arXiv subjects

Rui Meng

Publications and source records attributed to Rui Meng.

At least 19 recordsLinked to original sources

arXiv2026

EmbeddingGemma 2: Unified and Efficient Multimodal Representations

We present EmbeddingGemma 2, an open multimodal embedding model which natively represents text, images, audio, video, and combinations thereof in a unified space. Our approach introduces improvements such as simplified per-layer embeddings and a multi-stage curriculum with modality and objective progression which ensure strong quality across unimodal, cross-modal, and mixed-modal evaluations. Notably, in comprehensive embedding benchmarks across all modalities, EmbeddingGemma 2 establishes a new Pareto frontier, outperforming both specialist models and models two to twelve times larger. Furthermore, our design offers flexibility for resource-constrained environments, supporting selective modality encoder loading, adjustable vision token budgets, truncatable embedding dimensions, and quantized weights. We present ablations analyzing key decisions and release EmbeddingGemma 2 to promote further research.

cs.LG↗

arXiv2026

Evolving Intelligent Complex Systems via Intellicise Networks: Architecture, Technologies, and Pathways

Future engineering infrastructures are evolving into large-scale, open, heterogeneous, and wirelessly interconnected complex systems. These systems present significant challenges in optimizing network resource utilization, managing high-dimensional information spaces, and accommodating diverse business requirements. Intellicise networks, characterized by Intent-driven operation, semantic-native capability, and distributed intelligence, offer a promising paradigm for enabling such intelligent complex systems. We provide a systematic exploration of future intelligent complex systems from the perspective of intellicise networks. Specifically, we propose a cross-domain intelligent communication network architecture based on intellicise networks, grounded in information theory, systems theory, game theory, and cybernetics. The architecture comprises a cross-layer organizational framework, multi-functional planes, and novel information flows. The cross-layer framework defines the vertical evolution from perception and cognition to decision, while the control, user, data, computation, intelligence, and security planes deliver horizontal intellicise capabilities. Moreover, data, knowledge, model, and task flows interconnect the various layers and planes, forming a closed-loop process that derives simplicity from high-level intelligene while concurrently pursuing enhanced. Building on this architecture, we review key enabling technologies, tracing their evolution from semantic extraction to intent understanding, from heterogeneous resource integration to self-configuration and self-optimization, from generative artificial intelligence (AI) to agentic AI, and from embodied AI to symbodied AI. Additionally, we present a case study on intellicise networks for embodied agent communications and discuss representative applications and services for intelligent complex systems.

eess.SP↗

arXiv2026

Object-Oriented Harness: Evolving Agent Harnesses through Object-Oriented Design

Agent harnesses are executable code that specifies the prompts, tools, and control flow around a language model. A single global harness can struggle with heterogeneous tasks that require different execution strategies. In this work, we introduce Object-Oriented Harness (OO-Harness), a harness optimization framework that evolves a hierarchy of executable harness classes using object-oriented design. During training, an evolution agent uses execution records to create a family of specialized harnesses, each implemented as a subclass of the base or an existing specialized harness via inheritance. At test time, a harness router uses routing lessons accumulated during training to select a specific harness for each task. Across three long-horizon benchmarks spanning business workflows, workspace tasks, and tool use, OO-Harness achieves the best performance among the compared methods, with an absolute improvement of up to 15% over the base. Extensive ablation studies examine the contributions of its core components.

cs.CL↗

arXiv2026

SemVideoSte: Agentic AI for Cross-Modal Semantic Video Steganography Communication

We propose an agentic artificial intelligence (AI)-driven cross-modal semantic video steganography communication (SemVideoSte) scheme for secure cross-modal information transmission. Agentic AI interprets the service context to determine a semantically matched public prompt and coordinates video generation, message embedding, semantic transmission, and message recovery process. A multi-mask mechanism combines variational autoencoder (VAE) reconstruction robustness, semantic transmission robustness, and latent separability to select reliable embedding positions. Meanwhile, a private digital token controls the random seed and guidance offset to construct positive and negative terminal latent states, enabling coverless message embedding during the final diffusion sampling step. Simulation results on open-source datasets demonstrate that SemVideoSte achieves high embedding capacity, reliable message recovery, and strong semantic consistency under different channel conditions. In addition, eavesdroppers without the correct digital token cannot reliably recover the secret message, and the generated stego videos effectively resist the steganalysis detection. We further establish a prototype testbed to verify the practical feasibility of SemVideoSte.

cs.CR↗

arXiv2026

DuoDiscover: Scaling Autonomous Discovery Beyond the Evaluation Bottleneck

Evaluation has been central to advances in AI, and it is equally critical to autonomous discovery. Yet scaling autonomous discovery to real-world engineering challenges faces a fundamental bottleneck: the cost of evaluating progress. In real settings, a single true evaluation, such as a full benchmark or serving trace, can take hours, so a search can afford only a handful of true evaluations. We present DuoDiscover, a search system that jointly discovers what to optimize and how to evaluate progress efficiently, through three mechanisms. First, each worker in a fleet of coding agents co-discovers an optimization target and a fast, tailored proxy evaluator, which we call a yardstick. The agent then hill-climbs the yardstick rather than the true evaluation. Second, since search can exploit gaps between a yardstick and the true objective, fidelity tests continually recheck each yardstick's ranking and invalidate any that drift. Third, the fleet is bottom-up, with no central decomposition of the problem. A lightweight orchestrator spawns and terminates agents, briefs running ones, and seeds new agents with others' confirmed work. DuoDiscover enables autonomous optimization from software to chip architecture. We demonstrate DuoDiscover on two repository-level benchmarks and four GPU and TPU kernel suites covering hundreds of kernels. On SWE-fficiency, it doubles the average speedup of human experts. On SOL-Exec, it ranks first on 121 of 235 kernels and in the top three on 212. On the heavily optimized vLLM and SQLite engines, DuoDiscover achieves 1.92$\times$ and 1.38$\times$ gains over the base code. On a CPU microarchitecture optimization task with evaluations of up to 28 hours, DuoDiscover surpasses both AlphaEvolve-based ArchAgent v2 and the best human design, with 1.65$\times$ ArchAgent v2's gain in 16.7 days versus more than 40 days, advancing the frontier of autonomous chip design.

cs.CL↗

arXiv2026

AIM: Agentic Idea Management for Automated Research

Frontier LLMs are increasingly used to automate scientific research through iterative search. We distinguish idea-driven search from solution-driven search and identify three core challenges: organizing evolving research ideas, selecting promising directions, and maintaining alignment between ideas and their implementations. To address these challenges, we introduce the Agentic Idea Manager (AIM), a fully autonomous framework for managing and exploring research directions in idea-driven automated research. Inspired by Bayesian optimization, AIM uses an Agentic Surrogate and an Agentic Acquisition mechanism to organize discovered ideas and guide their selection. A Solution Auditor maintains idea-solution integrity, while a Resource Planner adaptively allocates the remaining experimental budget across parallel search branches. Experiments on 10 AutoLab benchmark tasks show that AIM surpasses the strongest baseline by 1.6 percentage points on System Optimization tasks and 4.9 percentage points on long-horizon Model Development & CUDA tasks. Notably, AIM reaches the best baseline performance up to 3.1x faster in wall-clock time. We further provide a theoretical analysis of when searching over ideas becomes beneficial. Our analysis shows that explicit idea-level allocation makes semantic coverage directly controllable, and that broader coverage becomes increasingly valuable when competitive research directions are sparse among many plausible alternatives. Project Page: https://imhgchoi.github.io/agentic-idea-manager/

cs.AI↗

arXiv2026

Dual-Branch Vector-Quantization-Aided Satellite Digital Semantic Communication with Index Compression for High-Resolution RSI Over AFDM

High-resolution remote sensing imagery (RSI) transmission is constrained by satellite-ground bandwidth and channel impairments, yet existing methods struggle to simultaneously achieve extreme compression and robust transmission. To address this, we propose a dual-branch vector-quantization aided satellite digital semantic communication (DVQ-SDSC) framework for RSI transmission over affine frequency division multiplexing (AFDM), whose bandwidth savings arise from two interrelated aspects. First, at the source-coding level, a dual-branch framework is developed to unify deep joint semantic coding, VQ-aided index transmission, channel estimation and adaption in an end-to-end architecture; departing from symmetric encoder designs, the codec is recast as an asymmetric dual-branch architecture that separately processes the high-frequency residuals and the low-frequency structural semantics, with gated fusion and channel-adaptive reconstruction jointly restoring the semantic content. Second, at the index-coding level, we develop a principal component analysis (PCA)-aided codebook reordering to align index topology with latent correlations, and devise group differential pulse-code modulation (G-DPCM) to encode prediction residuals rather than absolute indices, lowering the index bitrate while locally isolating clipping and channel errors. A two-stage training strategy further decouples channel impairments from the semantic codec. FAIR1M experiments over 3GPP NTN-TDL-D demonstrate that DVQ-SDSC with G-DPCM index coding outperforms the conventional JPEG-LDPC scheme at the base rate of 0.0625 bits per pixel (BPP), and that G-DPCM applies directly to the trained codec without retraining, yielding additional index compression at no extra cost.

cs.IT↗

arXiv2026

ScientistTwo: Pioneering the Human Knowledge Frontier with Autonomous AI

Scientific discovery is defined by the ability to identify the boundaries of existing knowledge and venture into unexplored territory. The ultimate vision for AI in science is problem-driven autonomous research: given a fundamental challenge by a human expert, the AI independently navigates the scientific landscape, uncovers theoretical and empirical bottlenecks, and systematically expands the frontier of knowledge. In this paper, we introduce ScientistTwo, a fully autonomous multi-agent framework designed to realize this vision. Specifically, ScientistTwo takes an initial problem as input, establishes state-of-the-art baselines, formulates novel hypotheses, and coordinates specialized agents to orchestrate an end-to-end discovery cycle without human intervention. Moreover, the framework rigorously conducts experiments using diverse datasets and metrics, refines methodologies through automated ablation studies, and validates research findings via a closed-loop simulated peer-review rebuttal engine. To evaluate ScientistTwo's capabilities against the highest standards of human scientific achievement, we benchmark it across papers accepted at top-tier conferences such as ICLR, ICML, and NeurIPS. As a result, ScientistTwo autonomously generates expert-level, publishable papers and fully verified, executable codebases. Its solutions consistently outperform human state-of-the-art models, and achieve higher average review ratings than human-authored papers under automated AI review agents. These results show that ScientistTwo is not merely an assistive tool but an autonomous scientific pioneer capable of pushing the frontiers of human discovery. Project website: https://scientist-two.github.io/

cs.AI↗

arXiv2026

PaperBanana-Interact: Scientific Diagram Refinement with Multi-Turn Human Feedback

Recent efforts have aimed to automate scientific diagram generation from paper content (Lin et al., 2026; Zhu et al., 2026a). However, fully satisfying an author's visual and communicative preferences in a single turn is challenging: in our formative user study (N = 14), all participants requested further revisions after viewing an initial draft, and 86% of them rated the refined diagrams as more satisfactory. Despite the clear demand, the multi-turn workflow remains largely underexplored. To bridge this gap, we present MTPaperBananaBench, a benchmark for multi-turn diagram generation containing 292 images annotated with 3,518 user requirements. To reduce expensive human studies and enable scalable benchmarking, we construct a user simulator that, at each turn, identifies unsatisfied requirements and converts k of them into natural language feedback. Evaluating both requirement satisfaction and overall diagram quality reveals two key failure modes shared across baseline multiturn systems: (1) quality drift, where diagram quality progressively declines over turns, and (2) forgetting, where previously implemented features are lost in subsequent turns. To address these issues, we introduce PaperBanana-Interact, a multi-agent system that refines diagrams via an internal critique-and-refine loop. PaperBanana-Interact consistently improves rather than degrades diagram quality across turns, outperforming baselines by 11.9-18.6 points in quality score and reducing forgetting by 3.7-6.2 points.

cs.CL↗

arXiv2026

Towards Semantic Internet of Everything in the Age of Agentic AI

Semantic communication improves task effectiveness by transmitting task-relevant information. However, most existing schemes remain organized as task-specific, end-to-end pipelines, which are difficult to reuse across models, applications, and deployment environments. Against this background, we propose the Semantic Internet of Everything (SIoE), a composable service architecture that represents heterogeneous communication and artificial intelligence (AI) functions as capability-profiled services and coordinates them according to application objectives. SIoE comprises three planes: a task and service plane, an agentic orchestration plane, and a semantic capability plane. In this framework, task requirements are captured via a semantic service-level agreement (SLA), while an agentic planner discovers and composes candidate capabilities under deterministic compatibility, resource, privacy, and policy validation. Feedback from the communication, semantic, and task levels enables continuous adaptation and replanning. A lightweight vehicle-to-everything case study illustrates profile-grounded capability planning under explicit service constraints. The results demonstrate the feasibility of decoupling service objectives from fixed communication implementations and also highlight key open challenges, including semantic SLA design, capability interoperability, scalable planning, and trustworthy execution.

eess.SP↗

arXiv2026

QV-PIC: Query-Aware Visual Position-Independent Caching for Efficient RAG Serving

Retrieval-Augmented Generation (RAG) repeatedly prefills identical text chunks across queries, incurring redundant computations. Position-Independent Caching (PIC) mitigates it by reusing precomputed Key-Value (KV) across positions, but its efficiency is constrained by the large volume of text tokens. Rendering text chunks as images can compress the text into fewer visual tokens, but the rendered-image PIC suffers more severe quality degradation than the text PIC. This representation-specific gap primarily arises from contextual mismatches across independently compiled caches and the loss of fine-grained textual evidence during visual compression. Existing PIC repair methods mainly address the former through selective recomputation, but they incur online computation and cannot recover lost textual details. We propose QV-PIC, a query-aware dual-resolution PIC reuse framework guided by model-native templates. Offline, QV-PIC compiles visual caches under the model's native chat-template prefix, improving PIC quality without online recomputation. Online, it preserves global context with low resolution and restores fine-grained textual evidence within a high-resolution budget by cumulative query relevance scores, retaining the efficiency benefit of visual compression. Across six tasks, QV-PIC improves average F1 by 21.6 points over vanilla rendered-image PIC, closes the gap to vanilla text PIC, and surpasses optimized text PIC by 2.58 F1 while reducing TTFT by 17.2\%. Relative to full prefill, it cuts TTFT by 83.8%.

cs.CL↗

arXiv2026

MMEB-V3: Measuring the Performance Gaps of Omni-Modality Embedding Models

Multimodal embedding models aim to map heterogeneous inputs, such as text, images, videos, and audio, into a shared semantic space. However, existing methods and benchmarks remain largely limited to partial modality coverage, making it difficult to systematically evaluate full-modality representation learning. In this work, we take a step toward the full-modality setting. We introduce MMEB-V3, a comprehensive benchmark that evaluates embeddings across text, image, video, audio, as well as agent-centric scenarios. To enable more fine-grained diagnosis, we further construct OmniSET (Omni-modality Semantic Equivalence Tuples), where semantically equivalent instances are represented across modalities, allowing us to disentangle semantic similarity from modality effects. Through experiments on MMEB-V3, we conduct a systematic analysis of full-modality embeddings and identify three key findings: (1) models often fail to retrieve the intended target modality; (2) cross-modal retrieval is highly asymmetric and dominated by query-modality bias; and (3) instruction-induced shifts are either insufficient or misaligned with the target modality, and therefore do not reliably improve retrieval. These results indicate that current multimodal embeddings are not yet capable of reliably enforcing modality constraints specified by instructions, and consequently fail to exhibit consistent modality-aware retrieval behavior. We hope MMEB-V3 provides a useful benchmark for understanding and diagnosing these limitations, and for guiding future research on full-modality embeddings.

cs.IR↗

arXiv2026

Why Does Feedback-Augmented Self-Distillation Fail to Improve Retrieval-Interleaved Search Agents?

On-policy self-distillation (OPSD) offers a promising approach for training large language models without relying on a separate teacher model. However, its effectiveness on complex agentic tasks remains largely unexplored. In this work, we instantiate Feedback-Augmented Self-Distillation (FA-SD), a self-distillation algorithm for agentic search that leverages successful demonstrations as privileged information. We identify that models can rely on recurring reasoning-and-search output templates, producing trajectories that appear diverse but are largely agnostic to the input question, making the KL-based self-distillation signal uninformative. We term this phenomenon decoding collapse, a failure mode that can be missed by existing evaluation metrics. To understand its underlying cause, we show that although the self-teacher achieves stronger performance, learning remains inherently unstable due to inconsistent supervision signals. We further decompose this inconsistency into model inconsistency and prompt inconsistency, and show that the latter can significantly degrade the quality of the supervision signal, limiting the effectiveness of self-teacher learning. To mitigate this inconsistency, we introduce an exponential moving average (EMA) teacher to stabilize the self-teacher and provide more consistent supervision signals. Although the EMA teacher requires a warm-up phase during which performance may temporarily regress, it ultimately improves model performance by providing more stable supervision.

cs.AI↗

arXiv2026

SemDPLA: Semantic Communication-based Distributed Physical-Layer Authentication for 6G-enabled Dense IoT

With the rapid development of 6G, increasingly dense device connectivity imposes more strict requirements on multi-users Physical-Layer Authentication (PLA). Compared with cryptography-based methods, PLA enables lightweight authentication by using the uniqueness of wireless channels. However, existing PLA schemes in dense wireless scenarios often suffer from weak fingerprint discriminability and limited computation and communication resources. To address these challenges, we propose a Semantic Communication-based Distributed PLA (SemDPLA) framework. The framework constructs fused central semantic Channel State Information (CSI) fingerprints by fusing semantic information from a central node and multiple distributed nodes. Specifically, we introduce semantic communication to reduce the impact of low Signal Noise Ratio (SNR) and the consumption of communication resource during the transmission from distributed nodes to central node. Furthermore, we propose an ArcFace-based classification method and a semantic fingerprint-oriented distributed voting consistency mechanism to enhance device classification accuracy. Simulation results demonstrate that the proposed SemDPLA scheme performs better than single-node authentication, decision fusion, raw-CSI transmission, and feature fusion baselines. It achieves equal error rates (EERs) of 8.6% at 0 dB and 3.4% at 20 dB. It also achieves classification accuracies of 91.4% at 0 dB and above 95.8% from 5 to 20 dB. Moreover, SemDPLA is robust in low SNR environment and under attacks of abnormal nodes.

eess.SP↗

arXiv2026

Semantic-based Internet of Embodied Intelligence: Visions and Frontiers

Recent advances in generative artificial intelligence (AI) and embodied intelligence (EI) enable autonomous agents to interact with the physical world. However, scaling these systems into networks of multiple agents, namely the Internet of EI (IoEI), faces critical bottlenecks. These include the overhead of massive multimodal data transmission and the decoupling of logical reasoning from physical constraints. To address these challenges, we envision the Semantic-based IoEI (SIoEI), which leverages semantic information as a unified metric throughout the agent lifecycle. We systematically define four key dimensions of EI: perception, intelligence, control, and communication. We further elaborate how semantic empowerment revolutionizes environmental perception, cognition and task planning, action generation and robust control, and communication and networking. We also present a case study to verify that, the semantic-empowered end-to-end process significantly improves channel robustness and reduces end-to-end latency for EI. Finally, we outline critical open research directions for the SIoEI paradigm.

eess.SP↗

arXiv2026

CompRank: Efficient LLM Reranking via Token-Level Compression and Decoding-Free Scoring

Large language model (LLM) rerankers have become an important component of modern retrieval and retrieval-augmented generation pipelines, but their high computational cost limits their applicability to long candidate lists. In this paper, we propose \textbf{CompRank}, a token-efficient reranking framework that reduces redundant computation by aligning reranker design with the sparsity of ranking signals. CompRank decouples document representations from candidate order and query context, enabling reusable document-side states; applies segment-wise token compression to reduce query--document interaction cost; and introduces a CopyNet-style objective that directly aligns attention-based document scoring with training supervision. Experiments on seven BEIR datasets show that CompRank achieves strong reranking performance while retaining only 10.2\% of document tokens, reaching an average NDCG@10 of 39.2 compared with 39.7 under full-token attention. Further scaling experiments on TREC-COVID show that CompRank remains stable when evaluated on candidate lists of up to 500 documents after training on 30-document lists, while achieving $4.9\times$--$9.5\times$ end-to-end speedup over generation-based listwise reranking and approximately $1.3\times$ speedup over the full-token CompRank variant. These results suggest that token-level compression and decoding-free attention scoring provide an effective path toward scalable LLM-based reranking.

cs.IR↗

arXiv2026

Gemini Embedding 2: A Native Multimodal Embedding Model from Gemini

We introduce Gemini Embedding 2, a native multimodal embedding model that allows embedding video, audio, image, and text modalities in a unified representation space. We leverage the multimodal capabilities of Gemini to produce embeddings for arbitrary combinations of interleaved inputs across all these modalities that generalize well across a wide variety of tasks. Applying large-scale contrastive learning in a multi-task multi-stage training setup, we achieve state-of-the-art performance on key embedding benchmarks including unimodal, cross-modal, and multimodal retrieval spanning a diverse set of tasks. We show that our embedding model demonstrates strong performance (with a score of 62.9 R@1 on MSCOCO, 68.8 NDCG@10 on Vatex, 69.9 on MTEB multilingual and 84.0 on MTEB Code) across a variety of tasks surpassing the performance of specialized models. These unified capabilities make Gemini Embedding 2 a promising candidate for downstream use cases such as RAG, recommendation and search. Furthermore, its robust zero-shot performance across distinct fields - from astronomy and bioscience to fine arts and the culinary arts - establishes it as a highly reliable, out-of-the-box representation even for specialized domains.

cs.CV↗

arXiv2026

ScientistOne: Towards Human-Level Autonomous Research via Chain-of-Evidence

Autonomous research agents produce competitive solutions and professional-looking manuscripts, yet their outputs contain verifiability failures undetectable by surface-level evaluation: fabricated citations, unreproducible scores, and method descriptions that diverge from the implementation. We address this through three contributions. First, Chain-of-Evidence (CoE), a verifiability framework requiring every claim to be traceable to its evidence source. Second, ScientistOne, an end-to-end autonomous research system that maintains evidence chains by construction throughout literature review, solution discovery, and paper writing. Third, CoE Audit, a post-hoc audit whose four integrity checks -- score verification, specification violation, reference verification, and method-code alignment -- apply uniformly to all systems. Across 75 papers spanning five systems and five frontier research tasks, every baseline exhibits at least one systematic failure mode: hallucinated reference rates reach 21%, score verification passes in as few as 42% of papers, and method-code alignment ranges from 20% to 80%. ScientistOne achieves zero hallucinated references (0/337), perfect score verification (12/12), and the highest method-code alignment (14/15), while matching or exceeding human expert performance on all five tasks. ScientistOne further generalizes to six additional tasks spanning medical imaging, fine-grained recognition, 3D perception, and language modeling, achieving state-of-the-art on Parameter Golf and gold medals on MLE-Bench tasks where baselines fail entirely.

cs.AI↗