Searcharxiv⌕ Search

arXiv subjects

Zirui Zhou

Publications and source records attributed to Zirui Zhou.

At least 19 recordsLinked to original sources
arXiv2026

Seeing Through the Illusion: Test-Time Scaling for World Action Models

World action models (WAMs) jointly generate actions and imagined futures, but a plausible prediction need not be realized by its co-generated action. We introduce WISE (\emph{World Action Inference Scaling through Evaluation}), a pre-execution selector that scores each sampled rollout on action validity and temporal regularity, inverse-dynamics consistency between the proposed action and the imagined motion, and instruction-conditioned task progress, and executes the highest-scoring candidate while keeping the WAM frozen. With four candidates per decision, WISE raises closed-loop success from 6.0\% to 24.0\% on RoboLab with Cosmos~3 Edge and from 75.1\% to 78.3\% mean success on LIBERO with DreamZero, and exceeds the best existing selector in aggregate by 4.0 and 0.9 points. Component ablations and agreement analyses show that the three scores rank candidates differently and that task progress carries the largest weight in the fused decision.

cs.RO↗
arXiv2026

RoboSnap: One-Shot Real-to-Sim Scene Generation for Generalizable Robot Learning and Evaluation

Recovering real-world scenes as interactive simulation environments can enable generalizable robot learning and reproducible policy evaluation. However, constructing scenes that are both physically stable and visually faithful remains slow and expensive. In this work, we present RoboSnap, a real-to-sim framework that turns a single RGB image into a simulation-ready scene. The key idea is a layered design that separates the physics-critical interaction area from the surrounding visual context: collision-aware foreground assets are refined for stable robot interaction, while a 3D Gaussian splatting visual layer preserves faithful background appearance under novel views. Experiments on DROID scenes and real-robot tasks show that RoboSnap achieves reliable trajectory replay in the recovered scenes, supports task-specific synthetic data generation for policy training, and yields meaningful sim-real correlation for policy evaluation. To further support real-to-sim research, we introduce DROID-Sim, a real-to-sim companion dataset constructed from 564 real-world scenes in DROID. Extensive experiments suggest that the value of real-to-sim methods lies not only in high-fidelity visual reconstruction, but in turning real environments into reusable infrastructure for robot learning and evaluation.

cs.RO↗
arXiv2026

HEDGE: Gated Hypergraph Memory for Scalable Multi-Agent Debate

Multi-agent debate systems require agents to retain and revisit arguments across multiple rounds, yet limited context windows and unstructured memories make relevant information increasingly difficult to recover. Existing flat and pairwise-graph memories also struggle to capture higher-order relations, including multi-agent agreement, conflicting evidence, and evolving consensus. We introduce HEDGE, a selective memory framework that represents debate trajectories as semantic hypergraphs. Hyperedges connect related claims across agents, while intra-hyperedge relational graphs preserve fine-grained argumentative structure. A learned memory selection module (MSM), trained through supervised learning and reinforcement learning, routes each memory candidate to STORE, BACKUP, or OMIT, retaining useful information while limiting redundancy. Across seven diverse benchmarks, HEDGE consistently outperforms strong single-agent, multi-agent, and memory baselines, improving accuracy by 5.0 points on average over the strongest multi-agent system with GPT-4o-mini. The gains reach 8.2 points on conflict-heavy tasks, while HEDGE reduces memory usage by 48% and the number of debate rounds by 34%.

cs.CL↗
arXiv2026

LLM-Guided Evolution of Symbolic Acceleration Inequality Families for MILP

Integer programming (IP) is central to many combinatorial optimization tasks but remains challenging due to its NP-hard nature. A practical way to improve IP solvers is to manually design inequality families that exploit problem structure. However, this creative process requires deep expertise and has therefore been difficult to automate. Our proposed framework, Symbolic Acceleration via Generative Evolution (SAGE), automates symbolic discovery of acceleration inequality families at the modeling level. It reasons over a symbolic IP formulation and a natural language description of the problem to discover a reusable set of acceleration inequality families that can be instantiated for each problem instance. These acceleration inequality families are selected through empirical admissibility and impact screening rather than mathematical proofs. SAGE (i) initializes a population of candidate acceleration inequality families via an initializer agent that uses an LLM; (ii) it screens candidates for admissibility on a small development admissibility set for solution feasibility and the separation of the relaxation solution; and (iii) it iteratively refines the population of inequalities through evolutionary crossover and mutation agents. Across seven MILP benchmarks, SAGE reduces optimality gaps by up to $62.1\%$ relative to baseline MILP formulations solved with a fixed time budget and reaches target gaps up to $5.51\times$ faster according to shifted geometric mean speedup. Comprehensive ablations show its robustness across different LLM backends and across solvers and internal cut settings. Code: https://github.com/milad1378yz/sage

cs.AI↗
arXiv2026

Organon : From Evolutionary Search to Recursive Self-Improving Agents for Scientific Discovery

Scientific discovery does not follow a single fixed methodology: research methods vary across domains and evolve through experimentation as investigations unfold. Although LLM-guided evolutionary search has advanced automated discovery, existing frameworks typically rely on fixed search procedures or restrict adaptation to predefined components, limiting their ability to address diverse research problems and adapt to challenges that emerge at runtime. We introduce Organon, a recursively self-improving agent for scientific discovery. During execution, a single persistent agent analyzes its accumulated experience and autonomously revises its own search strategies, population management, diagnostic tools, memory, and worker coordination. These revisions take effect at runtime, allowing the agent to continuously adapt its approach as research progresses. Importantly, the agent can also diagnose and correct flaws in its own adaptation procedures, enabling recursive improvement of the mechanisms that guide its future self-improvement. We evaluate Organon across 21 computational discovery tasks spanning algorithm design, GPU kernel optimization, mathematical construction, and systems optimization. On ALE-Bench Lite, Organon achieves the best performance among competing search frameworks on all 10 tasks, improving mean held-out performance by 20.5% over EvoX. On KernelBench, it achieves the best results on three of four workloads, with speedups of up to 1.67x over the strongest competing method. Analysis of the search trajectories further reveals how the agent develops task-specific search strategies, builds new diagnostic capabilities, and corrects or reverses its own earlier revisions. We implement Organon using DeepSeek Harness and Cordis's spatiotemporal composability. Code is available at https://github.com/cheikh025/Organon.

cs.NE↗
arXiv2026

Prefix-Tuning for Arbitrary Output Sequences on Pretrained Transformers

Prefix-tuning is a parameter-efficient fine-tuning method for large language models (LLMs) that steers model outputs by prepending task-specific prefix representations to the input sequence. Despite its practical success, the theoretical understanding of the expressiveness of prefixes remains incomplete. In particular, the following question remains open: Given a pretrained autoregressive Transformer, for any input-output sequence pair $(\mathtt{x},\mathtt{y})$, does there always exist a prefix $\mathcal{S}$ such that prepending $\mathcal{S}$ to $\mathtt{x}$ induces the model to generate the target output $\mathtt{y}$? In this work, we prove that under the assumption that $\mathtt{y}$ is in the range of the pretrained Transformer, such a prefix $\mathcal{S}$ exists and can be explicitly computed by solving multiple linear systems. We further show that this assumption is necessary, in the sense that such a prefix may not exist when the assumption fails, thereby providing a complete answer to the above question. Our analysis is constructive and the obtained theoretical results hold under fairly general assumptions on model architecture. Extensive experiments on GPT-2 and Qwen3-4B demonstrate that our algorithm produces prefixes with strong control over model outputs.

cs.CL↗
arXiv2026

VLMs are Native Action Models: Unleashing the Manipulation Capabilities of VLMs with Explicit Action-Space Grounding

Pretrained vision-language models (VLMs) already exhibit strong visual, spatial and reasoning capabilities relevant to robotic manipulation. The key challenge in turning pretrained VLMs into generalist manipulation policies is to expose these capabilities through an appropriate action interface. Vision-language-action (VLA) models provide a unified trainable policy, but typically introduce learned action tokens, action heads, or action experts whose semantics must be learned from robot data pretraining. In contrast, modular foundation-model-based manipulation methods can exploit pretrained VLM capabilities for zero-shot manipulation, but rely on external planning and control scaffolds that are not end-to-end trainable as a unified continuous-action policy. In this work, we introduce EASG, an Explicit Action-Space Grounding framework that bridges these two paradigms. By projecting a tool center point (TCP)-centered action frame into the wrist camera observation and expressing actions in the form of VLM's native textual generation interface, EASG makes the semantics of a continuous robot action space directly observable in the VLM's perceptual context. This allows pretrained VLMs to directly generate executable robot actions using their existing visual-spatial reasoning capabilities, without robot-action pretraining or additional action modules. Besides, the same action interface supports direct fine-tuning with robot trajectories, allowing the models to continuously improve beyond their manipulation capabilities unleashed by EASG. Overall, EASG enables zero-shot manipulation and generalization across object positions, categories, and robot embodiments, while supporting effective fine-tuning in the same continuous action space. These results demonstrate that, given an appropriate action interface, pretrained VLMs can themselves serve as native action models for robotic manipulation.

cs.RO↗
arXiv2026

DecomVoxel: Harnessing 3D-Native Priors with Guided In-situ Denoising Optimization for Decompositional Scene Reconstruction

Decompositional scene reconstruction aims to reconstruct high-quality objects and background, yet existing methods still struggle with the level of quality under heavy occlusions. While generative priors offer a potential solution, 2D image-based priors often suffer from multi-view inconsistency due to a lack of 3D awareness. Conversely, 3D-native priors provide stronger structural inductive biases but frequently lead to spatial drift and misalignment within complex scenes. To address these issues, we propose DecomVoxel, formulating object completion as a guided in-situ denoising optimization that bridges 3D-native priors with neural scene reconstruction. Our framework introduces a reformulated epsilon-based distillation loss to ensure stable latent refinement, alongside adaptive spatial guidance that utilizes occupied and vacant anchors with temporal annealing to suppress generative hallucinations and mitigate spatial drift. Experiments on Replica and ScanNet++ show that DecomVoxel significantly outperforms state-of-the-art methods while faithfully preserving the original spatial layout, structural fidelity, and style-consistent texture. Our method pushes the boundary of decompositional reconstruction by delivering high-quality textured meshes with clean topology, geometry, and appearance, providing a robust solution for the decompositional reconstruction of complex real-world scenes. Code is available at https://github.com/DecomVoxel/DecomVoxel.

cs.CV↗
arXiv2026

SEDIMA: Cross-Run Hierarchical Insight Memory for Evolutionary Search Agents

Large language model (LLM)-driven evolutionary search is a powerful paradigm for automated program and algorithm discovery, yet existing systems are largely memoryless: each run explores from scratch, so agents repeatedly rediscover the same improvements and re-encounter the same dead ends. We introduce SEDIMA, a persistent hierarchical insight memory for evolutionary search agents. SEDIMA distills raw traces into natural-language insights, clusters them by semantic similarity using attention-weighted centroids, and retrieves relevant guidance to condition future mutations, accumulating transferable knowledge across runs and problems rather than within a single trajectory. As a drop-in module that leaves the search operators unmodified, SEDIMA improves average final performance by 5.5% on AlgoTune and 6.6% on ALE-Bench LITE under a fixed budget of 100 evaluated candidates. Under OpenEvolve, SEDIMA requires 32.3% fewer iterations on average to reach baseline-best performance across the five evaluated backbones.

cs.NE↗
arXiv2026

Remote epitaxy beyond polarity

Remote epitaxy through a monolayer two-dimensional material-covered substrate establishes a crystallographic registry across the van der Waals (vdW) surface that enables the epitaxial growth, lift-off and transfer of single-crystalline films. A central belief in remote epitaxy is that the substrate facilitating the phenomenon must be a material with strong ionicity, as the interatomic electrostatic potential fluctuation in covalent and metallic materials is substantially attenuated by two-dimensional materials. Here, we show remote epitaxy is possible when the substrate is a metallic or covalently bonded material and experimentally demonstrate non-polar remote homo- and heteroepitaxy across a wide range of material systems, including both metals and semiconductors. The achieved non-polar remote interactions are designed and engineered by harnessing substrate conductivity and vicinal surface step-edge density. These findings indicate that remote epitaxy is universal and applicable to ionic, metallic, and covalent materials, expanding its capabilities and stimulating a plethora of new fundamental scientific questions about the mechanism of remote epitaxy.

cond-mat.mtrl-sci↗
arXiv2026

AIvilization v0: Toward Large-Scale Artificial Social Simulation with a Unified Agent Architecture and Adaptive Agent Profiles

AIvilization v0 is a publicly deployed large-scale artificial society that couples a resource-constrained sandbox with a unified LLM-agent architecture, aiming to sustain long-horizon autonomy while remaining executable under a rapidly changing environment. To mitigate the tension between goal stability and reactive correctness, keeping long-horizon objectives on course while each action remains valid in a fast-changing shared world, we introduce (i) a hierarchical branch-thinking planner that decomposes life goals into parallel objective branches and uses simulation-guided validation plus tiered re-planning to ensure feasibility; (ii) an adaptive agent profile with dual-process memory that separates short-term execution traces from long-term semantic consolidation, enabling persistent yet evolving identity; and (iii) a human-in-the-loop steering interface that injects long-horizon objectives and short commands at appropriate abstraction levels, with effects propagated through memory instead of brittle prompt overrides. The environment integrates physiological survival costs, non-substitutable multi-tier production, an AMM-based price mechanism, and a gated education-occupation system. In a large-scale public deployment with tens of thousands of agents, high-frequency transactions from the platform's mature phase reveal stable markets that reproduce key stylized facts of real economies and structured wealth stratification driven by education and access constraints. At the agent level, portraits evolve coherently over long horizons, and human steering is associated with measurably larger short-horizon profile updates. Controlled ablation experiments complement the deployment evidence, showing that our agent architecture is robust in multi-objective, long-horizon settings.

cs.MA↗
arXiv2026

MASPRM: Multi-Agent System Process Reward Model

Inference-time search over multi-agent systems (MAS) wastes compute when it cannot identify which agent's intermediate message advanced progress. We present the Multi-Agent System Process Reward Model (MASPRM), which scores routed transcripts (ordered sequences of messages between agents) and acts as an inference controller for step-level beam search (SBS) and Monte Carlo Tree Search (MCTS). MASPRM is trained from multi-agent MCTS rollouts labeled only with terminal outcome rewards, without human step-level annotations. We evaluate on GSM8K, MATH, MMLU, and LogiQA. Under matched scorer size and comparable MCTS budget, MASPRM exceeds a size-matched ORM by $+2.0$ to $+3.0$ points at 1.5B and $+4.1$ to $+14.5$ at 7B across all four benchmarks, with additional scorer-scaling gains over policy likelihood at 7B (avg $+13.4$ under MCTS). MASPRM also improves ranking quality, reducing Hit@1 to Hit@5 gaps by up to $10.3$ points, with the largest gains under stepwise search that uses intermediate decisions. Code: https://github.com/milad1378yz/MASPRM

cs.MA↗
arXiv2026

IntAttention: A Fully Integer Attention Pipeline for Efficient Edge Inference

Deploying Transformer models on edge devices is limited by latency and energy budgets. While INT8 quantization effectively accelerates the primary matrix multiplications, it exposes the softmax-related path as the dominant bottleneck. This stage incurs a costly dequantize -> softmax -> requantize detour, which can account for up to 65% of total attention latency and disrupts the end-to-end integer dataflow critical for edge hardware efficiency. To address this limitation, we present IntAttention, the first fully integer attention pipeline that serves as a training-free drop-in replacement. At the core of our approach lies IndexSoftmax, a hardware-friendly operator that replaces floating-point exponentials entirely within the integer domain. IntAttention integrates sparsity-aware clipping, a 32-entry lookup table approximation, and direct integer normalization, thereby eliminating datatype conversion overhead along the attention path. Experiments on Armv8 CPUs show that our method achieves up to 3.7x speedup and 61% energy reduction over FP16 baselines, and up to 2.0x speedup over conventional INT8 attention pipelines. Across diverse language and vision models, as well as additional reasoning and long-context evaluations, IntAttention maintains strong overall fidelity and demonstrates a more favorable trade-off than existing LUT-based softmax approximations. Code is available at https://github.com/WanliZhong/IntAttention

cs.LG↗
arXiv2026

COAgents: Multi-Agent Framework to Learn and Navigate Routing Problems Search Space

Although Vehicle Routing Problems (VRP) are essential to many real-world systems, they remain computationally intractable at scale due to their combinatorial complexity. Traditional heuristics rely on handcrafted rules for local improvements and occasional \textit{jumps} to escape local minima, but often struggle to generalize across diverse instances. We introduce \textbf{COAgents}, a cooperative multi-agent framework that models the search process as a graph: nodes represent solutions, and edges correspond to either local refinements or large perturbations for diversification (i.e., jumps). A \textit{Partial Search Graph} (PSG) is dynamically constructed during search, enabling COAgents to train a Node Selection Agent and a Move Selection Agent to guide intensification, and a Jump Agent to trigger well-timed explorations of new regions. Unlike end-to-end learning approaches, COAgents cleanly separates problem-agnostic search control from compact domain-specific encoding, facilitating adaptability across tasks. Extensive experiments on the CVRP and VRPTW benchmarks show that COAgents remains competitive with several learn-to-search baselines on CVRP and sets a new state of the art among learning-based methods on the more challenging VRPTW instances, reducing the gap to the best-known solutions by 14\% at $N\!=\!100$ and 44\% at $N\!=\!50$ relative to the strongest neural solver (POMO), and by 21\% and 40\% respectively relative to ALNS. Code is available at https://github.com/mahdims/COAgents.

cs.AI↗
arXiv2026

Latent Heuristic Search: Continuous Optimization for Automated Algorithm Design

The integration of Large Language Models (LLMs) into evolutionary frameworks has established a new paradigm for automated heuristic discovery. Despite their promise, these methods typically search in the discrete space of program syntax, relying on stochastic sampling to navigate a highly non-convex optimization landscape. This work proposes a continuous heuristic discovery framework that shifts optimization to a learned latent manifold. We employ an encoder to map discrete programs into continuous embeddings and train a differentiable surrogate model to predict performance, enabling gradient-based search. To regularize the optimization trajectory, an invertible normalizing flow maps these embeddings to a structured Gaussian prior, where we perform gradient ascent. The resulting optimized latent vectors are projected through a learned mapper into soft prompts, which condition a frozen LLM to synthesize novel executable heuristics. We evaluate the proposed method on the Traveling Salesman Problem (TSP), the Capacitated Vehicle Routing Problem (CVRP), the Knapsack Problem (KSP), and Online Bin Packing (OBP). Empirical results demonstrate that continuous latent-space optimization achieves performance competitive with state-of-the-art discrete evolutionary baselines while offering a complementary methodological alternative for automated algorithm design. The implementation code is available at \url{https://github.com/cheikh025/LHS}.

cs.AI↗
arXiv2026

Universal effect of ammonia pressure on synthesis of colloidal metal nitrides in molten salts

Metal nitrides represent a large class of materials with extensive applications in optoelectronics, energy, and healthcare technologies. For example, GaN and related nitride semiconductors are key materials for solid-state lighting and high-power electronics, TiN and other early transition metal nitrides (TMNs) are widely used in wear-resistant alloys, tool coatings, catalysts and medical implants. Strong metal-nitrogen bonds grant nitrides structural rigidity as well as chemical and thermal stability. However, the covalency of metal-nitrogen bonds necessitates high temperatures to synthesize crystalline metal nitrides. Common synthetic routes include high-temperature solid-state nitridation, crystal growth in supercritical ammonia, molecular-beam epitaxy (MBE), reactive sputtering, and chemical vapor deposition (CVD). The solution synthesis of colloidal nitride nanocrystals (NCs) is rare and particularly challenging because commonly used solvents and surfactants decompose at temperatures far below those required for crystallization of most metal nitrides. Here we report a general approach to solution synthesis of colloidal metal nitride NCs by reacting metal halides and ammonia dissolved in molten inorganic salts at elevated pressures. Successful syntheses of colloidal TiN, VN, GaN, NbN, Mo2N, Ta3N5, W2N, as well as ternary Ti1-xVxN NCs, are demonstrated. These NCs expand the scope of solution-processable technologically important materials.

cond-mat.mtrl-sci↗
arXiv2026

Atomic Alignment in PbS Nanocrystal Superlattices with Compact Inorganic Ligands via Reversible Oriented Attachment of Nanocrystals

Nanocrystals (NCs) serve as versatile building blocks for the creation of functional materials, with NC self-assembly offering opportunities to enable novel material properties. Here, we demonstrate that PbS NCs functionalized with strongly negatively charged metal chalcogenide complex (MCC) ligands, such as $Sn_2S_6^{4-}$ and $AsS_4^{3-}$, can self-assemble into all-inorganic superlattices with both long-range superlattice translational and atomic-lattice orientational order. Structural characterizations reveal that the NCs adopt unexpected edge-to-edge alignment, and numerical simulation clarifies that orientational order is thermodynamically stabilized by many-body ion correlations originating from the dense electrolyte. Furthermore, we show that the superlattices of $Sn_2S_6^{4-}$-functionalized PbS NCs can be fully disassembled back into the colloidal state, which is highly unusual for orientationally attached superlattices with atomic-lattice alignment. The reversible oriented attachment of NCs, enabling their dynamic assembly and disassembly into effectively single-crystalline superstructures, offers a pathway toward designing reconfigurable materials with adaptive and controllable electronic and optoelectronic properties.

cond-mat.mtrl-sci↗
arXiv2025

3D Scene Change Modeling With Consistent Multi-View Aggregation

Change detection plays a vital role in scene monitoring, exploration, and continual reconstruction. Existing 3D change detection methods often exhibit spatial inconsistency in the detected changes and fail to explicitly separate pre- and post-change states. To address these limitations, we propose SCaR-3D, a novel 3D scene change detection framework that identifies object-level changes from a dense-view pre-change image sequence and sparse-view post-change images. Our approach consists of a signed-distance-based 2D differencing module followed by multi-view aggregation with voting and pruning, leveraging the consistent nature of 3DGS to robustly separate pre- and post-change states. We further develop a continual scene reconstruction strategy that selectively updates dynamic regions while preserving the unchanged areas. We also contribute CCS3D, a challenging synthetic dataset that allows flexible combinations of 3D change types to support controlled evaluations. Extensive experiments demonstrate that our method achieves both high accuracy and efficiency, outperforming existing methods.

cs.CV↗