arXiv subjects

Rui Lang

Publications and source records attributed to Rui Lang.

4 recordsLinked to original sources

arXiv2026

Replacing Large Language Models with Jev Decision Models for Low-Latency Edge Service Orchestration

Natural-language service requests can require a language-model decision before execution starts, consuming part of the request's latency budget. We integrate Jev's decision-oriented application programming interface (API) into edge service orchestration to reduce this overhead while retaining service completion. The integration extracts four bounded intent fields and applies a shared validator, admission policy, and scheduler, accounting for decision waiting throughout the request timeline. We compare Jev with a short, structured-output DeepSeek deployment using live API measurements followed by modeled execution, and then a real two-node optical character recognition (OCR) service with self-hosted Qwen and rule-based references. Across three consecutive measurement blocks, Jev reduces median client decision latency by 15.9-26.5%. In eight paired OCR conditions, Jev matches DeepSeek's correct, on-time completion count in seven and exceeds it in one. Without caching, median end-to-end latency on requests completed correctly by both systems is 11.1-25.3% lower; API fees per correct completion are 69.0-70.6% lower. Repeated-request caching largely removes the latency difference. The results demonstrate that decision-model substitution can lower both response latency and API fees in the tested service path, and identify fresh interpretation as the main opportunity for latency savings.

cs.DC↗

arXiv2026

Fast Intent-Driven Service Orchestration with Jev for 6G Edge Networks

Intent-driven services envisioned for sixth-generation (6G) edge networks must translate changing requirements into executable contracts while wireless requests continue to arrive. Interpretation consumes part of the same deadline budget as transmission, queueing, and execution. We evaluate Jev as a decision model for this stage, asking whether faster contract activation improves service timeliness while preserving interpretation quality. Contracts specify permitted execution locations, deadlines, and priorities; a numerical scheduler uses them to allocate edge work. The evaluation combines live Jev, DeepSeek, Gemini, and self-hosted Qwen responses with packet-level New Radio simulation, mobility events, shared edge queues, and a real image-reading service. Cached interpretation with numerical scheduling substantially improves completion over direct model-selected placement. With correct interpretation on all evaluated contracts, Jev reduces median decision latency by 22.4% against DeepSeek and 61.9% against Gemini; completion in the modeled update scenarios rises by 3.50 and 8.35 percentage points. A separate comparison with direct-attribute Qwen retains a 53.0% latency reduction and a 4.78-point completion gain. Exchanging recorded model and scheduling waits reproduces the comparator completion summaries for all 12 trajectories in each of two modeled comparisons. In the real image service, Jev retains faster decisions and completes 459 requests correctly and on time, compared with 463 for DeepSeek and 435 for Qwen out of 1,080 each. These findings connect fast intent decisions to timely edge service execution and support using Jev at contract activation, with numerical scheduling adapting placements to current network and computing state.

cs.NI↗

arXiv2026

PlanTwin: Privacy-Preserving Planning Abstractions for Cloud-Assisted LLM Agents

Cloud-hosted large language models (LLMs) have become the de facto planners in agentic systems, coordinating tools and guiding execution over local environments. In many deployments, however, the environment being planned over is private, containing source code, files, credentials, and metadata that cannot be exposed to the cloud. Existing solutions address adjacent concerns, such as execution isolation, access control, or confidential inference, but they do not control what cloud planners observe during planning: within the permitted scope, \textit{raw environment state is still exposed}. We introduce PlanTwin, a privacy-preserving architecture for cloud-assisted planning without exposing raw local context. The key idea is to project the real environment into a \textit{planning-oriented digital twin}: a schema-constrained and de-identified abstract graph that preserves planning-relevant structure while removing reconstructable details. The cloud planner operates solely on this sanitized twin through a bounded capability interface, while a local gatekeeper enforces safety policies and cumulative disclosure budgets. We further formalize the privacy-utility trade-off as a capability granularity problem, define architectural privacy goals using $(k,δ)$-anonymity and $ε$-unlinkability, and mitigate compositional leakage through multi-turn disclosure control. We implement PlanTwin as middleware between local agents and cloud planners and evaluate it on 60 agentic tasks across ten domains with four cloud planners. PlanTwin achieves full sensitive-item non-disclosure (SND = 1.0) while maintaining planning quality close to full-context systems: three of four planners achieve PQS $> 0.79$, and the full pipeline incurs less than 2.2\% utility loss.

cs.CR↗

arXiv2022

NRTR: Neuron Reconstruction with Transformer from 3D Optical Microscopy Images

The neuron reconstruction from raw Optical Microscopy (OM) image stacks is the basis of neuroscience. Manual annotation and semi-automatic neuron tracing algorithms are time-consuming and inefficient. Existing deep learning neuron reconstruction methods, although demonstrating exemplary performance, greatly demand complex rule-based components. Therefore, a crucial challenge is designing an end-to-end neuron reconstruction method that makes the overall framework simpler and model training easier. We propose a Neuron Reconstruction Transformer (NRTR) that, discarding the complex rule-based components, views neuron reconstruction as a direct set-prediction problem. To the best of our knowledge, NRTR is the first image-to-set deep learning model for end-to-end neuron reconstruction. In experiments using the BigNeuron and VISoR-40 datasets, NRTR achieves excellent neuron reconstruction results for comprehensive benchmarks and outperforms competitive baselines. Results of extensive experiments indicate that NRTR is effective at showing that neuron reconstruction is viewed as a set-prediction problem, which makes end-to-end model training available.

cs.CV↗