arXiv subjects

Lu Li

Publications and source records attributed to Lu Li.

At least 19 recordsLinked to original sources

arXiv2026

Automated, auditable dating of the Babylonian Astronomical Diaries with language models and astronomical tools

The Babylonian Astronomical Diaries record the sky over nearly six centuries. Many tablets are broken and have lost their date formula. Experts dated many of them one tablet at a time in the standard edition, Astronomical Diaries and Related Texts from Babylonia (ADART). They used the recorded sky and other evidence, such as year formulas and royal names. Our system automates the astronomical part for all tablets with the same rules. A rule-based parser turns the recorded observations into clauses, short statements with fixed fields. A language model checks them against the text and adds observations that the parser missed. Astronomical tools test the clauses in all 701 years from 626 BCE to AD 75. The system ranks the years by how well they agree with the clauses. The system withholds published dates until its results are final, and the output lists the passages and computed positions behind each candidate year. On 120 dated tablets, the years tied at the top of the ranking include the published year in 96% of cases. For 86 of them, this top group is a single year, the published year in 84. In 119 of the 120, the reported candidate set contains it. For 56 undated entries of ADART IV, the system gives ranked candidate years. For each tablet, the results show how far astronomy alone narrows the year. An interactive page at https://tingyuansen.github.io/Babylonian_Dating/ follows each tablet and links the code, protocol and results.

astro-ph.IM↗

arXiv2026

FactReview: Evidence-Grounded Peer Review with Execution-Based Claim Verification

AI is accelerating scientific research, but providing timely and trustworthy feedback remains challenging. Such feedback requires evaluating specific claims against verified evidence, helping researchers and autonomous research systems determine which findings can guide subsequent work. Thorough verification is time-consuming, and reviewers under time pressure may overlook or misjudge critical issues. Existing AI reviewers have yet to fill this gap because they often assess claims without adequately verifying the supporting evidence. We introduce FactReview, a framework that verifies claims in papers to provide more trustworthy feedback. It checks claims against the literature and examines derivations, while also executing the paper's code to test empirical claims. Across 300 papers, FactReview ranks best among all evaluated systems on five of seven metrics, achieving 6.8% false support and 79.2% classification accuracy. In a study of 36 papers, reviewers using FactReview detect 9.4 issues per paper compared with 5.6 independently, while spending 74.1% less time. FactReview thus supports more reliable peer review, with the potential to accelerate research iteration. The code is public at https://github.com/DEFENSE-SEU/FactReview.

cs.AI↗

arXiv2026

Scaling Latent Reasoning via Looped Language Models

Modern LLMs are trained to "think" primarily via explicit text generation, such as chain-of-thought (CoT), which defers reasoning to post-training and under-leverages pre-training data. We present and open-source Ouro, named after the recursive Ouroboros, a family of pre-trained Looped Language Models (LoopLM) that instead build reasoning into the pre-training phase through (i) iterative computation in latent space, (ii) an entropy-regularized objective for learned depth allocation, and (iii) scaling to 7.7T tokens. Ouro 1.4B and 2.6B models enjoy superior performance that match the results of up to 12B SOTA LLMs across a wide range of benchmarks. Through controlled experiments, we show this advantage stems not from increased knowledge capacity, but from superior knowledge manipulation capabilities. We also show that LoopLM yields reasoning traces more aligned with final outputs than explicit CoT. We hope our results show the potential of LoopLM as a novel scaling direction in the reasoning era. Our model is available here: http://ouro-llm.github.io.

cs.CL↗

arXiv2026

From Observation to Intervention: Mastering Bus Holding Control from Historical Uncontrolled Trajectories with Offline Reinforcement Learning

Bus operations are inherently unstable---buses initially dispatched at equal headways will inevitably end up bunching up. This issue exacerbates passenger waiting times, creates load imbalances between buses, and undermines service reliability. Reinforcement learning (RL) has proven highly effective for mitigating bus bunching and outperforms traditional model-based approaches. However, existing RL approaches require agents to continuously interact with the environment, which potentially disrupts the bus system before effective policies are learned. This paper proposes an Offline Graph-based Multi-agent Actor-Critic (Offline-GMAC) method that trains bus holding policies using only readily available logged GPS data collected during normal uncontrolled operations, thereby eliminating the need for disruptive online interaction or exploration. The method employs general graph neural networks for multi-agent credit assignment. A key innovation of our method is a data augmentation scheme that synthesizes unperformed holding actions from historical uncontrolled bus trajectories. The scheme exploits the equivalence between stochastic link travel time delays and deliberate holding interventions. Under mild assumptions, we prove that the 1-Wasserstein distance between synthesized and target state transition distributions is bounded above by a term that scales linearly with the standard deviation of link travel times plus a tunable synthesis-dependent term. Evaluations using real-world bus data demonstrate that the proposed offline method achieves performance comparable to, and in some cases, even better than its online counterpart.

econ.GN↗

arXiv2026

A unified framework of energy-stable splitting exponential integrators for damped Hamiltonian systems

In this work, we study long-time numerical integration of Hamiltonian systems subject to linear perturbations. We assume the energy-compatible linear-damping condition, which ensures that the physical energy is nonincreasing along the linear damping subflow. Since the conservative and dissipative effects cannot always be merged into a single gradient-structured dissipation and classical energy-stable methods developed for gradient flows cannot directly extend to this setting, we propose a unified framework of two efficient and energy-stable splitting exponential integrators (SEI) to separately handle the dissipative and conservative parts: SEISAV (SEI based on the scalar auxiliary variable) and SEILM (SEI based on Lagrange multiplier). The SEISAV scheme composes the exact damping subflow with an exponential integrator based on the SAV update for the Hamiltonian subflow and requires solving only a scalar linear equation at each time step. We prove a discrete modified-energy dissipation law for SEISAV that mirrors the continuous energy-dissipation mechanism. To enforce decay of the original energy rather than a modified one, we further develop SEILM by incorporating a Lagrange-multiplier formulation within the splitting exponential framework, leading to solving only a scalar nonlinear equation at each time step. Numerical experiments on representative linearly damped Hamiltonian models confirm the designed second-order accuracy, energy stability, and competitive efficiency relative to state-of-the-art schemes, indicating that the proposed framework provides a simple and robust approach for simulating linearly perturbed Hamiltonian dynamics.

math.NA↗

arXiv2026

Point It, Strike It: Direction-Conditioned Dynamic Manipulation of Deformable Linear Objects

Goal-conditioned dynamic manipulation of deformable linear objects has mainly specified goals as positions for a rope tip to reach. Many tasks, however, depend on how the tip arrives. We therefore study single-swing rope striking with goals that specify the tip's 3D position and arrival direction, across the workspace and on different ropes. This is challenging because rope dynamics are hard to model, no demonstrations exist, distinct swings reach the same goal with different reliability, and the sim-to-real gap extends beyond the rope. To address these challenges, we extend the state-of-the-art DLO simulator DeformX with GPU acceleration, a stable Cosserat rod solver, and a cross-flow aerodynamic model, yielding DeformX2.0, which is more than $20{,}000\times$ faster. We then propose TRACE (Trace-rooted Adaptive Cross-Entropy), which generates striking data by warm-starting each new target from the stored swing whose tip path passes closest to it. Its cost penalizes rope bending and abrupt tip motion to favor repeatable swings. A conditional flow-matching policy trained on this data reaches 92.1% accuracy in simulation. Finally, we propose RECAP (Residual Calibration Policy), which fits the simulator's rope and rig parameters to a few calibration swings and adapts actions with a correction policy trained in simulation. On a real robot, across three ropes, RECAP raises success within 5cm from 72% to 87% for position goals, and within 10cm and 10° from 50% to 79% for goals that also specify the arrival direction.

cs.RO↗

arXiv2026

Beyond Isolation: Unlocking Reinforcement Learning Component Synergy for Sample-Efficient Continuous Control

Reinforcement learning systems are significantly more complex than other machine learning paradigms due to inherent properties, causing RL system design to jointly account for many tightly coupled factors. Despite advances in individual algorithmic components, their functional interdependencies remain underexplored: do they exhibit mutual synergy or counterproductive interference? To bridge this gap, we conduct a systematic investigation and find that the efficacy of different components exhibits significant task-dependency, and naively stacking state-of-the-art techniques does not necessarily yield performance gains; instead, it often triggers emergent challenges, such as compounded non-stationarity. Building upon these findings, we distill a suite of actionable insights into the principled coordination of these components. Guided by these insights, we propose ROSER, an RL framework that coordinates three critical dimensions: Model-based Representation, Optimization Stability, and Experience Replay. Across diverse continuous-control benchmarks, ROSER consistently outperforms vanilla baselines and achieves 17.60% gains over naive stack. Our findings underscore the necessity of a holistic perspective in RL system design and paves the way for developing sample-efficient agents.

cs.LG↗

arXiv2026

PIONEER: Bayesian Joint Modelling of Mechanistic Tumour Growth and Time-to-Event Endpoints for Dynamic Prediction of Ongoing Oncology Trials

High-stakes decisions in oncology clinical trials must often be made while survival data remains immature: progression-free survival (PFS) and overall survival (OS) are heavily censored, few events have accumulated, and the primary endpoint may be months or years from reading out. What is available at interim data cut-offs is information-rich longitudinal tumour measurements and baseline covariates. We present PIONEER, a Bayesian joint modelling framework that couples a mechanistic two-component state-space submodel of longitudinal tumour size dynamics to a multistate proportional-hazard submodel for competing clinical events, fitted simultaneously under a single posterior. The mechanistic submodel infers latent per-patient tumour trajectories - decomposed into treatment-responsive and refractory compartments with Gompertz-attenuated growth - from sparse, noisy sum-of-longest-diameter (SLD) observations. These latent trajectories feed the multistate hazard as time-varying covariates, while the event data simultaneously refines the tumour dynamics through the joint likelihood. All clinical endpoints (PFS, OS, objective response rate) are derived from the joint posterior in a single forward simulation pass, propagating full parameter uncertainty without any two-stage plug-in. Applied to a case study in extensive-stage small-cell lung cancer (two trials, N = 497), leave-future-out cross-validation demonstrates that at month 4 of enrolment (9 patients) the model produces calibrated PFS forecasts covering the mature month-19 Kaplan-Meier curve, and at month 11 (39 patients) the OS forecast converges - representing at least 8 months of advance forecasting with properly quantified uncertainty. We hope this work paves the way for broader adoption of Bayesian mechanistic state-space frameworks in clinical development, enabling earlier and more informed decision-making from immature trial data.

stat.AP↗

arXiv2026

Experimental investigation of altermagnetic order in Cr-doped FeSb2

Altermagnets are a class of materials with compensated magnetic moments, in which spin sublattices are related by specific rotational symmetries other than inversion or translation. This allows time-reversal symmetry to be broken without a net magnetization. Cr-doped FeSb2 has been theoretically proposed as a candidate d-wave altermagnetic system, yet its magnetic ground state has remained unresolved. Here, we synthesize single crystals of Fe1-xCrxSb2 and investigate their electrical transport and magnetic properties, with a focus on Fe0.85Cr0.15Sb2. Magnetization measurements suggest spin-compensated ordering below ~3.5 K, where magnetic moments align along the crystallographic b-direction. Transport measurements reveal a crossover from large positive to negative magnetoresistance, while an anomalous Hall response emerges below 5 K, indicating time-reversal symmetry breaking. Muon spin relaxation measurements confirm that the magnetic ordering below 3.5 K is bulk in nature. The absence of coherent oscillations in zero-field μSR spectra and of magnetic Bragg intensity in single-crystal neutron diffraction establishes that the magnetically ordered state is short-range or disordered, rather than collinear altermagnetic order. These results demonstrate that Cr-doping alone breaks time-reversal symmetry without stabilizing long-range altermagnetic order in FeSb2.

cond-mat.mtrl-sci↗

arXiv2026

The Three Regimes of Offline-to-Online Reinforcement Learning

Offline-to-online reinforcement learning (RL) has emerged as a practical paradigm that leverages offline datasets for pretraining and online interactions for fine-tuning. However, its empirical behavior is highly inconsistent: design choices of online fine-tuning that work well in one setting can fail completely in another. Guided by the stability--plasticity principle, we propose a framework that can explain this inconsistency: We argue that efficient fine-tuning must preserve the utility of the stronger offline prior, whether that is the pretrained policy or the offline dataset, while maintaining sufficient plasticity. This perspective identifies three regimes of online fine-tuning, each requiring distinct stability properties. We validate this framework through a large-scale empirical study, finding that the results strongly align with its predictions in 45 out of 63 cases, with only 3 opposite mismatches. This work provides a framework for guiding design choices in offline-to-online RL based on the relative performance of the offline dataset and the pretrained policy.

cs.LG↗

arXiv2026

CSympNet-ID: conformal-symplectic map learning for linearly damped Hamiltonian systems

Learning dissipative dynamics from discrete observations is essential for reliable long-horizon prediction and physically meaningful parameter identification. For linearly damped Hamiltonian systems, the exact flow is generally not symplectic but conformally symplectic, contracting the canonical symplectic form by a scalar factor that reflects the net dissipation. We propose Conformal Symplectic Networks with damping identification (CSympNet-ID), a discrete-time map-learning framework that learns the one-step flow map directly from snapshot pairs while enforcing exact discrete conformal symplecticity by construction, without penalty terms or projection. The architecture composes an exact symplectic neural core with explicit diagonal scaling layers whose factors are parameterized exponentially by a scalar damping-rate parameter, thereby guaranteeing positivity and interpretability of the learned dissipation factor. We establish a scaling-conjugacy factorization for conformal symplectic maps and derive a pointwise-in-step density result for CSympNet-ID. We evaluate an irregular-step damped oscillator, a damped spring-mass chain, a damped nonlinear cubic oscillator, and additional high-dimensional extensions. CSympNet-ID gives the most favorable overall results among the compared models in the reported experiments, particularly in data-scarce regimes, target contraction-law recovery, and high-dimensional tests where unstructured baselines degrade rapidly.

cs.LG↗

arXiv2026

SVD-Based Graph Fractional Fourier Transform on Directed Graphs and Its Application

Real-world signals frequently reside on directed Cartesian product graphs, including digital images, sensor networks, and meteorological temperature records. Designing a transform method suitable for processing such multi-dimensional graph signals within the fractional Fourier transform domain remains a critical challenge in graph signal processing (GSP). This paper proposes two novel graph fractional Fourier transforms (GFRFTs) for multi-dimensional signals defined on such directed product graphs and comprehensively investigates their denoising capabilities. Our contributions are fourfold: (1) We propose two distinct two-dimensional GFRFTs based on singular value decompositions of some fractional Laplacian matrices; (2) We generalize these transforms to multi-dimensional graph fractional Fourier transforms (MGFRFTs), establishing a powerful fractional domain analysis framework for multi-dimensional GSP; (3) We investigate the signal reconstruction capability of our proposed GFRFTs, as well as their computational complexity; and (4) We validate the practical utility of our approach through denoising experiments on real-world meteorological temperature datasets.

eess.SP↗

arXiv2026

Beyond One-Size-Fits-All: Diagnosis-Driven Online Reinforcement Learning with Offline Priors

Online reinforcement learning (RL) agents increasingly depend on knowledge acquired offline to achieve practical efficiency. Originally studied in offline-to-online RL, this paradigm now spans foundation model post-training and embodied intelligence, with prior types expanding from offline datasets and pre-trained policies to increasingly diverse knowledge sources such as multimodal foundation models and generative world models. Offline priors have become central to how deep RL is developed and deployed. However, this reliance introduces a challenge that the prevailing benchmark-driven paradigm cannot resolve: because prior validity varies across deployments and shifts during training, no single approach to managing it is universally optimal, and benchmark rankings offer limited guidance for real-world deployments. Rather than pursuing universal solutions, we argue that the field should shift to diagnosis-driven tension management, in which deployment-specific evidence guides how the learner relates to its priors throughout training, enabling both flexible and adaptive deployment. We support this position with a framework characterizing how priors reshape online optimization through three functional roles, controlled experiments demonstrating help-or-hurt reversals, cross-domain evidence from foundation model post-training to embodied intelligence, and engagement with five substantive counterarguments.

cs.LG↗

arXiv2026

DeformX: A Versatile Co-Simulation Framework for Deformable Linear Objects

Deformable linear objects (DLOs) such as wires, cables, and ropes are common in robotic manipulation tasks, yet simulating them with both visual realism and physical accuracy remains challenging. Existing visual simulation methods typically rely on procedural geometric primitives that lack physically grounded deformation behavior, while physics-based approaches with robot learning support often approximate DLOs as rigid-link chains or generic soft bodies, failing to accurately capture the bending, twisting, and shear mechanics of slender elastic structures. In this work, we introduce DeformX, a co-simulation framework that integrates a dedicated Cosserat rod physics engine with NVIDIA Isaac Sim, enabling DLO simulations that are both physically faithful and visually realistic. Our Cosserat rod engine simulates the dynamics and self-collisions of DLOs, and contact interactions with arbitrary free-form meshes. To achieve high-fidelity visualization, we employ mesh skinning to map discrete rod deformations onto imported CAD models. To the best of our knowledge, DeformX is the one of the first frameworks for DLO simulation that unifies realistic visualization, principled physics, and compatibility with robot learning pipelines. We demonstrate its versatility across synthetic data generation and policy learning for DLO manipulation, and validate visual and physical fidelity through comparisons against real-world experiments. Notably, fine-tuning Segment Anything Model 3 (SAM3) on DeformX-generated data yields a 10.2% mAP@75 improvement in real-image wire segmentation, and a rope-swinging policy trained entirely in DeformX achieves a mean target-hitting error of 6.6 cm on a UR5e manipulator in real-world trials, highlighting its strong sim-to-real transfer capability.

cs.RO↗

arXiv2026

Representation Learning Enables Scalable Multitask Deep Reinforcement Learning

Scaling reinforcement learning (RL) to diverse multitask settings remains a central challenge. While recent advances in model-based RL achieve strong performance, they rely on planning and complex training pipelines, making it unclear which components are essential for scalability. We revisit this question and argue that the primary driver of scalable multitask RL is not model-based control, but \emph{representation learning}. In particular, we show that combining predictive, model-based representations with high-capacity value function approximation is sufficient to achieve strong performance, even without planning. We evaluate a simple model-free algorithm, MR.Q, coupled with auxiliary predictive objectives into a scalable actor-critic architecture. This approach outperforms a recent world-model-based method and a range of deep RL baselines across a diverse suite of multitask continuous control tasks, while significantly reducing computational overhead and improving wall-clock efficiency. We observe consistent improvements with increased model capacity and show through ablations that predictive representation learning is critical for performance.

cs.LG↗

arXiv2026

Evolution of the stellar mass function in open clusters from a universal and unsegregated initial state

The stellar mass function (MF) and its spatial variation (mass segregation) within star clusters encode signatures of early formation physics and subsequent secular evolution. Yet, a coherent evolutionary picture remains elusive due to conflicting reports regarding the universality of the initial mass function (IMF) and the prevalence of primordial mass segregation. These discrepancies often arise from unresolved binaries, field contamination, and completeness bias. Here, we resolve these issues by analyzing 163 high-fidelity open clusters via a Bayesian forward-modeling framework. We reveal a remarkably simple initial state: young clusters ($\lesssim 300$ Myr) share a mean IMF slope of $-2.29$ in the mass range $M \geq 0.5 M_\odot$, consistent with the Salpeter slope but with an intrinsic scatter of 0.17, and exhibit minimal mass segregation at the onset of gas-free evolution ($\sim$10 Myr). This broadly universal "zero-point" for secular evolution disfavors star-forming scenarios that predict strong primordial segregation or significant IMF variations, and suggests that chaotic cluster assembly and gas expulsion efficiently erase any mild primordial inhomogeneities. By tracing the evolutionary sequence from $10^7$ to $10^{9.8}$ yr, we demonstrate that dynamical processing operates on distinct timescales: mass segregation proceeds rapidly via internal relaxation, whereas global MF flattening due to tidal evaporation becomes dominant only after $\sim$600 Myr. These findings impose robust observational constraints on the physics of star formation and early feedback and establish an empirical baseline for modeling secular stellar dynamics.

astro-ph.GA↗

arXiv2026

When 2D Tasks Meet 1D Serialization: On Serialization Friction in Structured Tasks

In the LLM era, many symbolic and structured problems are presented to models through 1D text serialization. Yet some such problems are natively two-dimensional: their relevant relations, such as row--column correspondence or spatial adjacency, are defined by position in a 2D layout rather than by sequential order. This raises a representational question: does preserving the same symbolic entries in a 1D sequence also preserve the relational structure needed for computation? We study this issue through the lens of serialization friction: the representational mismatch in which the same underlying task instances and entries are still present, but relations that depend on layout become implicit under 1D serialization. The study uses a controlled synthetic testbed of three tasks: matrix transpose, Conway's Game of Life, and LU decomposition. In each task, the same instances are presented either as 1D text serialization or as their native 2D layout rendered as an image. Across this testbed, 1D serialization degrades more sharply as task size grows, and errors under serialization exhibit spatially structured patterns, suggesting that this presentation choice is consequential within our testbed. To further interpret these results, we add supplementary analyses that include a within-visual probe and an additional comparison of the two input presentations under the mixed-training transpose setting. These findings suggest that, for layout-defined tasks, reducing inputs to 1D serialization is not a neutral choice of representation.

cs.CL↗

arXiv2026

Non-local low energy neutral excitations in a strongly disordered triangular Mott magnet Cr$_3$Se$_2$Br$_5$

Understanding if low-energy excitations can remain itinerant in the presence of strong disorder remains a central challenge in frustrated quantum magnets, where disorder is generally expected to localize excitations through Anderson-like mechanisms. Here we report the emergence of charge-neutral itinerant excitations in a van der Waals compound Cr$_3$Se$_2$Br$_5$, a strongly disordered $S = 3/2$ Mott insulator with a frustrated triangular lattice. Structural analysis reveals substantial intrinsic disorder arising from Cr-site deficiency and Se/Br-site mixing, which appear to be fixed and cannot be readily tuned. No long-range magnetic order or conventional glassy behavior is observed. In addition to its highly insulating nature, the magnetic specific heat C_mag/T and thermal conductivity \k{appa}_xx/T both exhibit linear temperature dependencies with substantial finite intercepts. In particular, a sizeable field-independent residual term $κ/T \approx 0.03~\mathrm{W\,m^{-1}\,K^{-2}}$ is observed, providing compelling evidence of itinerant low-energy excitations that carry entropy without charge. These findings conceptually advance our understanding of quantum matter by demonstrating a rare regime where the interplay of disorder, frustration, and electronic correlations actively reshapes the nature of low-energy excitations, allowing itinerant neutral excitations to coexist with strong intrinsic disorder.

cond-mat.str-el↗