arXiv subjects

Rui Zhu

Publications and source records attributed to Rui Zhu.

At least 19 recordsLinked to original sources

arXiv2026

Beyond Verified Answers: Solver-Informed Self-Distillation for Bootstrapping Operations Research Language Models

Modern large language models (LLMs) can translate natural-language descriptions into operations research (OR) formulations. Post-training techniques including reinforcement learning and on-policy self-distillation have further improved this capability. However, three limitations remain in training LLMs for OR formulations. First, training commonly relies on synthetic formulations validated by human experts or stronger models, constraining scalable supervision. Second, credit assignment is either coarse or costly: outcome rewards score an entire trajectory without locating the responsible modeling decision, whereas process-level supervision requires an additional evaluator. Third, privileged self-distillation can induce style mismatch by using solver context unavailable at deployment. We find that a model can improve from solver-artifact feedback generated by its own rollouts, making self-distillation a practical, evaluator-free source of dense supervision. Therefore, we propose SOLID: Solver-Informed On-Policy LearnIng through Self-Distillation, a novel framework for self-improving OR language models without verified answers or external evaluators. SOLID executes candidate programs from multiple rollouts, clusters their objectives, and selects a majority-group artifact as a pseudo-reference. The model then performs updates using group-relative advantages and dense self-supervision signals. Across multiple OR benchmarks, SOLID improves solution accuracy for both general-purpose and OR-tuned models over outcome-only group-relative training. These results show that solver artifacts can support scalable self-improvement without trusted answers.

math.OC↗

arXiv2026

GenFirst: Generation Before Reconstruction for Stable End-to-End Latent Generative Modeling

Latent generative models typically follow a two-stage pipeline, training a variational autoencoder for reconstruction and then a generative model on the frozen latent space. Since reconstruction-optimized latents are not necessarily generation-friendly, jointly training both models is an appealing alternative. However, direct end-to-end training remains challenging, as it is prone to latent collapse and faces a generation-reconstruction conflict. We revisit this problem by analyzing how different objectives shape the latent space and identify two key insights. First, the entropy term in the Kullback-Leibler divergence objective is essential for preventing collapse: reconstruction and prior fitting tend to shrink the posterior, while entropy preserves non-degenerate latent uncertainty. Second, reconstruction and generation exhibit asymmetric learning dynamics: reconstruction is fast and strongly supervised, whereas generation is slower and harder to optimize. Based on these insights, we achieve the first direct end-to-end training without latent collapse and propose GenFirst, a simple generation-before-reconstruction strategy. The generative objective first shapes the latent space under weak reconstruction pressure, after which reconstruction is progressively strengthened to recover visual details. We validate GenFirst with continuous autoregressive priors with exact likelihoods and SiT priors with implicit likelihoods. With our end-to-end objective and GenFirst, SiT achieves a gFID of 0.97 with CFG and 1.45 without CFG on ImageNet-256, while MMDiT reaches a GenEval score of 0.90 on text-to-image generation. Beyond image generation, we extend the framework to shared visual latents for generation and representation learning, and to continuous unified text-image generation. These results demonstrate the generality of stable end-to-end latent learning across generative priors and modalities.

cs.CV↗

arXiv2026

Gradient-extrapolation-based distributed mirror descent algorithm for multi-cluster aggregative games

This paper studies a class of multi-cluster aggregative games characterized by the coexistence of cooperation and competition, where each agent's cost function depends on its own strategy and the aggregate of all agents' strategies. To address the Nash equilibrium seeking problem for such games in the non-Euclidean setting, a distributed mirror descent algorithm with gradient extrapolation is proposed over time-varying intra-cluster and inter-cluster networks. The mirror descent framework employs a general Bregman divergence as the distance measure, providing greater flexibility than Euclidean-based methods, while gradient extrapolation exploits historical gradient information to improve convergence performance. Under the restricted strong monotonicity characterized by the Bregman divergence, the convergence of the proposed algorithm is established, and it achieves the $\mathcal{O}(1/k)$ convergence rate with the appropriately selected step-size and parameters. Finally, the effectiveness of the proposed algorithm is verified by an example on the demand response of energy systems.

cs.GT↗

arXiv2026

AUCH-Net: Action Unit-Based Consistency-Aware Hypergraph Network for Cross-Domain Few-Shot Facial Expression Recognition

Recently, cross-domain few-shot facial expression recognition (CF-FER) has received considerable attention. However, the performance of existing CF-FER methods is still unsatisfactory due to inferior transferable feature learning under large domain discrepancy and limited target samples. Fortunately, the action units (AUs), which indicate the movements of different facial muscles, provide consistent conceptual semantics for describing expressions within and across domains. Inspired by this, we propose a novel Action Unit-based Consistency-aware Hypergraph Network (AUCH-Net), which constructs consistency-aware hypergraphs on AUs, for CF-FER. Specifically, AUCH-Net presents a new AU feature learning (AFL) module and a new visual feature learning (VFL) module. The AFL module learns AU features under the guidance of a novel relation consistency loss and an AU regularization loss, while the VFL module learns visual features supervised by a relation consistency loss and a classification loss. By learning consistent AU features, AUCH-Net effectively models the connections between AUs and expression categories. As a result, we can bridge the gap between fine-grained facial variations and high-level expression categories, greatly facilitating the learning of transferable feature representations.Extensive experiments on both in-the-lab and in-the-wild datasets show that our method consistently outperforms several state-of-the-art methods. Our results clearly show that modeling the relationships among AUs holds significant potential for FER under cross-domain few-shot scenarios.

cs.CV↗

arXiv2026

Demographically-Informed Heat-Mortality Risk Curves via Risk Graph Neural Networks

Estimating heat-related mortality risk is a core task in environmental epidemiology, typically addressed with Distributed Lag Non-linear Models (DLNMs); interpretable exposure-response surfaces fitted to temperature-mortality time series. DLNMs are effective but ignore demographic and geographic context, despite well-established relevance to heat vulnerability. We propose Risk Graph Neural Networks (RGNNs), a hierarchical GNN encoder that uses granular census features to optimise DLNM coefficient vectors, preserving interpretable risk curve outputs while substantially improving predictive calibration. Evaluated across 10 regions of England and Wales on two unprecedented heat years, RGNN variants maintain both lower point-errors and near-nominal uncertainty coverage during the 2022 heatwave where baselines collapse.

cs.LG↗

arXiv2026

Far-Infrared Star Formation Rates of Quasar Host Galaxies from Multiwavelength Spectral Energy Distribution Decomposition

Reliable star formation rates (SFRs) are essential for studying the connection between black hole growth and quasar host galaxies. We study the far-infrared (FIR) SFRs and the host galaxy properties of 202 SDSS and PG quasars at $0.02<z\lesssim0.8$, spanning $\log({\rm SFR}_{\rm FIR}/M_\odot\,{\rm yr}^{-1})\simeq-0.45$--$2.76$, using multiwavelength spectral energy distribution (SED) decomposition. The photometry covers wavelengths from the optical to the FIR and is supplemented by JCMT/SCUBA-2 observations at 450 and 850 $\mu$m. We model the SEDs with CIGALE and AGNfitter and adopt multiple cold dust templates to quantify systematic uncertainties. The median model-dependent scatter among the five FIR SFR estimates is $0.14$ dex, and AGNfitter gives FIR SFRs lower than the mean CIGALE estimate by a median of $0.09$ dex. For the 58 quasars with SCUBA-2 coverage, including SCUBA-2 data changes the adopted FIR SFR by only $\sim$0.01 dex on average but can affect individual sources with limited Herschel coverage or radio-loud emission. Within our FIR-constrained sample, many quasar hosts lie on or above the star-forming main sequence, but the redshift-dependent FIR selection of the SDSS subsample limits conclusions about the full quasar-host population. We find no clear correlation between the main-sequence (MS) offset and the direct Eddington ratio, while the offset is positively related to the infrared-based $L_{\rm tor}/L_{\rm Edd}$ proxy. The minimum radiation field intensity in the dust model, $U_{\rm min}$, increases with bolometric luminosity and dust temperature. WISE W2 (4.6 $\mu$m) and W3 (12 $\mu$m) combined with Herschel bands can also provide useful empirical indicators of $f_{\rm AGN}$.

astro-ph.GA↗

arXiv2026

Testing [O II] $\lambda3727$ as a Star Formation Rate Tracer in Quasar Host Galaxies

The [O II] $\lambda3727$ emission line is a widely used star formation rate (SFR) tracer. However, its application to type I quasars is not straightforward, because the line can be affected by dust extinction, metallicity and contamination from the AGN narrow-line region (NLR). We test the reliability of [O II] SFRs using a sample of 202 SDSS and PG quasars, by comparing [O II] SFRs and reference far-infrared (FIR) SFRs derived from multiwavelength SED decomposition. We measure [O II], [O III], and narrow Balmer emission lines by spectral fitting. Then, we calculate [O II] SFRs after correcting dust extinction and metallicity. We then compare these SFRs with the FIR SFRs, both with and without subtracting the AGN contribution estimated from [O III]. After this correction, the median offset between [O II] and FIR SFRs is $-0.20\pm0.72$ dex for the full analysis sample and $-0.17\pm0.69$ dex for sources with S/N $>5$ in both [O II] and [O III]. Without subtracting the AGN contribution, the corresponding offsets are $0.00\pm0.69$ and $0.12\pm0.66$ dex. We conclude that [O II] is useful as a statistical SFR tracer for quasar host galaxies, but individual objects still require careful treatment of AGN contamination, extinction, metallicity, aperture effects, and redshift-dependent systematics.

astro-ph.GA↗

arXiv2026

CRISP: Pre-LLM Yet Text-Driven Visual Token Pruning for Efficient LVLM Inference

Large Vision-Language Models (LVLMs) typically require processing hundreds to thousands of visual tokens, leading to substantial inference overhead. Existing visual token pruning methods either operate before the LLM using text-agnostic heuristics or prune inside the LLM at the cost of efficiency and noisy cross-modal attention. To address these limitations, we propose CRISP, a pre-LLM yet text-driven visual token pruning framework that preserves both instruction-relevant evidence and essential scene context. CRISP works in a two-stage pipeline: Stage 1 first identifies text-aligned visual tokens, and Stage 2 enhances contextual completeness through semantic diversity. Extensive experiments on LLaVA-1.5 and LLaVA-NeXT demonstrate that CRISP achieves superior performance retention under aggressive pruning ratios, maintaining up to 99.5% accuracy while reducing inference cost and latency by more than 2 times. CRISP serves as a practical solution for efficient LVLM inference, especially in resource-constrained scenarios.

cs.CV↗

arXiv2026

Continuous Query for Top-$K$ Maximal Sum Intervals over Streaming Data

The continuous identification of top-$k$ maximal sum intervals using a sliding window over a data stream is a critical operation for applications in IoT and beyond. A maximal sum interval is a non-overlapping, contiguous subsequence with the maximal sum in a sequence of signed values. Existing algorithms are ill-suited for streaming contexts: they either exhaustively enumerate all intervals even for small $k$ values, or depend on indexes that require frequent and costly restructuring. We propose a novel partition-based strategy. Our core insight is a partitioning scheme that guarantees that any maximal sum interval is fully contained within a single partition, enabling independent and parallel processing. This design provides two key advantages: it enables safe pruning of partitions that cannot contribute to top-$k$ results, drastically narrowing the search space, and it enables efficient, incremental maintenance of the maximal sum intervals in each partition. We develop algorithms for partition construction, incremental partition updates, and partition-based top-$k$ maximal sum interval search. Extensive experiments on real and synthetic datasets demonstrate that our approach significantly improves efficiency.

cs.DB↗

arXiv2026

LeRoPE: Learnable RoPE Frequencies Improve Language Modeling

Rotary Positional Encodings (RoPE) are currently the most popular positional encodings used in modern language models. RoPE rotates two-dimensional chunks of query and key vectors, operating as a function of their relative positional offset. The position-wise rates of rotation in RoPE typically follow a geometric sequence specified by a fixed base-frequency hyperparameter. Prior work has improved performance by either increasing this parameter to slow rotation or by applying RoPE to only a subset of QK dimensions. In this work we modify RoPE by learning a scalar per frequency, treating frequencies as learnable parameters rather than hyperparameters. We validate Learned RoPE by training a ladder of language models from scratch, ranging from 52M to 2.5B parameters. We observe and analyze the emergence of a high-norm, positional LeRoPE band. LeRoPE consistently outperforms RoPE and partial RoPE across all scales, with RoPE requiring 3.4% more compute (FLOPs) to match LeRoPE at the largest scale.

cs.LG↗

arXiv2026

Can Agents Generalize to the Open World? Unveiling the Fragility of Static Training in Tool Use

While Large Language Model (LLM) agents demonstrate proficiency in static benchmarks, their deployment in real-world scenarios is hindered by the dynamic nature of user queries, tool sets, and interaction dynamics. To address this generalization gap, we formalize OpenAgent (Tool-Use Agent in Open-World), a problem setting characterized by distributional shifts across query, action, observation, and domain dimensions. To systematically diagnose its impact, we construct a controlled sandbox environment where we define fine-grained environmental shifts across a four-tier hierarchy, Perception, Interaction, Reasoning, and Internalization, and conduct a comprehensive series of experiments. Our analysis yields a series of key insights, demonstrating that agents trained via both Supervised Fine-Tuning(SFT) and Reinforcement Learning suffer from varying degrees of performance degradation when confronting open environmental shifts. Building on these insights, we propose Perturbation-Augmented Fine-Tuning, a disturbance-based intervention strategy for SFT that lays the foundation for enhancing agent robustness and utility in realistic environments. Our code will be released at: https://github. com/LAMDA-NeSy/OpenAgent.

cs.AI↗

arXiv2026

Continuous Latent Diffusion Language Model

Large language models have achieved remarkable success under the autoregressive paradigm, yet high-quality text generation need not be tied to a fixed left-to-right order. Existing alternatives still struggle to jointly achieve generation efficiency, scalable representation learning, and effective global semantic modeling. We propose Cola DLM, a hierarchical latent diffusion language model that frames text generation through hierarchical information decomposition. Cola DLM first learns a stable text-to-latent mapping with a Text VAE, then models a global semantic prior in continuous latent space with a block-causal DiT, and finally generates text through conditional decoding. From a unified Markov-path perspective, its diffusion process performs latent prior transport rather than token-level observation recovery, thereby separating global semantic organization from local textual realization. This design yields a more flexible non-autoregressive inductive bias, supports semantic compression and prior fitting in continuous space, and naturally extends to other continuous modalities. Through experiments spanning 4 research questions, 8 benchmarks, strictly matched ~2B-parameter autoregressive and LLaDA baselines, and scaling curves up to about 2000 EFLOPs, we identify an effective overall configuration of Cola DLM and verify its strong scaling behavior for text generation. Taken together, the results establish hierarchical continuous latent prior modeling as a principled alternative to strictly token-level language modeling, where generation quality and scaling behavior may better reflect model capability than likelihood, while also suggesting a concrete path toward unified modeling across discrete text and continuous modalities.

cs.CL↗

arXiv2026

How to Compress KV Cache in RL Post-Training? Shadow Mask Distillation for Memory-Efficient Alignment

Reinforcement Learning (RL) has emerged as a crucial paradigm for unlocking the advanced reasoning capabilities of Large Language Models (LLMs), encompassing frameworks like RLHF and RLAIF. Regardless of the specific optimization algorithm (e.g., PPO, GRPO, or Online DPO), online RL inherently requires an exploratory trajectory generation (rollout) phase. However, for long-context reasoning tasks, this rollout phase imposes a severe ``memory wall'' due to the exorbitant Key-Value (KV) cache footprint. While applying KV cache compression during rollouts mitigates this memory overhead, it induces a critical off-policy bias. Although modern KV compression is often nearly lossless during standard inference, even minuscule approximation errors are drastically amplified by the inherent instability of RL optimization. Specifically, the sampler generates responses under a sparse context, whereas the learner updates parameters using the full, dense context. Existing statistical solutions, such as importance reweighting, struggle to correct this magnified bias, suffering from high gradient variance and severe sample inefficiency.

cs.LG↗

arXiv2026

Can LLMs Predict Polymer Physics Just by Reading Synthesis and Processing Prose?

Can large language models predict physical and mechanical polymer properties simply by reading unstructured scientific prose? Polymer performance is rarely determined by chemical structure alone; identical nominal polymers can exhibit drastically different behaviors depending on their synthesis route, processing history, morphology, and testing conditions. Yet, state-of-the-art polymer property models typically rely on structure-only representations -- such as SMILES or molecular graphs -- which strip away this vital experimental context. In this work, we introduce \textbf{PolyLM}, a natural-language-only, process- and condition-aware framework that predicts materials performance directly from full-text literature. By circumventing structural inputs entirely, PolyLM preserves the nuanced, unstructured descriptions of synthesis and processing reported by domain scientists. To train this framework, we curated an unprecedented, literature-scale dataset encompassing 185,000 scientific papers and over 276,400 unique polymer samples across 22 physical, mechanical, and thermal properties. We fine-tuned a massive 9-billion-parameter language model (Qwen3.5-9B) using Low-Rank Adaptation (LoRA) and task-level uncertainty weighting. Evaluated on 68,283 held-out observations, the model achieves remarkably high predictive accuracy, establishing new state-of-the-art benchmarks for complex properties. Across the 22 diverse targets, the model achieves a median $R^2$ of 0.74, with predictions for key thermal, mechanical, and physicochemical properties frequently surpassing an $R^2$ of 0.80. These results unequivocally demonstrate that natural language is a powerful, highly scalable interface for realistic materials performance prediction.

cs.LG↗

arXiv2026

Unveiling the Atomistic Mechanisms of Shear-Induced LDA$\leftrightarrow$HDA Transformations and Shear Banding in Amorphous Silicon under High Pressures

Large-scale molecular dynamics simulations of shear deformation under constant pressures of amorphous silicon, PT from low-density-amorphous (LDA) to high-density-amorphous (HDA) Si, and formation of shear bands (SBs) are performed using the state-of-the-art Gaussian Approximation Potential. The simulations reveal that LDA$\leftrightarrow$HDA shear-induced PTs occur simultaneously until reaching steady state. The developed mechanism-based analytical model well describes shear-strain-governed kinetics and steady states at all pressures, independent of shear stresses. Shear reduces the pressure for initiation and completion of LDA$\rightarrow$HDA PT by $4.36$ and $5.10$ GPa, respectively. Without PT at low pressure, shear-banding occurs, which is partially suppressed by PT at higher pressure with uniform deformation-PT at $9.8$ GPa. Despite the much larger shear and expected fraction of HDA, surprising sharp drop in the HDA atomic fraction within the SB was discovered. In bulk, Si deforms by atomic rearrangement in localized shear transformation zones with high nonaffine displacements, which trigger nucleation of HDA clusters within LDA and, concurrently, of LDA clusters within HDA, without growth and coalescence. In SB, a turbulent-like flow with swirls is revealed, which promotes reverse PT from HDA$\rightarrow$LDA more effectively. Transformation-induced plasticity in amorphous Si is revealed. The findings open up basic research into the mechanisms and kinetics of plastic strain-induced PTs in amorphous materials under high pressure, with numerous important applications.

cond-mat.mtrl-sci↗

arXiv2026

Geographic Bias and Diversity in AI Evaluation

Among the many challenges hindering the responsible development and deployment of AI, arguably none has faced more intense scrutiny than bias in its various forms. This underscores the widespread concerns across AI researchers that model outputs, e.g., from generative AI, may encode structural distributional imbalances (stemming from training data or model design) that may amplify social inequality or introduce systemic distortions across application domains ranging from biodiversity to disaster mitigation. Yet, relatively little work has investigated the geographical nature of bias or developed measurable benchmarks for what it means for (generative) AI to be unbiased. In this chapter, we investigate this issue through a literature review. As foundation models are reshaping the landscape of bias research, we examine work spanning both the pre-generative AI and generative AI periods. First, we identify a range of geographic biases. These biases span from representation bias in the training data and regional disparities in the factual recall of language models to the tendency of generative AI to over-proportionally favor prototypical places (called defaults). Then, we showcase how recent studies address the latter bias by evaluating geographic diversity in the outputs of generative AI across various cognitive levels, parameter settings, and output modalities.

cs.CY↗

arXiv2026

ResPrune: Text-Conditioned Subspace Reconstruction for Visual Token Pruning in Large Vision-Language Models

Large Vision-Language Models (LVLMs) rely on dense visual tokens to capture fine-grained visual information, but processing all these tokens incurs substantial computational and memory overhead during inference. To address this issue, we propose ResPrune, a training-free visual token pruning framework that enables efficient LVLM inference by selecting a compact yet informative subset of visual tokens. ResPrune formulates visual token pruning as a subspace reconstruction problem and employs a greedy subspace expansion strategy guided by residual energy, allowing it to preserve the geometric structure of the original visual token space. To further incorporate cross modal alignment, the selection process is conditioned on textual relevance, encouraging the retention of tokens that are both informative and instruction-relevant. The proposed method is lightweight and model-agnostic, and can be seamlessly integrated into existing LVLM pipelines without retraining or architectural modifications. Extensive experiments on multiple LVLM backbones, including LLaVA-1.5, LLaVA-NeXT, and Qwen2.5-VL, demonstrate that ResPrune consistently outperforms existing pruning approaches across a wide range of benchmarks, while achieving effective reductions in computation, memory consumption, and inference latency.

cs.LG↗

arXiv2026

Geography According to ChatGPT -- How Generative AI Represents and Reasons about Geography

Understanding how AI will represent and reason about geography should be a key concern for all of us, as the broader public increasingly interacts with spaces and places through these systems. Similarly, in line with the nature of foundation models, our own research often relies on pre-trained models. Hence, understanding what world AI systems construct is as important as evaluating their accuracy, including factual recall. To motivate the need for such studies, we provide three illustrative vignettes, i.e., exploratory probes, in the hope that they will spark lively discussions and follow-up work: (1) Do models form strong defaults, and how brittle are model outputs to minute syntactic variations? (2) Can distributional shifts resurface from the composition of individually benign tasks, e.g., when using AI systems to create personas? (3) Do we overlook deeper questions of understanding when solely focusing on the ability of systems to recall facts such as geographic principles?

cs.AI↗