arXiv subjects

Zhe Zhao

Publications and source records attributed to Zhe Zhao.

At least 37 records · Page 2Linked to original sources

arXiv2026

Science Earth: Towards A Planet-Scale Operating System for AI-Native Scientific Discovery

Scientific discovery demands intelligence, perseverance, and serendipity across vast search spaces. Today, top scientific capabilities remain siloed--one AI system for biological analysis, another for clinical reasoning, mathematical derivation, or materials simulation--and no pre-designed team can anticipate every skill a question will need. Science Earth is a planet-scale scientific runtime in which any capability--a simulation cluster, a wet-lab robot, a proof engine, a single-cell pipeline--can connect to any other, with collaboration structure emerging from the question itself. Its underlying EACN protocol lets capabilities discover one another, negotiate task ownership, and adjudicate across incompatible evidentiary standards without prior knowledge of who will meet whom. This shifts the organizing challenge from workflow design to open-ended connectivity. Two runs validate this under structurally distinct conditions. In a trans-Pacific higher-order Kuramoto synchronization study, agents identified and corrected a closure-ratio assumption in Ott-Antonsen analytic theory that fails outside the Lorentzian limit, within thirty minutes. In an eight-agent single-cell run on the 4.88M-cell Kang 2024 pan-cancer atlas, heterogeneous capabilities coupled over a 64.9-hour window with one structural external instruction, producing three new result layers and anchoring findings against an independent wet-lab study on an adjacent CCR8- TIGIT+ Treg subset. These cases are a first empirical reading, not a benchmark sweep. They show that when AI capabilities are truly connectable and coordination emerges from the problem, scientific reasoning becomes a distributed, self-correcting process--a step towards scaling AI-native discovery to the planet.

cs.AI↗

arXiv2026

Shift-Left High-Level Synthesis Verification via Knowledge-Augmented LLM Agent

High-Level Synthesis (HLS) relies on transforming original C specifications into synthesizable HLS-oriented C (HLS-C) implementations. Functional consistency verification between original C specifications and HLS-C implementations is a critical yet labor-intensive task in HLS design flows. While Large Language Models (LLMs) have recently shown promise in automated testbench generation, their stochastic nature often leads to insufficient coverage, inconsistent verification environments, and unreliable equivalence checking results. To address these limitations, we propose a knowledge-augmented, agent-driven shift-left verification framework for automated functional consistency checking between original C and HLS-C implementations before synthesis. The framework introduces a Dual-Tier Consistency Checking mechanism that jointly enforces static structural alignment and dynamic behavioral equivalence between paired testbenches, while integrating symbolic execution and coverage-driven refinement to improve verification completeness. Furthermore, we construct a heterogeneous HLS Verification Knowledge Graph to provide topology-aware reasoning priors for testbench generation, and design an autonomous verification agent to orchestrate iterative refinement and failure diagnosis across heterogeneous toolchains. Experimental results on 107 HLS benchmark pairs demonstrate that the proposed framework achieves 0.9826 average coverage and 0.9533 dynamic consistency, outperforming representative AST-based, retrieval-augmented, and iterative agent-based baselines. https://github.com/cz-5f/HLS-LeVeri.git

cs.AR↗

arXiv2026

Wisdom of Committee: Diverse Distillation from Large Foundation Models and Domain Experts

Knowledge distillation from foundation models to compact domain models is challenging due to substantial gaps in capacity, architecture, and modality. For example, in our experiments, distilling from a 76M-parameter language model to a 2M-parameter recommender closes less than 40% of the performance gap between the undistilled student and the teacher. We show that introducing domain-specific experts -- which share the student's architectural characteristics -- alongside the foundation model as a diverse teacher committee significantly improves transfer. However, standard multi-teacher methods fail to exploit this diversity: naively combining heterogeneous teachers can degrade performance below single-teacher distillation. To address this, we propose DiverseDistill, an interactive distillation framework that employs a learnable Question-Answer mechanism to generate teacher-conditioned queries and align heterogeneous teacher outputs into the student's representation space. Unlike methods requiring gradient-based co-optimization or architectural modification of teachers, DiverseDistill operates with frozen teachers using only forward-pass inference through their intermediate layers: no parameter updates, no co-training, and no architectural surgery. A dynamic teacher importance mechanism further reduces training cost by filtering low-relevance teachers per sample (e.g., ~30% fewer forward passes with no quality loss for recommendation tasks), while the entire Distillation Module is discarded after training, adding zero inference overhead. Evaluations on recommendation (38x compression) and vision (3.6x compression) tasks demonstrate that DiverseDistill recovers 73-114% of the teacher-student performance gap, consistently outperforming all single- and multi-teacher baselines.

cs.LG↗

arXiv2026

Bridging Creative Intent and Visual Quality: Creator-Driven Recurrent Video Generation with Agentic Feedback Loops

Generative AI has made content creation increasingly accessible, but many AI-generated videos lack narrative coherence and creative direction, issues that become more substantial at longer durations. Unlike coding, where AI generation benefits from reliable feedback and techniques such as recurrent self-improvement, video generation requires subjective feedback about plot, scenes, and narrative, which naturally motivates approaches that incorporate human creative direction. We introduce CHIEF, a human-AI co-creation video generation framework that places the creator at the center of human-in-the-loop iterative video refinement, and supports them by providing automatic subjective feedback. The creator incorporates their creative direction by driving each iteration, while their revisions are incorporated by a specialized refiner agent. The feedback loop is generated by persona-conditioned multimodal LLMs that watch generated videos and produce subjective critique from the audience perspectives, providing feedback that self-evaluation alone cannot capture. To test the effectiveness of our proposed framework, we work with high school and college students with no prior filmmaking experience to create videos, from short 1-minute videos to a complete short 10-minute film with a complicated plot.

cs.CV↗

arXiv2026

Leveraging Metric Depth for Relative Depth Prediction

We present our solution to the 2025 SoccerNet Monocular Depth Estimation Competition Challenge. Predicting the relative depth in football scenarios is challenging, especially with only thousands of training samples available. To address this issue, our method leverages the powerful zero-shot capabilities of models pretrained on large-scale datasets to learn metric depth for effective relative depth prediction, achieving a score of $2.68 \times 10^{-3}$ on the challenge set.

cs.CV↗

arXiv2026

Training Prompt Matters: State-Adaptive Optimization for Robust Fine-Tuning

While prompt engineering is instrumental in maximizing the capabilities of Large Language Models (LLMs) during inference, the role of prompts during training remains critically underexplored. Prevailing fine-tuning paradigms typically treat training prompts as mere surface forms, assuming that semantically equivalent instructions yield identical learning outcomes. However, we reveal that this equivalence is deceptive: while paraphrased prompts often lead to comparable in-task performance, they induce drastically different cross-task impacts regarding catastrophic forgetting and generalization. Crucially, these impacts are positively correlated across tasks, indicating the existence of superior prompts that consistently yield better performance. Furthermore, we discover that these superior prompts can be robustly identified by task loss prior to learning. Leveraging these insights, we introduce State-Adaptive Prompt Optimization (SAPO), a lightweight yet effective training strategy that shifts task formulation from a static input to a dynamic, state-adaptive variable. Comprehensive experiments on diverse benchmarks confirm its effectiveness, which significantly mitigates forgetting while improving generalization, achieving substantial performance gains over state-of-the-art methods. These results provide insights into how training prompts shape learning dynamics and offer a practical recipe for robust fine-tuning. Our code is available at https://github.com/Eric8932/SAPO.

cs.CL↗

arXiv2026

Scaling Agentic Capabilities via Grounded Interaction Synthesis

General agentic intelligence hinges on the ability to interact with diverse real-world tools to complete complex tasks, a capability fundamentally tied to the quality of interaction data. To bypass the prohibitive costs of human annotation, prevailing paradigms depend entirely on Large Language Models (LLMs) to scale the synthesis of agentic environments and tasks. However, such unconstrained generation often degenerates into biased random sampling of LLMs' internal priors, failing to capture the diversity and difficulty of real-world domains or construct high-fidelity, long-horizon tasks. In this work, we introduce Grounded Agentic Interaction Synthesis (GAIS), a framework that automates the scalable construction of diverse environments and complex tasks via a two-phase grounding mechanism. Specifically, we construct protocol-anchored environments derived from real-world Model Context Protocol (MCP) servers to ensure functional diversity and difficulty. Subsequently, we employ structure-guided planning to navigate these environments, actively enforcing logical dependencies and adversarial policies to generate complex tasks. Experiments on BFCL, $τ^2$-Bench, and ACEBench demonstrate that GAIS-synthesized data significantly outperforms state-of-the-art baselines, enabling base models to match or even surpass their official instruction-tuned counterparts. Furthermore, GAIS exhibits superior data efficiency and scalability, achieving exceptional capabilities with significantly less data while maintaining continuous growth where baselines stagnate. Our code and dataset are publicly available at https://github.com/Eric8932/GAIS.

cs.CL↗

arXiv2026

Triaging Threats to Specialized Guardrails

Building robust safety guardrails is essential for deploying Large Language Models across diverse real-world applications. However, this goal remains challenging because safety risks span heterogeneous threat domains, while existing datasets cover only fragmented risk subsets and rely on inconsistent taxonomies. Consequently, it remains unclear whether current guardrails can generalize beyond narrow evaluation settings. To better understand the robustness of guardrail models, we first introduce GuardZoo, a unified human-annotated benchmark with 32,460 samples covering 15 distinct unsafe categories. Evaluation on GuardZoo reveals that monolithic guardrails suffer from task interference: different threat domains require distinct decision boundaries that are difficult to compress into a single model. We therefore propose RouteGuard, a router-expert framework that triages each conversation to specialized expert guardrails for threat-specific detection. Experiments show that RouteGuard improves fine-grained threat detection over strong guardrail baselines, generalizes better under out-of-domain evaluation, and supports flexible modular expansion to emerging threats.

cs.CR↗

arXiv2026

FedMPT: Federated Multi-label Prompt Tuning of Vision-Language Models

Multi-Label Recognition (MLR) based on Vision-Language Models (VLMs) aims to leverage their pre-trained knowledge to better adapt complex recognition scenarios, thereby enhancing model robustness. However, for realistic decentralized applications requiring federated learning, adapting VLMs to each client that possesses private and heterogeneous data can cause the model to overfit spurious label correlations, consequently triggering irrelevant categories when encountering new samples. To tackle this problem, we reconsider the federated learning for MLR with a causal model, in which we adopt a front-door adjustment and decouple the MLR modeling process by intermediate variables that magnify the oracle label co-occurrence. Guided by our analysis, we propose our FedMPT, the first method specifically designed for federated MLR. The core idea of FedMPT is to leverage generalizable conditions to steer federated MLR to mitigate erroneous label activations. To achieve this, FedMPT introduces an Large Language Model (LLM)-driven pipeline to decipher the underlying conditions that govern label dependencies. Furthermore, we introduce an optimal transport between the condition-enriched prompts and the image patches to uncover multiple region-level semantics. Finally, we generate synergistic predictions from different conditions with a crafted gating mechanism. Experiments on multiple benchmark datasets show that our proposed approach achieves competitive results and outperforms SOTA methods under varied settings.

cs.AI↗

arXiv2026

AQuaUI: Visual Token Reduction for GUI Agents with Adaptive Quadtrees

Large Multimodal Models (LMMs) have recently emerged as promising backbones for GUI-agent models, where high-resolution GUI screenshots are introduced to the prompts at each iteration step. However, these screenshots exhibit highly non-uniform spatial information density: large regions may carry little information and are visually homogeneous, while key text and icons may require high visual fidelity. Existing approaches to this problem either require additional training or rely on attention-based token compression, ignoring the structured layout and spatial redundancy of GUI screenshots. To fill the gap, this paper proposes AquaUI, a training-free inference-time token reduction method for GUI agent models that utilizes the non-uniform information density in screenshots. AQuaUI constructs an adaptive quadtree on each screenshot input and keeps one representative merged token per leaf of the quadtree. AQuaUI preserves the spatial positions of retained tokens throughout the pipeline to ensure that all position-encoding stages remain consistent. To further improve temporal consistency across multi-step GUI interactions, we propose a conditional quadtree algorithm that leverages the continuity between consecutive screenshots within a single request. Specifically, it refines the current quadtree using previous quadtrees as references, helping preserve fine-grained regions across static or mildly shifted GUI states. We implement AQuaUI on state-of-the-art GUI agent models and conduct experiments on standard grounding and navigational benchmarks. AQuaUI consistently shows improved accuracy-efficiency trade-offs over prior baselines. Notably, on GUI-Owl-1.5-32B-Instruct, AQuaUI achieves up to 13.22% speedup and 29.52% fewer visual tokens while retaining 99.06% of full-token performance, suggesting that the spatial redundancy of GUI screenshots can be exploited at inference without retraining.

cs.AI↗

arXiv2026

ModelLens: Finding the Best for Your Task from Myriads of Models

The open-source model ecosystem now contains hundreds of thousands of pretrained models, yet picking the best model for a new dataset is increasingly infeasible: new models and unbenchmarked datasets emerge continuously, leaving practitioners with no prior records on either side. Existing approaches handle only fragments of this in-the-wild setting: AutoML and transferability estimation select models from small predefined pools or require expensive per-model forward passes on the target dataset, while model routing presupposes a given candidate pool. We introduce ModelLens, a unified framework for model recommendation in the wild. Our key insight is that public leaderboard interactions, though scattered and noisy, collectively trace out an implicit atlas of model capabilities across heterogeneous evaluation settings, a signal rich enough to learn from directly. By learning a performance-aware latent space over model--dataset--metric tuples, ModelLens ranks unseen models on unseen datasets without running candidates on the target dataset. On a new benchmark of 1.62M evaluation records spanning 47K models and 9.6K datasets, ModelLens surpasses baselines that either rely on metadata alone or require running each candidate on the target dataset. Its recommended Top-K pools further improve multiple representative routing methods by up to 81% across diverse QA benchmarks. Case studies on recently released benchmarks further confirm generalization to both text and vision-language tasks.

cs.LG↗

arXiv2026

From Memorization to Creation: Evaluating the Cognitive Depth of LLM-Generated Educational Questions

While LLMs show promise in automating educational content creation, their ability to generate questions that stimulate higher-order thinking remains understudied. This work evaluates six widely-used LLMs through a Bloom's Taxonomy lens, focusing on their capacity to transcend rote memorization and achieve cognitive leaps. Using a hybrid human--AI evaluation protocol, we generate and analyze 20{,}700 questions across computer science, K--12 math, and social-science domains. Key contributions include: (1) a fine-grained prompting strategy that reduces question repetitiveness by 24.45\% for Qwen2.5-7B-Instruct, and increases the proportion of higher-order cognitive level outputs by 11.53\% for InternLM3-8B-Instruct; (2) quantitative metrics for cognitive shift intensity (CogShift) and category drift, revealing InternLM3's superior performance in multi-level transitions; (3) an interpretability analysis revealing metric-level correlations that enhance the transparency of Chain-of-Thought prompting. Our findings highlight the importance of cognitive-aware prompt design and provide benchmarks for deploying LLMs in personalized learning systems.

cs.HC↗

arXiv2026

DiffQEC: A versatile diffusion model for quantum error correction

Quantum computers could solve problems beyond the reach of classical devices, but this potential depends on quantum error correction (QEC) to protect fragile quantum states from noise. A central challenge in QEC is decoding: inferring likely physical errors from syndrome patterns generated by repeated stabilizer measurements. Existing decoders, including graph-based and neural approaches, typically return a single correction hypothesis and therefore discard the richer posterior structure of the error distribution conditioned on the observed syndrome. Here we recast QEC decoding as posterior inference using discrete denoising diffusion, exploiting the analogy between stochastic error accumulation and the forward diffusion process. We introduce DiffQEC, a generative decoder that combines a syndrome processor for multi-round spatial-temporal syndrome histories with syndrome feature modulation to condition denoising on the observed syndrome throughout inference. On experimental data from Google's superconducting quantum processor, DiffQEC reduces logical error rates by up to 10.2% relative to minimum-weight perfect matching and by about 5% relative to tensor-network decoding. These improvements persist for larger code distances up to 17 under depolarizing noise and for logical circuits of increasing depth. Beyond accuracy, the learned posterior provides confidence estimates for post-selection and reveals physically meaningful error structure, establishing posterior generative decoding as a practical framework for QEC.

quant-ph↗

arXiv2026

ResearchEVO: An End-to-End Framework for Automated Scientific Discovery and Documentation

An important recurring pattern in scientific breakthroughs is a two-stage process: an initial phase of undirected experimentation that yields an unexpected finding, followed by a retrospective phase that explains why the finding works and situates it within existing theory. We present ResearchEVO, an end-to-end framework that computationally instantiates this discover-then-explain paradigm. The Evolution Phase employs LLM-guided bi-dimensional co-evolution -- simultaneously optimizing both algorithmic logic and overall architecture -- to search the space of code implementations purely by fitness, without requiring any understanding of the solutions it produces. The Writing Phase then takes the best-performing algorithm and autonomously generates a complete, publication-ready research paper through sentence-level retrieval-augmented generation with explicit anti-hallucination verification and automated experiment design. To our knowledge, ResearchEVO is the first system to cover this full pipeline end to end: no prior work jointly performs principled algorithm evolution and literature-grounded scientific documentation. We validate the framework on two cross-disciplinary scientific problems -- Quantum Error Correction using real Google quantum hardware data, and Physics-Informed Neural Networks -- where the Evolution Phase discovered human-interpretable algorithmic mechanisms that had not been previously proposed in the respective domain literatures. In both cases, the Writing Phase autonomously produced compilable LaTeX manuscripts that correctly grounded these blind discoveries in existing theory via RAG, with zero fabricated citations.

cs.AI↗

arXiv2026

Testing black hole metrics with binary black hole inspirals

Gravitational wave astronomy has opened an unprecedented window onto tests of gravity and fundamental physics in the strong-field regime. In this study, we examine a series of well-motivated deviations from the classical Kerr solution of General Relativity and employ gravitational wave data to place constraints on possible deviations from the Kerr geometry. The method involves calculating the phase of gravitational waves using the effective one-body formalism and then applying the parameterized post-Einsteinian framework to constrain the parameters appearing in these scenarios beyond General Relativity. The effective one-body method, known for its capability to model complex gravitational waveforms, is used to compute the wave phase, and the post-Einsteinian framework allows for a flexible, model-independent approach to parameter estimation. We demonstrate that gravitational wave data provide evidence supporting the Kerr nature of black holes, showing no significant deviations from General Relativity, thereby affirming its validity within the current observational limits. We further assess the potential impact of orbital eccentricity and find that, within observationally allowed ranges, its contribution to the inferred deviations is subdominant. This work bridges theoretical waveform modeling with observational constraints, providing a pathway to test the no-hair theorem and probe the astrophysical viability of modified black holes.

gr-qc↗

arXiv2026

Automating Android Build Repair: Bridging the Reasoning-Execution Gap in LLM Agents with Domain-Specific Tools

Android is the largest mobile platform, yet automatically building applications remains a practical challenge. While Large Language Models (LLMs) show promise for code repair, their use for fixing Android build errors remains underexplored. To address this gap, we first introduce AndroidBuildBench, a benchmark of 1,019 build failures curated from the commit histories of 43 open-source Android projects. Each problem is paired with a verified solution from a subsequent commit, ensuring that fixes are feasible. Second, we propose GradleFixer, an LLM agent with domain-specific tools for inspecting and manipulating the Gradle build environment. GradleFixer achieves a resolve rate of 81.4% (pass@1), significantly outperforming a state-of-the-art coding agent that relies on a general-purpose shell. GradleFixer's success suggests that while LLMs possess the high-level knowledge to solve these failures, they struggle to translate this knowledge into effective low-level actions using a general-purpose shell. We demonstrate the effectiveness of a strategy we term Tool Bridging, which replaces general-purpose shell commands with domain-aware abstractions. We hypothesize this approach works through two mechanisms: 1) it provides tools in an API-like format that LLMs use more reliably, and 2) it constrains the action space to relevant operations. This approach bridges the gap between the model's high-level reasoning and effective low-level execution.

cs.SE↗

arXiv2026

CARPE: Context-Aware Image Representation Prioritization via Ensemble for Large Vision-Language Models

Large vision-language models (LVLMs) are typically trained using autoregressive language modeling objectives, which align visual representations with linguistic space. While effective for multimodal reasoning, this alignment can weaken vision-centric capabilities, causing LVLMs to underperform their base vision encoders on tasks such as image classification. To address this limitation, we propose Context-Aware Image Representation Prioritization via Ensemble (CARPE), a lightweight framework that integrates raw vision features with aligned LLM representations through vision-integration layers and a context-aware ensemble mechanism. This design enhances the model's ability to adaptively weight visual and textual modalities and enables the model to capture various aspects of image representations. Extensive experiments demonstrate that CARPE improves performance on both image classification and diverse vision-language benchmarks. Our results suggest that modality balancing plays a critical role in multimodal generalization by improving representation utilization within autoregressive LVLMs.

cs.CV↗

arXiv2026

Simulation-Based Inference for Probabilistic Galaxy Detection and Deblending

Stage-IV dark energy wide-field surveys, such as the Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST), will observe an unprecedented number density of galaxies. As a result, the majority of imaged galaxies will visually overlap, a phenomenon known as blending. Blending is expected to be a leading source of systematic error in astronomical measurements. To mitigate this systematic, we propose a new probabilistic method for detecting, deblending, and measuring the properties of galaxies, called the Bayesian Light Source Separator (BLISS). Given an astronomical survey image, BLISS uses convolutional neural networks to produce a probabilistic astronomical catalog by approximating the posterior distribution over the number of light sources, their centroids' locations, and their types (galaxy vs. star). BLISS additionally includes a denoising autoencoder to reconstruct unblended galaxy profiles. As a first step towards demonstrating the feasibility of BLISS for cosmological applications, we apply our method to simulated single-band images whose properties are representative of year-10 LSST coadds. First, we study each BLISS component independently and examine its probabilistic output as a function of SNR and degree of blending. Then, by propagating the probabilistic detections from BLISS to its deblender, we produce per-object flux posteriors. Using these posteriors yields a substantial improvement in aperture flux residuals relative to deterministic detections alone, particularly for highly blended and faint objects. These results highlight the potential of BLISS as a scalable, uncertainty-aware tool for mitigating blending-induced systematics in next-generation cosmological surveys.

astro-ph.IM↗