Author name cluster

Yu Qian

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

5 papers

1 author row

EAAI Journal 2026 Journal Article

Real-time railroad crossing surveillance with edge computing

Youzhi Tang
Yu Qian

Accurate and effective detection of anomalies at rail crossings is critical for improving railway safety. However, existing methods often struggle to reliably identify intrusions and threats. To address these challenges, this study introduces You Only Look Once-Region-based Convolutional Neural Network (YOLO-RCNN), a hybrid framework that incorporates YOLO-FG (foreground), a component specifically designed to detect and segment all objects within a scene, capabilities that conventional object detectors lack. The proposed model further integrates an RCNN with the Region of Interest Align (RoIAlign) mechanism, effectively classifying and tracking detected objects. This framework enables precise detecting, classifying, and tracking objects with high computational efficiency, and effectively overcomes the limitations of traditional “classification and tracking by detection” pipelines, significantly reducing computational overhead. Optimized for real-time applications on resource-constrained edge devices, YOLO-RCNN achieves an F-measure of 90. 49 % on the Change Detection Network (CDnet) 2014 dataset and demonstrates its effectiveness on the custom Railroad Crossing Dataset (RCD), achieving a mean average precision (mAP) of 54. 15 %, a Seg mAP of 43. 62 %, and a Higher Order Tracking Accuracy (HOTA) score of 63. 64 %. Deployment optimizations using TensorRT and oneAPI Threading Building Blocks (oneTBB) increased inference speed from 4. 69 Frames Per Second (FPS) to 54. 79 FPS on desktop systems and from 3. 39 FPS to 29. 19 FPS on Jetson AGX Orin, demonstrating its real-time applicability. These results underscore YOLO-RCNN's potential as a robust solution for efficient and reliable railroad crossing monitoring, as well as other real-time surveillance tasks. The RCD dataset is publicly released at: https: //www. kaggle. com/datasets/elvin1233/rail-crossing.

Details DOI

AAMAS Conference 2019 Conference Paper

A Multi-task Selected Learning Approach for Solving 3D Flexible Bin Packing Problem

Lu Duan
Haoyuan Hu
Yu Qian
Yu Gong
Xiaodong Zhang
Jiangwen Wei
Yinghui Xu

A 3D flexible bin packing problem (3D-FBPP) arises from the process of warehouse packing in e-commerce. An online customer’s order usually contains several items and needs to be packed as a whole before shipping. In particular, 5% of tens of millions of packages are using plastic wrapping as outer packaging every day, which brings pressure on the plastic surface minimization to save traditional logistics costs. Because of the huge practical significance, we focus on the issue of packing cuboid-shaped items orthogonally into a least-surface-area bin. The existing heuristic methods for classic 3D bin packing don’t work well for this particular NP-hard problem and designing a good problem-specific heuristic is nontrivial. In this paper, rather than designing heuristics, we propose a novel multi-task framework based on Selected Learning to learn a heuristic-like policy that generates the sequence and orientations of items to be packed simultaneously. Through comprehensive experiments on a large scale real-world transaction order dataset and online AB tests, we show: 1) our selected learning method trades off the imbalance and correlation among the tasks and significantly outperforms the single task Pointer Network and the multi-task network without selected learning; 2) our method obtains an average 5. 47% cost reduction than the well-designed greedy algorithm which is previously used in our online production system.

PDF

AAMAS Conference 2019 Conference Paper

Can Sophisticated Dispatching Strategy Acquired by Reinforcement Learning? : A Case Study in Dynamic Courier Dispatching System

Yujie Chen
Yu Qian
Yichen Yao
Zili Wu
Rongqi Li
Yinzhi Zhou
Haoyuan Hu
Yinghui Xu

In this paper, we study a courier dispatching problem (CDP) raised from an online pickup-service platform of Alibaba. The CDP aims to assign a set of couriers to serve pickup requests with stochastic spatial and temporal arrival rate among urban regions. The objective is to maximize the revenue of served requests given a limited number of couriers over a period of time. Many online algorithms such as dynamic matching and vehicle routing strategy from existing literature could be applied to tackle this problem. However, these methods rely on appropriately predefined optimization objectives at each decision point, which is hard in dynamic situations. This paper formulates the CDP as a Markov decision process (MDP) and proposes a data-driven approach to derive the optimal dispatching rule-set under different scenarios. Our method stacks multi-layer images of the spatial-and-temporal map and apply multi-agent reinforcement learning (MARL) techniques to evolve dispatching models. This method solves the learning inefficiency caused by traditional centralized MDP modeling. Through comprehensive experiments on both artificial dataset and real-world dataset, we show: 1) By utilizing historical data and considering long-term revenue gains, MARL achieves better performance than myopic online algorithms; 2) MARL is able to construct the mapping between complex scenarios to sophisticated decisions such as the dispatching rule. 3) MARL has the scalability to adopt in large-scale real-world scenarios. Proc. of the 18th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2019), N. Agmon, M. E. Taylor, E. Elkind, M. Veloso (eds.), May 13–17, 2019, Montreal, Canada. © 2019 International Foundation for Autonomous Agents and Multiagent Systems (www. ifaamas. org). All rights reserved.

PDF

AAMAS Conference 2016 Conference Paper

Boosting Nonparametric Policies

Yang Yu
Peng-Fei Hou
Qing Da
Yu Qian

Learning complex policies is a key step toward real-world applications of reinforcement learning. While boosting approaches have been widely applied in state-of-the-art supervised learning techniques to adaptively learn nonparametric functions, in reinforcement learning the boosting-style approaches have been little investigated. Only a few pieces of previous work explored this direction, however theoretical properties are still unclear and empirical performance is quite limited. In this paper, we propose the PolicyBoost method. It optimizes a finite-sample objective function, which leads to maximization of the expected total reward, by employing the GradientBoost approach. Experimental results verify the effectiveness as well as the robustness of PolicyBoost, even without feature engineering.

PDF

EAAI Journal 1993 Journal Article

Process modelling and optimization of systems with imprecise and conflicting equations

Yu Qian
Patrick Tessier
Guy A. Dumont

For industrial processes, mechanistic models are not always available due to incomplete knowledge and imprecise descriptions of the phenomena that take place in the process. However, through years of practical operation, empirical knowledge of these processes can be accumulated and represented by a set of imprecise and empirical equations. Unfortunately, these equations may sometimes be redundant and even contradictory. A fuzzy-logic-based modelling and optimization technique is proposed for representing uncertainty and approximation in relationships among process variables. The process model is represented on three levels: heuristic knowledge base, fuzzy equation sets, and fuzzy relation matrix. The model is used to maximize the degree of compatibility and minimize conflicts among the fuzzy equations via a genetic algorithm. This work illustrates a new, important feature of fuzzy modelling: the ability to handle conflict among system equations. The new approach has been applied to fuzzy optimization of pulp quality control of an industrial wood chip refiner.

Details DOI