Vector data trading is essential for cross-domain learning with vector databases, yet it remains largely unexplored. We study this problem under online learning, where sellers face uncertain retrieval costs and buyers provide stochastic feedback to posted prices. Three main challenges arise: (1) heterogeneous and partial feedback in configuration learning, (2) variable and complex feedback in pricing learning, and (3) inherent coupling between configuration and pricing decisions. We propose a hierarchical bandit framework that jointly optimizes retrieval configurations and pricing. Stage I employs contextual clustering with confidence-based exploration to learn effective configurations with logarithmic regret. Stage II adopts interval-based price selection with local Taylor approximation to estimate buyer responses and achieve sublinear regret. We establish theoretical guarantees with polynomial time complexity and validate the framework on four real-world datasets, demonstrating consistent improvements in cumulative reward and regret reduction compared with existing methods.

Natural and efficient interaction remains a critical challenge for virtual reality and augmented reality (VR/AR) systems. Vision-based gesture recognition suffers from high computational cost, sensitivity to lighting conditions, and privacy leakage concerns. Acoustic sensing provides an attractive alternative: by emitting inaudible high-frequency signals and capturing their reflections, channel impulse response (CIR) encodes how gestures perturb the acoustic field in a low-cost and user-transparent manner. However, existing CIR-based gesture recognition methods often rely on extensive training of models on large labeled datasets, making them unsuitable for few-shot VR scenarios. In this work, we propose the first framework that leverages large language models (LLMs) for CIR-based gesture recognition in VR/AR systems. Despite LLMs' strengths, it is non-trivial to achieve few-shot and zero-shot learning of CIR gestures due to their inconspicuous features. To tackle this challenge, we collect differential CIR rather than original CIR data. Moreover, we construct a real-world dataset collected from 10 participants performing 15 gestures across three categories (digits, letters, and shapes), with 10 repetitions each. We then conduct extensive experiments on this dataset using an LLM-adopted classifier. Results show that our LLM-based framework achieves accuracy comparable to classical machine learning baselines, while requiring no domain-specific retraining.

The performance of large language models (LLMs) and large multimodal models (LMMs) depends heavily on the quality and scale of their pre-training datasets. Recent research shows that large multimodal models trained on natural documents where images and text are interleaved outperform those trained only on image-text pairs across a wide range of benchmarks, leveraging advanced pre-trained models to enforce semantic alignment, image-sequence consistency, and textual coherence. For Arabic, however, the lack of high-quality multimodal datasets that preserve document structure has limited progress. In this paper, we present our pipeline Wasm for processing the Common Crawl dataset to create a new Arabic multimodal dataset that uniquely provides markdown output. Unlike existing Arabic corpora that focus solely on text extraction, our approach preserves the structural integrity of web content while maintaining flexibility for both text-only and multimodal pre-training scenarios. We provide a comprehensive comparative analysis of our data processing pipeline against those used for major existing datasets, highlighting the convergences in filtering strategies and justifying our specific design choices. To support future research, we publicly release a representative dataset dump along with the multimodal processing pipeline for Arabic.

This dissertation explores the application of multi-agent reinforcement learning (MARL) for handling deadlocks in intralogistics systems that rely on autonomous mobile robots (AMRs). AMRs enhance operational flexibility but also increase the risk of deadlocks, which degrade system throughput and reliability. Existing approaches often neglect deadlock handling in the planning phase and rely on rigid control rules that cannot adapt to dynamic operational conditions. To address these shortcomings, this work develops a structured methodology for integrating MARL into logistics planning and operational control. It introduces reference models that explicitly consider deadlock-capable multi-agent pathfinding (MAPF) problems, enabling systematic evaluation of MARL strategies. Using grid-based environments and an external simulation software, the study compares traditional deadlock handling strategies with MARL-based solutions, focusing on PPO and IMPALA algorithms under different training and execution modes. Findings reveal that MARL-based strategies, particularly when combined with centralized training and decentralized execution (CTDE), outperform rule-based methods in complex, congested environments. In simpler environments or those with ample spatial freedom, rule-based methods remain competitive due to their lower computational demands. These results highlight that MARL provides a flexible and scalable solution for deadlock handling in dynamic intralogistics scenarios, but requires careful tailoring to the operational context.

As a key medium for human interaction and information exchange, social networking services (SNS) pose unique challenges for large language models (LLMs): heterogeneous workloads, fast-shifting norms and slang, and multilingual, culturally diverse corpora that induce sharp distribution shift. Supervised fine-tuning (SFT) can specialize models but often triggers a ``seesaw'' between in-distribution gains and out-of-distribution robustness, especially for smaller models. To address these challenges, we introduce RedOne 2.0, an SNS-oriented LLM trained with a progressive, RL-prioritized post-training paradigm designed for rapid and stable adaptation. The pipeline consist in three stages: (1) Exploratory Learning on curated SNS corpora to establish initial alignment and identify systematic weaknesses; (2) Targeted Fine-Tuning that selectively applies SFT to the diagnosed gaps while mixing a small fraction of general data to mitigate forgetting; and (3) Refinement Learning that re-applies RL with SNS-centric signals to consolidate improvements and harmonize trade-offs across tasks. Across various tasks spanning three categories, our 4B scale model delivers an average improvements about 2.41 over the 7B sub-optimal baseline. Additionally, RedOne 2.0 achieves average performance lift about 8.74 from the base model with less than half the data required by SFT-centric method RedOne, evidencing superior data efficiency and stability at compact scales. Overall, RedOne 2.0 establishes a competitive, cost-effective baseline for domain-specific LLMs in SNS scenario, advancing capability without sacrificing robustness.

Large Language Models for Simulating Professions (SP-LLMs), particularly as teachers, are pivotal for personalized education. However, ensuring their professional competence and ethical safety is a critical challenge, as existing benchmarks fail to measure role-playing fidelity or address the unique teaching harms inherent in educational scenarios. To address this, we propose EduGuardBench, a dual-component benchmark. It assesses professional fidelity using a Role-playing Fidelity Score (RFS) while diagnosing harms specific to the teaching profession. It also probes safety vulnerabilities using persona-based adversarial prompts targeting both general harms and, particularly, academic misconduct, evaluated with metrics including Attack Success Rate (ASR) and a three-tier Refusal Quality assessment. Our extensive experiments on 14 leading models reveal a stark polarization in performance. While reasoning-oriented models generally show superior fidelity, incompetence remains the dominant failure mode across most models. The adversarial tests uncovered a counterintuitive scaling paradox, where mid-sized models can be the most vulnerable, challenging monotonic safety assumptions. Critically, we identified a powerful Educational Transformation Effect: the safest models excel at converting harmful requests into teachable moments by providing ideal Educational Refusals. This capacity is strongly negatively correlated with ASR, revealing a new dimension of advanced AI safety. EduGuardBench thus provides a reproducible framework that moves beyond siloed knowledge tests toward a holistic assessment of professional, ethical, and pedagogical alignment, uncovering complex dynamics essential for deploying trustworthy AI in education. See https://github.com/YL1N/EduGuardBench for Materials.

Multimodal large language models (MLLMs) have demonstrated remarkable capabilities in vision-language answering tasks. Despite their strengths, these models often encounter challenges in achieving complex reasoning tasks such as mathematical problem-solving. Previous works have focused on fine-tuning on specialized mathematical datasets. However, these datasets are typically distilled directly from teacher models, which capture only static reasoning patterns and leaving substantial gaps compared to student models. This reliance on fixed teacher-derived datasets not only restricts the model's ability to adapt to novel or more intricate questions that extend beyond the confines of the training data, but also lacks the iterative depth needed for robust generalization. To overcome these limitations, we propose \textbf{\method}, a \textbf{Math}ematical \textbf{S}elf-\textbf{E}volving framework for MLLMs. In contrast to traditional one-shot fine-tuning paradigms, \method iteratively refines the model through cycles of inference, reflection, and reward-based feedback. Specifically, we leverage iterative fine-tuning by incorporating correct reasoning paths derived from previous-stage inference and integrating reflections from a specialized Outcome Reward Model (ORM). To verify the effectiveness of \method, we evaluate it on a suite of challenging benchmarks, demonstrating significant performance gains over backbone models. Notably, our experimental results on MathVL-test surpass the leading open-source multimodal mathematical reasoning model QVQ. Our code and models are available at \texttt{https://zheny2751\allowbreak-dotcom.github.io/\allowbreak MathSE.github.io/}.

Query optimization has been studied using machine learning, reinforcement learning, and, more recently, graph-based convolutional networks. Ontology, as a structured, information-rich knowledge representation, can provide context, particularly in learning problems. This paper presents OntoTune, an ontology-based platform for enhancing learning for query optimization. By connecting SQL queries, database metadata, and statistics, the ontology developed in this research is promising in capturing relationships and important determinants of query performance. This research also develops a method to embed ontologies while preserving as much of the relationships and key information as possible, before feeding it into learning algorithms such as tree-based and graph-based convolutional networks. A case study shows how OntoTune's ontology-driven learning delivers performance gains compared with database system default query execution.

General-purpose large language models (LLMs) are increasingly deployed in verticals such as telecommunications, where adaptation is hindered by scarce, low-information-density corpora and tight mobile/edge constraints. We propose Data Trajectory Alignment (DTA), a two-phase, model-agnostic data curation framework that treats solution processes - not only final answers - as first-class supervision. Phase I (Initializing) synthesizes diverse, high-coverage candidates using an ensemble of strong teachers. Phase II (DTA) rewrites teacher solutions to align intermediate steps and presentation style with the target student's inductive biases and then performs signal-aware exemplar selection via agreement checks and reflection-based judging. Instantiated on telecommunications mathematics (e.g., link budgets, SNR/AMC selection, and power-control feasibility), DTA yields state-of-the-art (SOTA) accuracy on TELEMATH without enabling explicit "thinking" modes: 72.45% pass@1, surpassing distilled-only training by +17.65 points and outperforming a strong baseline (Qwen3-32B with thinking enabled) by +2.94 points. Token-shift analyses indicate that DTA concentrates gains on logical-structural discourse markers rather than merely amplifying domain nouns, indicating improved reasoning scaffolding. Under edge-like inference settings, DTA improves efficiency by reducing reliance on multi-sample voting and disabling expensive reasoning heuristics, cutting energy per output token by ~42% versus Qwen3-32B (thinking mode enabled) and end-to-end latency by ~60% versus Qwen3-32B (thinking mode disabled). These results demonstrate that aligning how solutions are produced enables compact, high-yield supervision that is effective for both accuracy and efficiency, offering a practical recipe for domain adaptation in low-resource verticals beyond telecom.

Abstract--Computational spectral imaging (CSI) achieves real-time hyperspectral imaging through co-designed optics and algorithms, but typical CSI methods suffer from a bulky footprint and limited fidelity. Therefore, Spectral Deconvolution imaging (SDI) methods based on PSF engineering have been proposed to achieve high-fidelity compact CSI design recently. However, the composite convolution-integration operations of SDI render the normal-equation coefficient matrix scene-dependent, which hampers the efficient exploitation of imaging priors and poses challenges for accurate reconstruction. T o tackle the inherent data-dependent operators in SDI, we introduce a Hierarchical Spatial-Spectral Aggregation Unfolding Framework (HSF AUF). By decomposing subproblems and projecting them into the frequency domain, HSF AUF transforms nonlinear processes into linear mappings, thereby enabling efficient solutions. Furthermore, to integrate spatial-spectral priors during iterative refinement, we propose a Spatial-Frequency Aggregation Transformer (SF A T), which explicitly aggregates information across spatial and frequency domains. By integrating SF A T into HSF AUF, we develop a Transformer-based deep unfolding method, Hierarchical Spatial-Frequency Aggregation Unfolding Transformer (HSF AUT), to solve the inverse problem of SDI. Systematic simulated and real experiments show that HSF AUT surpasses SOT A methods with cheaper memory and computational costs, while exhibiting optimal performance on different SDI systems. Hyperspectral images (HSIs) capture high-resolution spectra at each spatial location, providing a spectral representation that reveals the rich characteristics of various components and materials, offering a high-dimensional visual capability beyond human vision. Thus, HSIs have found widespread applications in fields such as medical diagnosis [1], remote sensing [2], [3], agricultural inspection [4], and machine vision [5]. However, early hyperspectral imaging techniques were constrained by 2D sensor, requiring spatial or spectral scanning that sacrificed temporal resolution for spectral resolution, restricting their use in dynamic scenes. To overcome these challenges, computational spectral imaging (CSI) [6] integrates optics, electronics, and algorithms to enhance imaging capabilities [7], [8], [9].