arXiv subjects

Yingying He

Publications and source records attributed to Yingying He.

5 recordsLinked to original sources

arXiv2026

RMSWeb: Reflection, Failure-Mode Mining, and Salvage-DS for Web Agent Reinforcement Learning

Compact web agents can reduce deployment cost, but training them poses challenges in both data collection and post-SFT reinforcement learning (RL). Successful trajectories are expensive to collect and often contain inefficient detours. After supervised fine-tuning (SFT), full trajectory corpora are dominated by routine states; moreover, when group-relative RL is applied to web actions, inadequately designed action-level rewards can yield weak or misleading relative updates, while groups rejected as unsuitable for such updates receive no fallback learning signal. We present RMSWeb, a three-part recipe for Qwen3-VL-Instruct at 8B and 32B. Reflection-conditioned retries increase collection yield and shorten successful trajectories; failure-mode mining concentrates offline RL on critical states exposed by the SFT policy; and Salvage-DS combines an action-semantic polarized reward, contrast-and-competence-gated dynamic sampling, and an action-only anchor for rejected groups. Policies trained with reflection-collected data use up to 19.7% fewer action steps on solved tasks. On WebVoyager, Online-Mind2Web, and WebTailBench, RMSWeb improves over SFT by 2.4-7.0 points at 8B and 1.2-7.7 points at 32B. Our 8B model also achieves the strongest reported Online-Mind2Web result among similarly sized open-weight models in our comparison and a leading reported accuracy-cost trade-off on WebVoyager and WebTailBench, with the caveat that external evaluation protocols differ.

cs.AI↗

arXiv2025

Improving Multi-Vehicle Perception Fusion with Millimeter-Wave Radar Assistance

Cooperative perception enables vehicles to share sensor readings and has become a new paradigm to improve driving safety, where the key enabling technology for realizing this vision is to real-time and accurately align and fuse the perceptions. Recent advances to align the views rely on high-density LiDAR data or fine-grained image feature representations, which however fail to meet the requirements of accuracy, real-time, and adaptability for autonomous driving. To this end, we present MMatch, a lightweight system that enables accurate and real-time perception fusion with mmWave radar point clouds. The key insight is that fine-grained spatial information provided by the radar present unique associations with all the vehicles even in two separate views. As a result, by capturing and understanding the unique local and global position of the targets in this association, we can quickly find out all the co-visible vehicles for view alignment. We implement MMatch on both the datasets collected from the CARLA platform and the real-world traffic with over 15,000 radar point cloud pairs. Experimental results show that MMatch achieves decimeter-level accuracy within 59ms, which significantly improves the reliability for autonomous driving.

cs.RO↗

arXiv2024

Semi-supervised learning via DQN for log anomaly detection

Log anomaly detection is a critical component in modern software system security and maintenance, serving as a crucial support and basis for system monitoring, operation, and troubleshooting. It aids operations personnel in timely identification and resolution of issues. However, current methods in log anomaly detection still face challenges such as underutilization of unlabeled data, imbalance between normal and anomaly class data, and high rates of false positives and false negatives, leading to insufficient effectiveness in anomaly recognition. In this study, we propose a semi-supervised log anomaly detection method named DQNLog, which integrates deep reinforcement learning to enhance anomaly detection performance by leveraging a small amount of labeled data and large-scale unlabeled data. To address issues of imbalanced data and insufficient labeling, we design a state transition function biased towards anomalies based on cosine similarity, aiming to capture semantic-similar anomalies rather than favoring the majority class. To enhance the model's capability in learning anomalies, we devise a joint reward function that encourages the model to utilize labeled anomalies and explore unlabeled anomalies, thereby reducing false positives and false negatives. Additionally, to prevent the model from deviating from normal trajectories due to misestimation, we introduce a regularization term in the loss function to ensure the model retains prior knowledge during updates. We evaluate DQNLog on three widely used datasets, demonstrating its ability to effectively utilize large-scale unlabeled data and achieve promising results across all experimental datasets.

cs.SE↗

arXiv2022

FORCE: A Framework of Rule-Based Conversational Recommender System

The conversational recommender systems (CRSs) have received extensive attention in recent years. However, most of the existing works focus on various deep learning models, which are largely limited by the requirement of large-scale human-annotated datasets. Such methods are not able to deal with the cold-start scenarios in industrial products. To alleviate the problem, we propose FORCE, a Framework Of Rule-based Conversational Recommender system that helps developers to quickly build CRS bots by simple configuration. We conduct experiments on two datasets in different languages and domains to verify its effectiveness and usability.

cs.IR↗

arXiv2021

Learning Neural Templates for Recommender Dialogue System

Though recent end-to-end neural models have shown promising progress on Conversational Recommender System (CRS), two key challenges still remain. First, the recommended items cannot be always incorporated into the generated replies precisely and appropriately. Second, only the items mentioned in the training corpus have a chance to be recommended in the conversation. To tackle these challenges, we introduce a novel framework called NTRD for recommender dialogue system that decouples the dialogue generation from the item recommendation. NTRD has two key components, i.e., response template generator and item selector. The former adopts an encoder-decoder model to generate a response template with slot locations tied to target items, while the latter fills in slot locations with the proper items using a sufficient attention mechanism. Our approach combines the strengths of both classical slot filling approaches (that are generally controllable) and modern neural NLG approaches (that are generally more natural and accurate). Extensive experiments on the benchmark ReDial show our NTRD significantly outperforms the previous state-of-the-art methods. Besides, our approach has the unique advantage to produce novel items that do not appear in the training set of dialogue corpus. The code is available at \url{https://github.com/jokieleung/NTRD}.

cs.CL↗