Xue Liu

Retrieval-augmented generation (RAG) is increasingly deployed in enterprise search and document-centric assistants, where responses must be grounded in long and complex source materials. In practice, verifying that generated answers faithfully reflect retrieved documents is difficult: large language models can check long contexts but are too slow and costly for interactive services, while lightweight classifiers operate within strict context limits and frequently miss evidence outside truncated passages. We present the design of a real-time verification component integrated into a production RAG pipeline that enables full-document grounding under latency constraints. The system processes documents up to 32K tokens and employs adaptive inference strategies to balance response time and verification coverage across workloads. We describe the architectural decisions, operational trade-offs, and evaluation methodology used to deploy the verifier, and show that full-context verification substantially improves detection of unsupported responses compared with truncated validation. Our experience highlights when long-context verification is necessary, why chunk-based checking often fails in real documents, and how latency budgets shape model design. These findings provide practical guidance for practitioners building reliable large-scale retrieval-augmented applications.

User-generated 3D content plays a crucial role in enabling truly immersive experiences within the Web3 metaverse. Although significant progress has been made in dynamic 3D content creation and animation using multi-camera setups, they typically require professional equipment and strict synchronization. In contrast, a dynamic and inclusive Web3 ecosystem is collectively built by digital assets from anyone, at any time, and from anywhere. Most Web3 participants are therefore decentralized and non-professional, and their contributed videos, even of the same scene, are usually captured asynchronously using handheld devices. In this paper, we present GSAlign, a novel framework that achieves high-fidelity dynamic 3D reconstruction from unstructured, user-generated videos in the Web3 metaverse. To address the temporal and spatial misalignments inherent in such data, GSAlign integrates three key modules: Epipolar-guided Temporal Alignment (ETA), Motion-centric Separated Reconstruction (MSR), and Background-guided Global Pose Alignment (BGPA). We detail the design of each module and their integration toward a practical end-to-end implementation. Our evaluation of GSAlign on real-world user-generated videos demonstrates robust reconstruction of dynamic 3D scenes despite unsynchronized captures, sparse views, and handheld camera motion.

Multimedia data has become the dominant traffic through the mobile network. In recent years, the forms of videos become more immersive and interactive, raising significant challenges for video streaming due to the limited network resources. Since a viewer’s viewport is limited, video frame partition and tile-based streaming stand out as a feasible solution. Existing works all follow the Tiling-on-Frame (ToF) strategy, i.e., partition each frame into tiles before encoding. However, there is a challenging trade-off. Coarse tiling leads to redundant content transmission, while fine tiling reduces compression ratio due to loss of spatial information. ToF inevitably struggles to balance the tiling granularity and compression efficiency. To tackle this challenge, we introduce a pixel-level framework called Tiling-on-Bitstream (ToB). It eliminates the redundant contents and achieves high compression efficiency. The key idea is that, instead of ToF, we first encode the entire video into bitstreams and then select the corresponding bitstream as requested for streaming; and at the receiver side, the bitstream is decoded to retrieve the requested video content. Such ToB framework enables a fine-grained partition at pixel level, avoiding redundancy in boundary. To efficiently implement a prototype of ToB, we design Neural-EGDL, a neural network-based codec algorithm. To demonstrate the power of our design, we also apply it to 360° videos.

Seeking a transformative memory scheme that grows in performance and capacity as the infrastructure expands.

This paper introduces Jami, a pioneering mobile communication platform that significantly departs from conventional online communication norms. Unlike traditional platforms that depend on centralized servers and intermediaries, Jami is constructed on a distributed architecture that harnesses the power of peer-to-peer (P2P) technology. This innovative design enables secure, end-to-end encrypted communication, ensuring that users maintain full control over their data without reliance on third-party entities. By prioritizing decentralization, Jami significantly enhances user privacy and security while democratizing the communication process, empowering individuals with unparalleled autonomy over their interactions. As an officially recognized GNU package, Jami upholds the highest standards of transparency, freedom, and trust, embodying the principles of open-source development. It fosters a collaborative, community-oriented development model, ensuring continuous innovation and adaptability. This paper delves into Jami's architectural intricacies and real-world deployment experiences, showcasing how its novel design challenges conventional paradigms and highlights its transformative potential to reshape the future landscape of digital communication. Through this exploration, the study aims to contribute to a deeper understanding of mobile distributed systems and their pivotal role in advancing secure, equitable, and user-focused technological solutions.

Large language models (LLMs) have shown strong potential across a variety of tasks, but their application in the telecom field remains challenging due to domain complexity, evolving standards, and specialized terminology. Therefore, general-domain LLMs may struggle to provide accurate and reliable outputs in this context, leading to increased hallucinations and reduced utility in telecom operations. To address these limitations, this work introduces KG-RAG—a novel framework that integrates knowledge graphs (KGs) with retrieval-augmented generation (RAG) to enhance LLMs for telecom-specific tasks. In particular, the KG provides a structured representation of domain knowledge derived from telecom standards and technical documents, while RAG enables dynamic retrieval of relevant facts to ground the model’s outputs. Such a combination improves factual accuracy, reduces hallucination, and ensures compliance with telecom specifications. Experimental results across benchmark datasets demonstrate that KG-RAG outperforms both LLM-only and standard RAG baselines, e.g., KG-RAG achieves an average accuracy improvement of 14.3% over RAG and 21.6% over LLM-only models. These results highlight KG-RAG’s effectiveness in producing accurate, reliable, and explainable outputs in complex telecom scenarios.