pairs. The same comparison windows support stage-level training. We train SCRIBE on Qwen3.8-27B using rewards measured at each stage's exit. Under matched evidence, SCRIBE achieves a certified rank interval of [1,2], with certified advantages over all evaluated published pipelines and the evaluated Claude and OpenAI
agents. Under same pool, SCRIBE matches the strongest published retriever and is certified above three published pipelines.
Read the full abstract ▾
All authors Zhaowei Han, Xiang Zhang, Lingxiao Guan, Danqi Hu, Kai Liu, Kevin Chang, Jie Liu
Neural and Evolutionary Computing
All On-Board: Fully On-Chip Neuromorphic Q-Learning with Embedded CartPole Simulation
Steven C. Nesbit, Giovanni T. Michel, Gerd J. Kunde et al. (+2 more)
As AI models grow in size and usage, their energy demands increase dramatically, raising sustainability and economic concerns. Neuromorphic hardware, inspired by the energy efficiency of the
brain, seeks to address this challenge by offering low-power, fast-processing alternatives to conventional computing. Such hardware is particularly well-suited to control systems deployed in resource-constrained environments, which are best trained via reinforcement learning (RL). This contribution presents the design and implementation of a fully on-chip, closed-loop Loihi 2 RL
agent. Our neuromorphic
circuit consists of a fully embedded Q-learning
algorithm and an on-chip simulation of the CartPole-v0 environment on Loihi 2. Our Q-learning
algorithm trained the same number of successful
agents as the CPU implementation in only half the execution time and with two orders of magnitude less dynamic power. These findings demonstrate the viability of RL on neuromorphic hardware and highlight its promise for building energy-efficient, real-time, embedded AI systems.
Read the full abstract ▾
All authors Steven C. Nesbit, Giovanni T. Michel, Gerd J. Kunde, Edward Kim, Andrew T. Sornborger
Computer Vision
One Perception, All Maneuvers: Directional Traffic Signal Understanding for Maneuver-Level Signal Intent Prediction
Ang Zou, Runzhe Zheng, Zhigang li et al. (+5 more)
Traffic lights are a key regulatory signal for autonomous driving at urban intersections, yet existing traffic signal perception is still predominantly formulated as instance-level detection or color recognition. Such formulations identify where traffic lights are and what colors they display, but leave a critical semantic gap before downstream planning: which ego maneuver is controlled by each visible signal and what dynamic permission the signal expresses for that maneuver. In this paper, we formulate Directional Traffic Signal Understanding, a decision-oriented task that predicts structured signal states for straight, left-turn, right-turn, and U-turn maneuvers from a front-view image. Each state contains the associated signal color and signal-implied passability. Based on OpenLane-V2, we provide a direction-level
benchmark with maneuver-level supervision and metrics for color recognition, passability, full-frame consistency, and safety-critical errors. A direction-aware baseline combines global context, localized traffic-light evidence, and maneuver-specific representations. Experiments show that direction-level modeling improves passability prediction over image-level classifiers and detection-oriented pipelines, particularly at complex multi-signal intersections. The resulting representation provides a direct and interpretable traffic-signal interface for downstream planning together with
topology, route, and surrounding-
agent information.
Read the full abstract ▾
All authors Ang Zou, Runzhe Zheng, Zhigang li, Zhen Yang, Han Xia, Xuewei Li, Zequn Qin, Xi Li