Search arXiv⌕ Search

arXiv subjects

Mengting Li

Publications and source records attributed to Mengting Li.

At least 19 recordsLinked to original sources
arXiv2026

From Solvers to Research: Large Language Model-Driven Formal Mathematics at the Research Frontier

Recent developments in AI for Mathematics (AI4Math), especially Large Language Model (LLM)-driven theorem provers, has achieved remarkable success in formal proof generation for well-defined mathematical problems through Interactive Theorem Proving (ITP) languages. However, current systems remain fundamentally limited in tackling frontier research mathematics, such as discovering new theorems or resolving open conjectures, which are often open-ended, under-specified, and involve multiple layers of abstraction. We argue that the next leap in AI4Math systems requires a decisive shift from predefined problem-solvers to research agents that can address frontier mathematical challenges with rigorous formal mathematical reasoning. In this position paper, we provide a systematic review of the field, covering datasets, auto-formalization, and proof synthesis. More importantly, we identify core limitations of existing systems in serving as mathematical research agents, examining issues across datasets, relational structure, mathematical exploration, tool ecosystem, and human-AI collaboration, outlining a strategic road-map for the future of AI4Math.

cs.CL↗
arXiv2026

OneLatent: Latent Reasoning for Efficient Foundation Recommendation Models

Large language models (LLMs) have demonstrated strong reasoning capabilities, motivating their use as the backbone of foundation recommendation models (FRMs). Existing methods enhance recommendations through explicit Chain-of-Thought (CoT) reasoning under a Think-then-Answer paradigm. However, explicit CoT incurs substantial inference overhead by generating lengthy reasoning traces and relies on manually designed templates that struggle to capture diverse, dynamic user interests. We propose OneLatent, an efficient latent reasoning framework that compresses explicit reasoning traces into several learnable latent tokens, enabling Latent-Reason-then-Answer inference without generating verbose traces. OneLatent first introduces Multi-View Adaptive CoT (MV-ACoT), which creates diverse, high-quality teacher-generated supervision by exploring user interests from multiple perspectives and automatically adapting reasoning complexity to each instance. Building on pretrained FRMs, it then uses a three-stage latent-token alignment paradigm to progressively internalize CoT traces into learnable latent tokens. Finally, a multistage curriculum-based post-training strategy activates latent-token reasoning for downstream recommendation tasks. Experiments on an industrial-scale Kuaishou dataset and the public Kuaishou LLM-Rec benchmark show that OneLatent consistently outperforms explicit CoT-based methods and traditional baselines. Compared with the Think and No-Think variants of FRMs, OneLatent improves SID@64 by 17.44% and 9.33%, respectively, while achieving over 17x higher online inference throughput. We further develop a production serving system for scalable, real-time FRM inference. An online A/B test in Kuaishou's local-services advertising scenario shows that deploying OneLatent with this system yields an estimated 9.6% revenue lift over strong online baselines, including OneRec and OneReason.

cs.IR↗
arXiv2026

MemRole: From Memories to Decisions for Long-Horizon Reasoning in Role-Playing Agents

LLM-based role-playing agents are fluent at imitating a character's style, yet struggle to make role-consistent decisions over long narrative horizons. Such decisions are shaped by experiences spanning many prior scenes, which are often partial and may point in different directions. Existing methods inject cross-scene memories as context, but do not reason about how past events shape a character's current motivations and behavior. We propose MemRole, a framework that makes this memory-to-decision process explicit. It introduces Long-Horizon Memory Thinking (LHMT), a reasoning schema that decomposes hidden reasoning into context understanding, memory activation, and response strategy planning, trained on a novel-derived dataset whose character-grounded, timeline-constrained memory pools limit each character to what it could plausibly know at the current narrative point. Beyond supervised fine-tuning, we propose Self-Refining On-Policy Self-Distillation (SR-OPSD): the model critiques its own rollouts against the reference under a data-driven rubric, and a critique-conditioned teacher view of the same model distills token-level guidance back into the policy. MemRole-32B improves over its backbone by 15.2 points on memory-grounded role-playing, performs on par with frontier LLMs such as GPT-5.4, and achieves the best character-motivation recognition accuracy on CroSS-MR.

cs.CL↗
arXiv2026

EvoRec: A Unified Self-Evolving Agentic Recommender with Dynamic Reflection Trees

Conventional recommender systems rely on fixed cascaded pipelines, limiting their ability to adapt to evolving user needs and feedback. Existing agentic recommenders introduce reasoning and tool use but often lack end-to-end decision control and systematic failure reflection. We propose EvoRec, a unified agentic recommendation framework that combines autonomous decision-making with reflection-driven self-evolution. A Need Agent infers user needs from long-term, short-term, and working memory. A Planning Agent adaptively allocates retrieval budgets between Behavior Continuation and Need Satisfaction, selects retrieval tools, and sets ranking-fusion weights. An Acting Executor executes the plan and records decision traces. Following user feedback, a Dynamic Reflection Tree identifies failures in need inference, planning, or previous repairs, proposes personalized corrections, and validates them through replay. Verified repairs update working memory and are distilled into reusable cross-user strategies, enabling continual improvement without parameter updates. Experiments on a Kuaishou local-services advertising dataset show a 15.7% Hit@50 improvement over the strongest baseline. Hierarchical reflection further improves Hit@50 by 13.1% among users with prior repairs. In a seven-day online A/B test, nearline EvoRec increases advertising revenue by 4.6% and CTCVR by 3.5%, with a 14.5% revenue gain for high-intent users. These results demonstrate the effectiveness of reflection-driven self-evolution in industrial-scale recommendation.

cs.IR↗
arXiv2026

Modality Disentangled Learning for Incomplete Multimodal Emotion Recognition: A Primitive Memory Distillation Perspective

Multimodal Emotion Recognition (MER) systems often suffer from missing modalities in real-world scenarios. Existing methods usually generate, align, or distill missing modalities as a whole, overlooking the heterogeneous nature of the information carried by each modality. Such holistic treatment mixes inferable shared semantics with uncertain modality-specific details, yielding unstable representations and degrading robustness. To address this issue, we propose the Primitive Memory Distillation (PriMD) framework. Unlike existing methods, PriMD takes an intra-modal perspective and focuses on how different types of information within a modality differ in recoverability. PriMD first disentangles cross-modal shared semantics from modality-specific representations, and then discretizes the latter into learnable semantic primitives to construct modality-specific memory banks. When modalities are missing, PriMD is a teacher-student framework that the student model uses the shared semantics of available modalities as queries to dynamically retrieve primitives. It compensates for missing modality-specific information within a constrained memory space and aligns with the teacher model. Extensive experiments on IEMOCAP, CMU-MOSI, and CMU-MOSEI demonstrate that PriMD achieves state-of-the-art performance and consistently stronger robustness across a wide range of missing-modality settings, while mitigating the instability caused by holistic feature inference. Our code and project website are available at https://www.jiaqizhang-sengoku.com/Code/PriMD and https://www.jiaqizhang-sengoku.com/PriMD/ , respectively.

cs.CV↗
arXiv2026

WhisperRec: Latent Reasoning for Efficient Foundation Recommendation Models

Large language models (LLMs) have demonstrated strong reasoning capabilities, motivating their adoption as backbones for foundation recommendation models (FRMs). Existing approaches typically enhance recommendation with explicit Chain-of-Thought (CoT) under the Think-then-Answer paradigm. However, generating lengthy rationales introduces substantial inference overhead, while fixed CoT templates struggle to model diverse, dynamic, and context-dependent user interests. We propose WhisperRec, an efficient latent reasoning framework for FRMs. WhisperRec compresses teacher-generated CoT into learnable latent reasoning tokens, enabling a Latent-Reason-then-Answer paradigm that performs reasoning in latent space without producing verbose rationales. This design retains decision-relevant reasoning information while avoiding the latency bottleneck of autoregressive rationale generation. Specifically, it first introduces Multi-View Adaptive CoT (MV-ACoT) to construct diverse, high-quality supervision from complementary perspectives on user interests. MV-ACoT also adapts reasoning complexity to each instance, applying lightweight analysis to clear cases and targeted multi-factor reasoning to challenging ones. Building on a pre-trained FRM, WhisperRec then employs a three-stage Latent Reasoning Alignment procedure to progressively internalize teacher CoT into latent representations. Finally, curriculum-based post-training activates latent-token reasoning for downstream recommendation while preserving standard recommendation capability. Experiments on an industrial-scale Kuaishou dataset and the public Kuaishou LLM-Rec benchmark show that WhisperRec consistently outperforms explicit-CoT methods and conventional baselines. Compared with explicit CoT Think and No-Think variants, WhisperRec improves SID@64 by 17.44% and 9.33%, respectively, and achieves over 10x higher online inference throughput.

cs.IR↗
arXiv2026

AECG: Asymmetric Experience Consolidation and Governance In Multi-Agent Systems

Large language model (LLM)-based multi-agent systems increasingly rely on memory to transform execution trajectories into reusable procedural knowledge. Yet repeated retrieval also makes memory errors persistent: memory pollution arises when outdated, weakly supported, or spuriously successful procedures become recurring components of future reasoning. Multi-agent execution introduces an additional structural risk. Scope collapse occurs when procedural knowledge escapes the coordination scope in which it was shown effective and is repeatedly reused at incompatible decision levels, allowing local errors to influence cascades of downstream decisions. Meanwhile, task-level failures provide ambiguous supervision because they rarely reveal which recalled knowledge was responsible. We introduce AECG, a framework for asymmetric experience consolidation and governance for multi-agent systems. AECG turns memory from static experience storage into a dynamic reliability-governance loop, preserving coordination scope and using multi-scale, confidence-aware reliability to detect degradation. It then combines degradation with downstream impact to prioritize high-risk knowledge under a bounded review budget, applies targeted interventions, and reactivates revised skills only after paired replay. Across three multi-agent frameworks and four benchmarks, AECG achieves the best score in 11 of 12 framework--benchmark settings and improves over the strongest competing memory method by as much as 10.23 percentage points; removing scope preservation reduces accuracy by up to 16.89 points. AECG thereby reframes multi-agent memory from passive accumulation into auditable reliability governance. Code is available at https://github.com/fenhg297/AECG

cs.AI↗
arXiv2026

TTS-Guard: Black-Box Ownership Verification of Text-to-Speech Models via Adaptive Adversarial Speaker-Pair Fingerprints

The rapid maturation of zero-shot Text-to-Speech (TTS) models has turned high-quality voice cloning into a widely available capability, raising acute concerns over unauthorised replication, fine-tuning and resale of proprietary speech models. Yet ownership verification for TTS remains largely open: speech is a continuous waveform whose perturbations are easily destroyed by routine signal processing, and the human auditory system imposes a much tighter perceptual budget than vision. We present \textbf{TTS-Guard}, a black-box ownership verification framework for TTS models built on \emph{adversarial speaker-pair fingerprints}. TTS-Guard(i) selects key speaker pairs in a \emph{dual} embedding space for architecture-agnostic stealth;(ii) optimises a perturbation through an \emph{adaptive curriculum} of shadow models covering fine-tuning, pruning, quantisation and distillation; and (iii) aggregates black-box queries into a calibrated \emph{Verification Confidence Score}. On five mainstream TTS systems, TTS-Guard reaches an average Fingerprint Success Rate of $96.4\%$ at a False Positive Rate of $5.8\%$, while preserving intelligibility and naturalness. The fingerprint remains effective against ten audio attacks, six model modifications, and two state-of-the-art adversarial purifiers.

cs.SD↗
arXiv2026

On-Site Beam Calibration for RIS-Aided Positioning Systems

High precision positioning is a key enabler for next-generation communication applications such as smart transportation and augmented reality. Reconfigurable intelligent surface (RIS) technology can enhance positioning by providing additional angular information and improving coverage under obstructed propagation conditions. However, true RIS beams can differ significantly from the simplified or ideal beam response models commonly used in RIS-aided positioning, leading to beam model mismatch and an elevated positioning error floor. This paper proposes an on-site RIS beam calibration framework that reduces this error floor by estimating a realistic 3D RIS beam response model from on-site measurements. The proposed calibration algorithm first extracts the RIS-reflected channel response from signals received by a calibration agent sampling the angular range of interest, using delay-domain sparse recovery, and then estimates the beam model parameters with a gradient-based estimator. To validate the proposed framework, 3D beam patterns under 66 phase modulations were measured and incorporated into simulations. With an angular sampling step of 1 deg, the calibrated model achieves an average beam response similarity of 88.5% with respect to the ground truth, compared with 43.7% for the ideal model. The probability that the absolute lower bound of the positioning error is below 0.5m increases from 0.52 without calibration to 0.74 after calibration, showing that on-site RIS beam calibration effectively reduces the positioning error floor caused by true beam model mismatch.

eess.SP↗
arXiv2026

DREAM: LLM-based Dynamic Role-playing via Event-Aware Memory Graph

Role-playing agents (RPAs) have emerged as a key application of large language models, enabling immersive and high-fidelity character simulation. Accurate role-playing of established characters requires not only stylistic imitation but also temporally consistent and causally grounded behavioral reasoning. However, existing RPAs primarily rely on static character descriptions and unstructured memory, limiting their ability to maintain long-term narrative and personality coherence. We introduce DREAM, a structured memory framework for role-playing agents inspired by the Activating Event-Belief-Consequence (ABC) cognitive model. DREAM transforms unstructured literary text into an Event-aware Memory Graph (EMG) that organizes character experiences into temporally ordered and causally linked event graph. This representation enables the construction of dynamic, dual-granularity character profiles that capture both stable personality traits and event-driven behavioral evolution. We further propose the Temporal Causal Memory (TCM) benchmark to evaluate temporal consistency and long-range causal narrative coherence. DREAM achieves state-of-the-art performance across CoSER, LIFECHOICE, and TCM, outperforming multiple strong baselines. Our approach demonstrates the effectiveness of structured memory in enhancing the interpretability and consistency of role-playing agents.

cs.CL↗
arXiv2026

Dynamic Generation of Multi-LLM Agents Communication Topologies with Graph Diffusion Models

The efficiency of multi-agent systems driven by large language models (LLMs) largely hinges on their communication topology. However, designing an optimal topology is a non-trivial challenge, as it requires balancing competing objectives such as task performance, communication cost, and robustness. Existing frameworks often rely on static or hand-crafted topologies, which inherently fail to adapt to diverse task requirements, leading to either excessive token consumption for simple problems or performance bottlenecks for complex ones. To address this challenge, we introduce a novel generative framework called \textit{Guided Topology Diffusion (GTD)}. Inspired by conditional discrete graph diffusion models, GTD formulates topology synthesis as an iterative construction process. At each step, the generation is steered by a lightweight proxy model that predicts multi-objective rewards (e.g., accuracy, utility, cost), enabling real-time, gradient-free optimization towards task-adaptive topologies. This iterative, guided synthesis process distinguishes GTD from single-step generative frameworks, enabling it to better navigate complex design trade-offs. We validated GTD across multiple benchmarks, and experiments show that this framework can generate highly task-adaptive, sparse, and efficient communication topologies, significantly outperforming existing methods in LLM agent collaboration.

cs.CL↗
arXiv2026

Mismatch Analysis and Cooperative Calibration of Array Beam Patterns for ISAC Systems

Integrated sensing and communication (ISAC) is a key technology for enabling a wide range of applications in future wireless systems. However, the sensing performance is often degraded by model mismatches caused by geometric errors (e.g., position and orientation) and hardware impairments (e.g., mutual coupling and amplifier non-linearity). This paper focuses on the angle estimation performance with antenna arrays and tackles the critical challenge of array beam pattern calibration for ISAC systems. To assess calibration quality from a sensing perspective, a novel performance metric that accounts for angle estimation error, rather than beam pattern similarity, is proposed and incorporated into a differentiable loss function. Additionally, a cooperative calibration framework is introduced, allowing multiple user equipments to iteratively optimize the beam pattern based on the proposed loss functions and local data, and collaboratively update global calibration parameters. The proposed models and algorithms are validated using real-world beam pattern measurements collected in an anechoic chamber. Experimental results show that the angle estimation error can be reduced from {$\textbf{1.01}^\circ$} to $\textbf{0.11}^\circ$ in 2D calibration scenarios, and from $\textbf{5.19}^\circ$ to $\textbf{0.86}^\circ$ in 3D calibration ones.

eess.SP↗
arXiv2025

RIS Beam Calibration for ISAC Systems: Modeling and Performance Analysis

High-accuracy localization is a key enabler for integrated sensing and communication (ISAC), playing an essential role in various applications such as autonomous driving. Antenna arrays and reconfigurable intelligent surface (RIS) are incorporated into these systems to achieve high angular resolution, assisting in the localization process. However, array and RIS beam patterns in practice often deviate from the idealized models used for algorithm design, leading to significant degradation in positioning accuracy. This mismatch highlights the need for beam calibration to bridge the gap between theoretical models and real-world hardware behavior. In this paper, we present and analyze three beam models considering several key non-idealities such as mutual coupling, non-ideal codebook, and measurement uncertainties. Based on the models, we then develop calibration algorithms to estimate the model parameters that can be used for future localization tasks. This work evaluates the effectiveness of the beam models and the calibration algorithms using both theoretical bounds and real-world beam pattern data from an RIS prototype. The simulation results show that the model incorporating combined impacts can accurately reconstruct measured beam patterns. This highlights the necessity of realistic beam modeling and calibration to achieve high-accuracy localization.

eess.SP↗
arXiv2025

Pilot-Based End-to-End Radio Positioning and Mapping for ISAC: Beyond Point-Based Landmarks

Integrated sensing and communication enables simultaneous communication and sensing tasks, including precise radio positioning and mapping, essential for future 6G networks. Current methods typically model environmental landmarks as isolated incidence points or small reflection areas, lacking detailed attributes essential for advanced environmental interpretation. This paper addresses these limitations by developing an end-to-end cooperative uplink framework involving multiple base stations and users. Our method uniquely estimates extended landmark objects and incorporates obstruction-based outlier removal to mitigate multi-bounce signal effects. Validation using realistic ray-tracing data demonstrates substantial improvements in the richness of the estimated environmental map.

eess.SP↗
arXiv2025

Joint Near-Field Sensing and Visibility Region Detection with Extremely Large Aperture Arrays

In this paper, we consider near-field localization and sensing with an extremely large aperture array under partial blockage of array antennas, where spherical wavefront and spatial non-stationarity are accounted for. We propose an Ising model to characterize the clustered sparsity feature of the blockage pattern, develop an algorithm based on alternating optimization for joint channel parameter estimation and visibility region detection, and further estimate the locations of the user and environmental scatterers. The simulation results confirm the effectiveness of the proposed algorithm compared to conventional methods.

eess.SP↗
arXiv2025

RIS-Aided Positioning Under Adverse Conditions: Interference from Unauthorized RIS

Positioning technology, which aims to determine the geometric information of a device in a global coordinate, is a key component in integrated sensing and communication systems. In addition to traditional active anchor-based positioning systems, reconfigurable intelligent surfaces (RIS) have shown great potential for enhancing system performance. However, their ability to manipulate electromagnetic waves and ease of deployment pose potential risks, as unauthorized RIS may be intentionally introduced to jeopardize the positioning service. Such an unauthorized RIS can cause unexpected interference in the original localization system, distorting the transmitted signals, and leading to degraded positioning accuracy. In this work, we investigate the scenario of RIS-aided positioning in the presence of interference from an unauthorized RIS. Theoretical lower bounds are employed to analyze the impact of unauthorized RIS on channel parameter estimation and positioning accuracy. Several codebook design strategies for unauthorized RIS are evaluated, and various system arrangements are discussed. The simulation results show that an unauthorized RIS path with a high channel gain or a delay similar to that of legitimate RIS paths leads to poor positioning performance. Furthermore, unauthorized RIS generates more effective interference when using directional beamforming codebooks compared to random codebooks.

eess.SP↗
arXiv2025

Optimizing Global Genomic Surveillance for Early Detection of Emerging SARS-CoV-2 Variants

Background: Global viral threats underscore the need for effective genomic surveillance, but high costs and uneven resource distribution hamper its implementation. Targeting surveillance to international travelers in major travel hubs may offer a more efficient strategy for the early detection of SARS-CoV-2 variants. Methods: We developed and calibrated a multiple-strain metapopulation model of global SARS-CoV-2 transmission using extensive epidemiological, phylogenetic, and high-resolution air travel data. We then compared baseline surveillance with various resource-allocation approaches that prioritize travelers, focusing on Omicron BA.1/BA.2 retrospectively and on hypothetical future variants under different emergence, transmission and vaccine effectiveness scenarios. Findings: Focusing existing surveillance resources on travelers at key global hubs significantly shortened detection delays without increasing total surveillance efforts. In retrospective analyses of Omicron BA.1/BA.2, traveler-targeted approaches consistently outperformed baseline strategies, even when overall resources were reduced. Simulations indicate that focusing surveillance on key travel hubs outperform baseline practices in detecting future variants, across different possible origins, even with reduced resources. This approach also remains effective in future pandemic scenarios with varying reproductive numbers and vaccine effectiveness. Interpretation: These findings provide a quantitative, cost-effective framework for strengthening global genomic surveillance. By reallocating resources toward international travelers in select travel hubs, early detection of emerging variants can be enhanced, informing rapid public health interventions and bolstering preparedness for future pandemics.

q-bio.PE↗
arXiv2024

Multiple-Frequency-Bands Channel Characterization for In-vehicle Wireless Networks

In-vehicle wireless networks are crucial for advancing smart transportation systems and enhancing interaction among vehicles and their occupants. However, there are limited studies in the current state of the art that investigate the in-vehicle channel characteristics in multiple frequency bands. In this paper, we present measurement campaigns conducted in a van and a car across below 7 GHz, millimeter-wave (mmWave), and sub-Terahertz (Sub-THz) bands. These campaigns aim to compare the channel characteristics for in-vehicle scenarios across various frequency bands. Channel impulse responses (CIRs) were measured at various locations distributed across the engine compartment of both the van and car. The CIR results reveal a high similarity in the delay properties between frequency bands below 7GHz and mmWave bands for the measurements in the engine bay. Sparse channels can be observed at Sub-THz bands in the engine bay scenarios. Channel spatial profiles in the passenger cabin of both the van and car are obtained by the directional scan sounding scheme for three bands. We compare the power angle delay profiles (PADPs) measured at different frequency bands in two line of sight (LOS) scenarios and one non-LOS (NLOS) scenario. Some major \added{multipath components (MPCs)} can be identified in all frequency bands and their trajectories are traced based on the geometry of the vehicles. The angular spread of arrival is also calculated for three scenarios. The analysis of channel characteristics in this paper can enhance our understanding of in-vehicle channels and foster the evolution of in-vehicle wireless networks.

eess.SP↗