arXiv subjects

Yichao Jin

Publications and source records attributed to Yichao Jin.

18 recordsLinked to original sources

arXiv2026

Perception, Layout, and Validation: Calibrated Confidence for Reliable Straight-Through Processing of Financial Documents

Straight-through processing (STP) on extracted key-value fields from financial documents without human review requires a calibrated probability together with a bounded guarantee on the residual error of the auto-approved tier. The emergence of modern Vision Language Models (VLMs) provides an out-of-the-box capability for extracting the key-values, but their verbalized confidence signals are unreliable and weakly track field correctness. This paper introduces a decomposed confidence layer along three interpretable channels, including perception, layout, and validation. Together with a final conformal risk control, the score can be used for reliable STP of financial documents. The method is validated on three public datasets covering real invoices, synthetic invoices, and ad-buy forms, using two different VLM families (Qwen3.6-27B and Gemini-3.1-Flash-Lite). Our decomposed score consistently improves the separation of correct from incorrect extractions, substantially raising the AUROC from 0.54-0.74 for VLM verbalized signals to 0.90-0.99 with contributions from all three designed channels. Crucially for industrial deployment, this enables usable STP. The native VLM confidence signals could clear only 0.1%-7.0% of fields under risk control at a target error of <10%. In contrast, the proposed method auto-approves 49-72% of fields while holding the empirical error of the accepted tier at or below the target.

cs.AI↗

arXiv2026

Detection, Attribution, Narration: An End-to-End Pipeline for Explainable Money Mule Identification

Money mule accounts are critical facilitators of financial fraud, yet detecting them at scale remains challenging due to the heterogeneous nature of transactional and behavioural data. We present an end-to-end pipeline for customer-level mule detection comprising three stages: (1) a LightGBM classifier trained on 280 engineered features spanning transaction patterns, account demographics, network topology, and temporal behaviour; (2) a TreeSHAP attribution layer that decomposes each prediction into feature contributions; and (3) a large language model (LLM) module that converts SHAP attributions into analyst-facing natural-language narratives. We evaluate across three open-weight LLM families and assess explanation quality through analyst feedback. In a live production deployment, the system achieves a yield rate of 89%, up from 61% under the incumbent rule-based system, with monthly alert volume expanding from 211 to 302, reflecting broader true-positive coverage rather than increased noise. This corresponds to a 60% incremental adverse detection beyond existing review workflows, substantially outperforming the rule-based approach. Qualitative feedback from analysts indicates that LLM-generated narratives reduce cognitive load during alert triage. We further discuss implications of deploying LLM-augmented explainability in regulated financial environments.

cs.CR↗

arXiv2026

A RAG-Enhanced Bi-Level Cognitive Orchestration Framework for LEO Satellite Networks

The rapid growth of remote sensing data in Low Earth Orbit (LEO) satellite networks is increasingly constrained by limited downlink capacity to terrestrial networks. Satellite edge computing alleviates this pressure by enabling in-orbit data processing. However, it introduces a new challenge of spatio-temporal resource fragmentation. Variations in onboard computing capability, constrained energy availability, and intermittent inter-satellite and satellite-ground connectivity lead to highly dynamic and uneven resource distribution, which degrades the performance of conventional static routing and scheduling approaches. To address this, we propose a Retrieval-Augmented Generation (RAG)-enhanced bi-level cognitive orchestration framework for knowledge-guided, multi-objective scheduling. The proposed framework explicitly decouples network control across two different operational scales: at the strategic upper level, a Large Language Model (LLM) leverages an offline-distilled Expert Knowledge Base (EKB) to dynamically infer preference weights based on a compact abstract-state descriptor of real-time network conditions. At the lower execution level, a fidelity-aware genetic scheduler utilizes these inferred weights to compute physically feasible, collision-free joint routing and task offloading schedules. Extensive evaluations on a high-fidelity Walker-Delta network testbed under mixed-criticality workloads demonstrate that the proposed framework effectively consolidates fragmented resources, achieving a 30.7% reduction in packet loss, a 30% improvement in energy efficiency over the most competitive learning-based baseline, and an 8.5% decrease in end-to-end latency, while maintaining robust performance under cascading node-failure scenarios.

cs.DC↗

arXiv2026

A Multistage Extraction Pipeline for Long Scanned Financial Documents: An Empirical Study in Industrial KYC Workflows

Structured information extraction from long, multilingual scanned financial documents is a core requirement in industrial KYC and compliance workflows. These documents are typically non machine readable, noisy, and visually heterogeneous. They usually span dozens of pages while containing only sparse task relevant information. Although recent vision-language models achieve strong benchmark performance, directly applying them end to end to full financial reports often leads to unreliable extraction under real world conditions. We present a multistage extraction framework that integrates image preprocessing, multilingual OCR, hybrid page-level retrieval, and compact VLM-based structured extraction. The design separates page localization from multimodal reasoning, enabling more accurate extraction from complex multipage documents. We evaluated the framework on 120 production KYC documents comprising about 3000 multilingual scanned pages. Across multiple OCR-VLM combinations, the proposed pipeline consistently outperforms direct PDF-to-VLM baselines, improving field-level accuracy by up to 31.9 percentage points. The best configuration, PaddleOCR with MiniCPM2.6, achieves 87.27 percent accuracy. Ablation studies show that page-level retrieval is the dominant factor in performance improvements, particularly for complex financial statements and non-English documents.

cs.CV↗

arXiv2026

Goal-oriented Semantic Communication for Robot Arm Reconstruction in Digital Twin: Feature and Temporal Selections

As one of the most promising technologies in industry, the Digital Twin (DT) facilitates real-time monitoring and predictive analysis for real-world systems by precisely reconstructing virtual replicas of physical entities. However, this reconstruction faces unprecedented challenges due to the everincreasing communication overhead, especially for digital robot arm reconstruction. To this end, we propose a novel goal-oriented semantic communication (GSC) framework to extract the GSC information for the robot arm reconstruction task in the DT, with the aim of minimising the communication load under the strict and relaxed reconstruction error constraints. Unlike the traditional reconstruction framework that periodically transmits a reconstruction message for real-time DT reconstruction, our framework implements a feature selection (FS) algorithm to extract the semantic information from the reconstruction message, and a deep reinforcement learning-based temporal selection algorithm to selectively transmit the semantic information over time. We validate our proposed GSC framework through both Pybullet simulations and lab experiments based on the Franka Research 3 robot arm. For a range of distinct robotic tasks, simulation results show that our framework can reduce the communication load by at least 59.5% under strict reconstruction error constraints and 80% under relaxed reconstruction error constraints, compared with traditional communication framework. Also, experimental results confirm the effectiveness of our framework, where the communication load is reduced by 53% in strict constraint case and 74% in relaxed constraint case. The demo is available at: https://youtu.be/2OdeHKxcgnk.

cs.RO↗

arXiv2025

Multi-Stage Field Extraction of Financial Documents with OCR and Compact Vision-Language Models

Financial documents are essential sources of information for regulators, auditors, and financial institutions, particularly for assessing the wealth and compliance of Small and Medium-sized Businesses. However, SMB documents are often difficult to parse. They are rarely born digital and instead are distributed as scanned images that are none machine readable. The scans themselves are low in resolution, affected by skew or rotation, and often contain noisy backgrounds. These documents also tend to be heterogeneous, mixing narratives, tables, figures, and multilingual content within the same report. Such characteristics pose major challenges for automated information extraction, especially when relying on end to end large Vision Language Models, which are computationally expensive, sensitive to noise, and slow when applied to files with hundreds of pages. We propose a multistage pipeline that leverages traditional image processing models and OCR extraction, together with compact VLMs for structured field extraction of large-scale financial documents. Our approach begins with image pre-processing, including segmentation, orientation detection, and size normalization. Multilingual OCR is then applied to recover page-level text. Upon analyzing the text information, pages are retrieved for coherent sections. Finally, compact VLMs are operated within these narrowed-down scopes to extract structured financial indicators. Our approach is evaluated using an internal corpus of multi-lingual, scanned financial documents. The results demonstrate that compact VLMs, together with a multistage pipeline, achieves 8.8 times higher field level accuracy relative to directly feeding the whole document into large VLMs, only at 0.7 percent of the GPU cost and 92.6 percent less end-to-end service latency.

cs.IR↗

arXiv2024

Modeling and Analysis of Multi-Line Orders in Multi-Tote Storage and Retrieval Autonomous Mobile Robot Systems

As warehouses are emphasizing space utilization and the ability to handle multi-line orders, multi-tote storage and retrieval (MTSR) autonomous mobile robot systems, where robots directly retrieve totes from high shelves, are becoming increasingly popular. This paper presents a novel shared-token, multi-class, semi-open queueing network model to account for multi-line orders with general distribution forms in MTSR systems. The numerical results obtained from solving the SOQN model are validated against discrete-event simulation, with most key performance metrics demonstrating high accuracy. In our experimental setting, results indicate a 12.5% reduction in the minimum number of robots needed to satisfy a specific order arrival rate using the closest retrieval sequence policy compared with the random policy. Increasing the number of tote buffer positions on a robot can greatly reduce the number of robots required in the warehouse.

cs.RO↗

arXiv2024

Polarization and quantum entanglement effects in $B^\pm_c\to J/ψ+π^\pm +π^0$ process

Motivated by the very recent observation of the $B^+_c\to J/ψ+π^+ +π^0$ decay using proton-proton collision data by the LHCb collaboration, we study the four-body angular distributions and the quantum entanglement effects in the $B^+_c\to J/ψ+π^+ +π^0$ associated with $J/ψ\to μ^++μ^-$. The helicity angular distributions are given in the QCD effective theory and the von Neumann entropy is obtained in $B^\pm_c\to J/ψ(\to μ^+μ^-)+ρ^\pm(\to π^\pm π^0)$ decay process.

hep-ph↗

arXiv2024

UMBRELLA: A One-stop Shop Bridging the Gap from Lab to Real-World IoT Experimentation

UMBRELLA is an open, large-scale IoT ecosystem deployed across South Gloucestershire, UK. It is intended to accelerate innovation across multiple technology domains. UMBRELLA is built to bridge the gap between existing specialised testbeds and address holistically real-world technological challenges in a System-of-Systems (SoS) fashion. UMBRELLA provides open access to real-world devices and infrastructure, enabling researchers and the industry to evaluate solutions for Smart Cities, Robotics, Wireless Communications, Edge Intelligence, and more. Key features include over 200 multi-sensor nodes installed on public infrastructure, a robotics arena with 20 mobile robots, a 5G network-in-a-box solution, and a unified backend platform for management, control and secure user access. The heterogeneity of hardware components, including diverse sensors, communication interfaces, and GPU-enabled edge devices, coupled with tools like digital twins, allows for comprehensive experimentation and benchmarking of innovative solutions not viable in lab environments. This paper provides a comprehensive overview of UMBRELLA's multi-domain architecture and capabilities, making it an ideal playground for Internet of Things (IoT) and Industrial IoT (IIoT) innovation. It discusses the challenges in designing, developing and operating UMBRELLA as an open, sustainable testbed and shares lessons learned to guide similar future initiatives. With its unique openness, heterogeneity, realism and tools, UMBRELLA aims to continue accelerating cutting-edge technology research, development and translation into real-world progress.

cs.NI↗

arXiv2022

A Dataset of Images of Public Streetlights with Operational Monitoring using Computer Vision Techniques

A dataset of street light images is presented. Our dataset consists of $\sim350\textrm{k}$ images, taken from 140 UMBRELLA nodes installed in the South Gloucestershire region in the UK. Each UMBRELLA node is installed on the pole of a lamppost and is equipped with a Raspberry Pi Camera Module v1 facing upwards towards the sky and lamppost light bulb. Each node collects an image at hourly intervals for 24h every day. The data collection spans for a period of six months. Each image taken is logged as a single entry in the dataset along with the Global Positioning System (GPS) coordinates of the lamppost. All entries in the dataset have been post-processed and labelled based on the operation of the lamppost, i.e., whether the lamppost is switched ON or OFF. The dataset can be used to train deep neural networks and generate pre-trained models providing feature representations for smart city CCTV applications, smart weather detection algorithms, or street infrastructure monitoring. The dataset can be found at \url{https://doi.org/10.5281/zenodo.6046758}.

cs.CV↗

arXiv2022

Securing Synchronous Flooding Communications: An Atomic-SDN Implementation

Synchronous Flooding (SF) protocols can enhance the wireless connectivity between Internet of Things (IoT) devices. However, existing SF solutions fail to introduce sufficient security measures due to strict time synchronisation requirements, making them vulnerable to malicious actions. Our paper presents a design paradigm for encrypted SF communications. We describe a mechanism for synchronising encryption parameters in a network-wide fashion. Our solution operates with minimal overhead and without compromising communication reliability. Evaluating our paradigm on a real-world, large-scale IoT testbed, we have proven that a communication layer impervious to a range of attacks is established without sacrificing the network performance.

cs.NI↗

arXiv2022

Reliable IoT Firmware Updates: A Large-scale Mesh Network Performance Investigation

Internet of Things (IoT) networks require regular firmware updates to ensure enhanced security and stability. As we move towards methodologies of codifying security and policy decisions and exchanging them over IoT large-scale deployments (security-as-a-code), these demands should be considered a routine operation. However, rolling out firmware updates to large-scale networks presents a crucial challenge for constrained wireless environments with large numbers of IoT devices. This paper initially investigates how the current state-of-the-art protocols operate in such adverse conditions by measuring various Quality-of-Service (QoS) Key Performance Indicators (KPIs) of the shared wireless medium. We later discuss how Concurrent Transmissions (CT) can extend the scalability of IoT protocols and ensure reliable firmware roll-outs over large geographical areas. Measuring KPIs such as the mesh join time, the throughput, and the number of nodes forming a network, we provide great insight into how an IoT environment will behave under a large-scale firmware roll-out. Finally, we conducted our performance investigation over the UMBRELLA platform, a real-world IoT testbed deployed in Bristol, UK. This ensures our findings represent a realistic IoT scenario and meet the strict QoS requirements of today's IoT applications.

cs.NI↗

arXiv2021

Recursive Periodicity Shifting for Semi-Persistent Scheduling of Time-Sensitive Communication in 5G

Various legacy and emerging industrial control applications create the requirement of periodic and time-sensitive communication (TSC) for 5G/6G networks. State-of-the-art semi-persistent scheduling (SPS) techniques fall short of meeting the requirements of this type of critical traffic due to periodicity misalignment between assignments and arriving packets that lead to significant waiting delays. To tackle this challenge, we develop a novel recursive periodicity shifting (RPS)-SPS scheme that provides an optimal scheduling policy by recursively aligning the period of assignments until the timing mismatch is minimized. RPS can be realized in 5G wireless networks with minimal modifications to the scheduling framework. Performance evaluation shows the effectiveness of the proposed scheme in terms of minimizing misalignment delay with arbitrary traffic periodicity.

cs.NI↗

arXiv2021

A Self-Configurable Grouping Method for Integrated Wi-SUN FAN and TSCH-based Networks

Recent applications in large-scale wireless mesh networks (WSN), e.g., Advanced Metering Infrastructure (AMI) scenarios, expect to support an extended number of nodes with higher throughput, which cannot be sufficiently supported by the current WSN protocols. Two prior protocols, Wi-SUN Field Area Network (Wi-SUN FAN) and IETF 6TiSCH standards, are popularly used that are respectively based on asynchronous Carrier Sense Multiple Access / Collision Avoidance (CSMA/CA) and IEEE 802.15.4 Time Scheduled Channel Hopping (TSCH). However, the former one with CSMA/CA can be prone to the Hidden Node Problem (HNP) that leads to a degradation of reliability, while the latter one with TSCH avoids HNP by using synchronously scheduling but requires massive control signalling that degrades the upper-bound of throughput. Accordingly, this paper tackles the challenge of how to improve the upper-bound of throughput without loss of reliability. To do so, we first present an in-depth evaluation of reliability and throughput between the two existing standards via system-level simulation. We then propose a self-configurable grouping (SCG) method to cluster nodes without using location information of each node. This SCG method eliminates around 99\% HNP in CSMA/CA, thus greatly improves its reliability with maintaining the relatively high upper-bound of throughput. Our results show that the SCG method almost doubles the network throughput compared with both 6TiSCH and Wi-SUN FAN in heavy traffic scenarios, while providing extremely high reliability of more than 99.999\%.

cs.NI↗

arXiv2020

6TiSCH++ with Bluetooth 5 and Concurrent Transmissions

Targeting dependable communications for industrial Internet of Things applications, IETF 6TiSCH provides mechanisms for efficient scheduling, routing, and forwarding of IPv6 traffic across low-power mesh networks. Yet, despite an overwhelming body of literature covering both centralized and distributed scheduling schemes for 6TiSCH, an effective control solution for large-scale multi-hop mesh networks remains an open challenge. This paper addresses this with a novel approach that eliminates much of the routing and link-layer overhead incurred by centralized schedulers, and provides a robust mechanism for data dissemination synchronization within 6TiSCH. Specifically, we leverage the physical layer (PHY) switching capabilities of modern low-power wireless platforms to build on recent work demonstrating the viability of Concurrent Transmission (CT)-based flooding protocols across the Bluetooth 5 (BT 5) PHYs. By switching the PHY and MAC layer at runtime, we inject a BT 5-based CT flood within a standard IEEE 802.15.4 TSCH slotframe, thereby providing a reliable, low-latency scheme for 6TiSCH control messaging. We present an analytical model and experimental evaluation showing how our solution not only exploits the BT 5 high data-rate PHY layers for rapid data dissemination, but can also provide reliable 6TiSCH association and synchronization even under external radio interference. We further discuss how the proposed technique can be used to address other open challenges within the standard.

cs.NI↗

arXiv2020

Hysia: Serving DNN-Based Video-to-Retail Applications in Cloud

Combining \underline{v}ideo streaming and online \underline{r}etailing (V2R) has been a growing trend recently. In this paper, we provide practitioners and researchers in multimedia with a cloud-based platform named Hysia for easy development and deployment of V2R applications. The system consists of: 1) a back-end infrastructure providing optimized V2R related services including data engine, model repository, model serving and content matching; and 2) an application layer which enables rapid V2R application prototyping. Hysia addresses industry and academic needs in large-scale multimedia by: 1) seamlessly integrating state-of-the-art libraries including NVIDIA video SDK, Facebook faiss, and gRPC; 2) efficiently utilizing GPU computation; and 3) allowing developers to bind new models easily to meet the rapidly changing deep learning (DL) techniques. On top of that, we implement an orchestrator for further optimizing DL model serving performance. Hysia has been released as an open source project on GitHub, and attracted considerable attention. We have published Hysia to DockerHub as an official image for seamless integration and deployment in current cloud environments.

cs.MM↗

arXiv2017

Towards Wi-Fi AP-Assisted Content Prefetching for On-Demand TV Series: A Reinforcement Learning Approach

The emergence of smart Wi-Fi APs (Access Point), which are equipped with huge storage space, opens a new research area on how to utilize these resources at the edge network to improve users' quality of experience (QoE) (e.g., a short startup delay and smooth playback). One important research interest in this area is content prefetching, which predicts and accurately fetches contents ahead of users' requests to shift the traffic away during peak periods. However, in practice, the different video watching patterns among users, and the varying network connection status lead to the time-varying server load, which eventually makes the content prefetching problem challenging. To understand this challenge, this paper first performs a large-scale measurement study on users' AP connection and TV series watching patterns using real-traces. Then, based on the obtained insights, we formulate the content prefetching problem as a Markov Decision Process (MDP). The objective is to strike a balance between the increased prefetching&storage cost incurred by incorrect prediction and the reduced content download delay because of successful prediction. A learning-based approach is proposed to solve this problem and another three algorithms are adopted as baselines. In particular, first, we investigate the performance lower bound by using a random algorithm, and the upper bound by using an ideal offline approach. Then, we present a heuristic algorithm as another baseline. Finally, we design a reinforcement learning algorithm that is more practical to work in the online manner. Through extensive trace-based experiments, we demonstrate the performance gain of our design. Remarkably, our learning-based algorithm achieves a better precision and hit ratio (e.g., 80%) with about 70% (resp. 50%) cost saving compared to the random (resp. heuristic) algorithm.

cs.MM↗

arXiv2014

Enhancing User Experience for Multi-Screen Social TV Streaming over Wireless Networks

Recently, multi-screen cloud social TV is invented to transform TV into social experience. People watching the same content on social TV may come from different locations, while freely interact with each other through text, image, audio and video. This crucial virtual living-room experience adds social aspects into existing performance metrics. In this paper, we parse social TV user experience into three elements (i.e., inter-user delay, video quality of experience (QoE), and resource efficiency), and provide a joint analytical framework to enhance user experience. Specifically, we propose a cloud-based optimal playback rate allocation scheme to maximize the overall QoE while upper bounding inter-user delay. Experiment results show that our algorithm achieves near-optimal tradeoff between inter-user delay and video quality, and demonstrates resilient performance even under very fast wireless channel fading.

cs.MM↗