Author name cluster

Xiaolong Li

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.

24 papers

2 author rows

AAAI Conference 2026 Conference Paper

A More Efficient Reduction from Outlier-Aware to Outlier-Free k-Median

Zhen Zhang
Han Peng
Limei Liu
Junyu Huang
Xiaolong Li
Qilong Feng

Given a non-negative integer \ell, the k-median with outliers problem extends the standard k-median problem by allowing the removal of up to \ell points and minimizing the clustering cost over the remaining ones. Algorithmic development in this setting remains an active area of research due to its relevance in processing noisy data. In this paper, we present a sampling-based reduction from the k-median with outliers problem to its outlier-free counterpart. The reduction incurs a multiplicative overhead of (kℓ⁻¹ + ε⁻¹)^O(ℓ) in the running time: it yields (kℓ⁻¹ + ε⁻¹)^O(ℓ) outlier-free instances, a solution to one of which can be directly transformed into a solution to the original instance with an arbitrarily small loss in the approximation ratio. This improves upon the previously known reduction with an overhead of ((k + ℓ)ε⁻¹)^O(ℓ). As applications, we obtain faster fixed-parameter tractable (FPT) algorithms with tight approximation guarantees for the k-median with outliers problem under various metric spaces. Furthermore, our approach naturally generalizes to constrained variants of the problem where additional constraints are imposed on the cluster sizes, and yields similar improvements in their FPT approximations.

PDF Details DOI

IROS Conference 2025 Conference Paper

A Multi-Task Learning System for Composites Defect Segmentation and Classification with TacRoller

Xiaolong Li
Tunwu Li
Zhenyu Lu 0001
Chao Zeng 0002
Guangliang Cheng
Chenguang Yang 0001

Due to non-destructive testing (NDT) techniques being both expensive and inconvenient in dynamic detection scenarios, innovative alternatives are urgently needed to address cost-efficiency and deployment challenges. We first design TacRoller, a tactile sensor roller for automated characterization of surface defects in composite materials, to address the dilemma. It collects tactile images of defects on the composite’s plies by capturing changes caused by deformation of the outer elastomer through the internal camera. It reduces the cost of inspection by 80% to 90% compared to NDT equipment like radiographic testing while ensuring detection efficiency. It takes 58. 86 seconds to complete a 35 cm×18 cm × 0. 5 mm dry-woven fabric. Moreover, we collect a total of 2, 744 images of samples of dry-woven fabric unidirectional prepreg through TacRoller to form a dataset, including wrinkles, foreign objects and debris (FODs), broken fibre, voids and healthy textures. Subsequently, we propose a multi-order gated aggregation (MOGA)-U-Net to tackle critical challenges of noise sensitivity and multi-scale defect recognition in tactile images, enabling robust segmentation and multi-category classification tasks. The results show that the MOGA-U-Net achieves a test dice coefficient of 76. 0% and classification accuracy of 98. 9%, outperforming DeepLabV3 and other benchmarks. By providing a scalable and effective NDT substitute, our system realises autonomous defect identification and classification on composites surface, thus improving quality control in the production of composites.

Details

AAAI Conference 2025 Conference Paper

IRMamba: Pixel Difference Mamba with Layer Restoration for Infrared Small Target Detection

Mingjin Zhang
Xiaolong Li
Fei Gao
Jie Guo

Infrared small target detection (IRSTD) focuses on identifying small targets in infrared images. Despite advancements with deep learning, challenges persist due to the IR long-range imaging mechanism, where targets are small, dim, and easily lost in noise and background clutter. Current deep learning methods struggle to suppress noise and background interference while preserving fine details, leading to missed detections and false alarms. To address these issues, we propose IRMamba, an encoder-decoder architecture featuring Pixel Difference Mamba (PDMamba) and a Layer Restoration Module (LRM). Specifically, PDMamba integrates the intensity and directional information of pixel differences between scanning positions and their central neighborhoods into the state equation of the state space model (SSM). This enhances target detail representation and suppresses background interference by capturing local 2D dependencies from a global perspective. In addition, LRM incorporates the double-depth image prior into the iterative convergence algorithm, and utilizes the inter-layer interrelationships to gradually reverse the separation of the target layer, achieving noise suppression and refined reconstruction of the image mask. Experiments conducted on multiple public datasets, including NUAA-SIRST, NUDT-SIRST, and IRSTD-1K, demonstrate the significant advantages of IRMamba over SOTA methods.

PDF Details DOI

NeurIPS Conference 2025 Conference Paper

SPC: Evolving Self-Play Critic via Adversarial Games for LLM Reasoning

Jiaqi Chen
Bang Zhang
Ruotian Ma
Peisong Wang
Xiaodan Liang
Zhaopeng Tu
Xiaolong Li
Kwan-Yee K. Wong

Evaluating the step-by-step reliability of large language model (LLM) reasoning, such as Chain-of-Thought, remains challenging due to the difficulty and cost of obtaining high-quality step-level supervision. In this paper, we introduce Self-Play Critic (SPC), a novel approach where a critic model evolves its ability to assess reasoning steps through adversarial self-play games, eliminating the need for manual step-level annotation. SPC involves fine-tuning two copies of a base model to play two roles, namely a "sneaky generator" that deliberately produces erroneous steps designed to be difficult to detect, and a "critic" that analyzes the correctness of reasoning steps. These two models engage in an adversarial game in which the generator aims to fool the critic, while the critic model seeks to identify the generator's errors. Using reinforcement learning based on the game outcomes, the models iteratively improve; the winner of each confrontation receives a positive reward and the loser receives a negative reward, driving continuous self-evolution. Experiments on three reasoning process benchmarks (ProcessBench, PRM800K, DeltaBench) demonstrate that our SPC progressively enhances its error detection capabilities (e. g. , accuracy increases from 70. 8% to 77. 7% on ProcessBench) and surpasses strong baselines, including distilled R1 model. Furthermore, SPC can guide the test-time search of diverse LLMs and significantly improve their mathematical reasoning performance on MATH500 and AIME2024, surpassing those guided by state-of-the-art process reward models.

PDF Details

NeurIPS Conference 2025 Conference Paper

SWE-SQL: Illuminating LLM Pathways to Solve User SQL Issues in Real-World Applications

Jinyang Li
Xiaolong Li
Ge Qu
Per Jacobsson
Bowen Qin
Binyuan Hui
Shuzheng Si
Nan Huo

Resolution of complex SQL issues persists as a significant bottleneck in real-world database applications. Current Large Language Models (LLMs), while adept at text-to-SQL translation, have not been rigorously evaluated on the more challenging task of debugging on SQL issues. In order to address this gap, we introduce BIRD-CRITIC, a new SQL issue debugging benchmark comprising 530 carefully curated PostgreSQL tasks ( BIRD-CRITIC-PG ) and 570 multi-dialect tasks ( BIRD-CRITIC-Multi ), which are distilled from authentic user issues and replayed within new environments to facilitate rigorous and contamination-free evaluation. Baseline evaluations on BIRD-CRITIC underscore the task's complexity, with the leading reasoning model O3-Mini achieving only 38. 87% success rate on BIRD-CRITIC-PG and 33. 33% on BIRD-CRITIC-Multi. Meanwhile, realizing open-source models for database tasks is crucial which can empower local development while safeguarding data privacy. Therefore, we present Six-Gym ( S ql-f IX -Gym), a training environment for elevating the capabilities of open-source models specifically for SQL issue debugging. This environment leverages SQL-Rewind strategy, which automatically generates executable issue-solution datasets by reverse-engineering issues from verified SQLs. However, popular trajectory-based fine-tuning methods do not explore substantial supervisory signals. We further propose f -Plan Boosting, which extracts high-level debugging plans automatically from SQL solutions, enabling the teacher LLMs to harvest and produce 73. 7% more successful trajectories for training. We integrate these components into an open-source agent, BIRD-Fixer. Based on Qwen-2. 5-Coder-14B, BIRD-Fixer raises its success rate to 38. 11% on BIRD-CRITIC-PG and 29. 65% on BIRD-CRITIC-Multi, surpassing many leading proprietary models such as Claude-3. 7-Sonnet and GPT-4. 1, marking a significant step toward democratizing sophisticated SQL-debugging capabilities for both research and industry.

PDF Details

NeurIPS Conference 2025 Conference Paper

The Lighthouse of Language: Enhancing LLM Agents via Critique-Guided Improvement

Ruihan Yang
Fanghua Ye
Jian Li
Siyu Yuan
Yikai Zhang
Zhaopeng Tu
Xiaolong Li
Deqing Yang

Large language models (LLMs) have recently transformed from text-based assistants to autonomous agents capable of planning, reasoning, and iteratively improving their actions. While numerical reward signals and verifiers can effectively rank candidate actions, they often provide limited contextual guidance. In contrast, natural language feedback better aligns with the generative capabilities of LLMs, providing richer and more actionable suggestions. However, parsing and implementing this feedback effectively can be challenging for LLM-based agents. In this work, we introduce Critique-Guided Improvement (CGI), a novel two-player framework, comprising an actor model that explores an environment and a critic model that generates detailed nature language feedback. By training the critic to produce fine-grained assessments and actionable revisions, and the actor to utilize these critiques, our approach promotes more robust exploration of alternative strategies while avoiding local optima. Experiments in three interactive environments show that CGI outperforms existing baselines by a substantial margin. Notably, even a small critic model surpasses GPT-4 in feedback quality. The resulting actor achieves state-of-the-art performance, demonstrating the power of explicit iterative guidance to enhance decision-making in LLM-based agents.

PDF Details

NeurIPS Conference 2025 Conference Paper

Two Experts Are All You Need for Steering Thinking: Reinforcing Cognitive Effort in MoE Reasoning Models Without Additional Training

Mengru Wang
Xingyu Chen
Yue Wang
Zhiwei He
Jiahao Xu
Tian Liang
Qiuzhi Liu
Yunzhi Yao

Mixture-of-Experts (MoE) architectures within Large Reasoning Models (LRMs) have achieved impressive reasoning capabilities by selectively activating experts to facilitate structured cognitive processes. Despite notable advances, existing reasoning models often suffer from cognitive inefficiencies like overthinking and underthinking. To address these limitations, we introduce a novel inference-time steering methodology called Reinforcing Cognitive Experts (RICE), designed to improve reasoning depth and efficiency without additional training or complex heuristics. Leveraging normalized Pointwise Mutual Information (nPMI), we systematically identify specialized experts, termed cognitive experts that orchestrate meta-level reasoning operations characterized by tokens like. Empirical evaluations with leading MoE-based LRMs (DeepSeek-R1 and Qwen3-235B) on rigorous quantitative and scientific reasoning benchmarks (AIME and GPQA Diamond) demonstrate noticeable and consistent improvements in reasoning accuracy, cognitive efficiency, and cross-domain generalization. Crucially, our lightweight approach substantially outperforms prevalent reasoning-steering techniques, such as prompt design and decoding constraints, while preserving the model's general instruction-following skills. These results highlight reinforcing cognitive experts as a promising, practical, and interpretable direction to enhance cognitive efficiency within advanced reasoning models.

PDF Details

AIIM Journal 2024 Journal Article

A label information fused medical image report generation framework

Shuifa Sun
Zhoujunsen Mei
Xiaolong Li
Tinglong Tang
Zhanglin Su
Yirong Wu

Medical imaging is an important tool for clinical diagnosis. Nevertheless, it is very time-consuming and error-prone for physicians to prepare imaging diagnosis reports. Therefore, it is necessary to develop some methods to generate medical imaging reports automatically. Currently, the task of medical imaging report generation is challenging in at least two aspects: (1) medical images are very similar to each other. The differences between normal and abnormal images and between different abnormal images are usually trivial; (2) unrelated or incorrect keywords describing abnormal findings in the generated reports lead to mis-communications. In this paper, we propose a medical image report generation framework composed of four modules, including a Transformer encoder, a MIX-MLP multi-label classification network, a co-attention mechanism (CAM) based semantic and visual feature fusion, and a hierarchical LSTM decoder. The Transformer encoder can be used to learn long-range dependencies between images and labels, effectively extract visual and semantic features of images, and establish long-term dependent relationships between visual and semantic information to accurately extract abnormal features from images. The MIX-MLP multi-label classification network, the co-attention mechanism and the hierarchical LSTM network can better identify abnormalities, achieving visual and text alignment fusion and multi-label diagnostic classification to better facilitate report generation. The results of the experiments performed on two widely used radiology report datasets, IU X-RAY and MIMIC-CXR, show that our proposed framework outperforms current report generation models in terms of both natural linguistic generation metrics and clinical efficacy assessment metrics. The code of this work is available online at https: //github. com/watersunhznu/LIFMRG.

Details DOI

AAAI Conference 2024 Conference Paper

On the Unstable Convergence Regime of Gradient Descent

Shuo Chen
Jiaying Peng
Xiaolong Li
Yao Zhao

Traditional gradient descent (GD) has been fully investigated for convex or L-smoothness functions, and it is widely utilized in current neural network optimization. The classical descent lemma ensures that for a function with L-smoothness, the GD trajectory converges stably towards the minimum when the learning rate is below 2 / L. This convergence is marked by a consistent reduction in the loss function throughout the iterations. However, recent experimental studies have demonstrated that even when the L-smoothness condition is not met, or if the learning rate is increased leading to oscillations in the loss function during iterations, the GD trajectory still exhibits convergence over the long run. This phenomenon is referred to as the unstable convergence regime of GD. In this paper, we present a theoretical perspective to offer a qualitative analysis of this phenomenon. The unstable convergence is in fact an inherent property of GD for general twice differentiable functions. Specifically, the forwardinvariance of GD is established, i.e., it ensures that any point within a local region will always remain within this region under GD iteration. Then, based on the forward-invariance, for the initialization outside an open set containing the local minimum, the loss function will oscillate at the first several iterations and then become monotonely decreasing after the GD trajectory jumped into the open set. This work theoretically clarifies the unstable convergence phenomenon of GD discussed in previous experimental works. The unstable convergence of GD mainly depends on the selection of the initialization, and it is actually inevitable due to the complex nature of loss function.

PDF Details DOI

NeurIPS Conference 2024 Conference Paper

Self-playing Adversarial Language Game Enhances LLM Reasoning

Pengyu Cheng
Yong Dai
Tianhao Hu
Han Xu
Zhisong Zhang
Lei Han
Nan Du
Xiaolong Li

We explore the potential of self-play training for large language models (LLMs) in a two-player adversarial language game called Adversarial Taboo. In this game, an attacker and a defender communicate around a target word only visible to the attacker. The attacker aims to induce the defender to speak the target word unconsciously, while the defender tries to infer the target word from the attacker's utterances. To win the game, both players must have sufficient knowledge about the target word and high-level reasoning ability to infer and express in this information-reserved conversation. Hence, we are curious about whether LLMs' reasoning ability can be further enhanced by Self-Playing this Adversarial language Game (SPAG). With this goal, we select several open-source LLMs and let each act as the attacker and play with a copy of itself as the defender on an extensive range of target words. Through reinforcement learning on the game outcomes, we observe that the LLMs' performances uniformly improve on a broad range of reasoning benchmarks. Furthermore, iteratively adopting this self-play process can continuously promote LLMs' reasoning abilities. The code is available at https: //github. com/Linear95/SPAG.

PDF Details DOI

JBHI Journal 2023 Journal Article

Natural Grasp Intention Recognition Based on Gaze in Human–Robot Interaction

Bo Yang
Jian Huang
Xinxing Chen
Xiaolong Li
Yasuhisa Hasegawa

Objective: While neuroscience research has established a link between vision and intention, studies on gaze data features for intention recognition are absent. The majority of existing gaze-based intention recognition approaches are based on deliberate long-term fixation and suffer from insufficient accuracy. In order to address the lack of features and insufficient accuracy in previous studies, the primary objective of this study is to suppress noise from human gaze data and extract useful features for recognizing grasp intention. Methods: We conduct gaze movement evaluation experiments to investigate the characteristics of gaze motion. The target-attracted gaze movement model (TAGMM) is proposed as a quantitative description of gaze movement based on the findings. A Kalman filter (KF) is used to reduce the noise in the gaze data based on TAGMM. We conduct gaze-based natural grasp intention recognition evaluation experiments to collect the subject's gaze data. Four types of features describing gaze point dispersion ( $f_{var}$ ), gaze point movement ( $f_{gm}$ ), head movement ( $f_{hm}$ ), and distance from the gaze points to objects ( $f_{d_{j}}$ ) are then proposed to recognize the subject's grasp intentions. With the proposed features, we perform intention recognition experiments, employing various classifiers, and the results are compared with different methods. Results: The statistical analysis reveals that the proposed features differ significantly across intentions, offering the possibility of employing these features to recognize grasp intentions. We demonstrated the intention recognition performance utilizing the TAGMM and the proposed features in within-subject and cross-subject experiments. The results indicate that the proposed method can recognize the intention with accuracy improvements of 44. 26% (within-subject) and 30. 67% (cross-subject) over the fixation-based method. The proposed method also consumes less time (34. 87 ms) to recognize the intention than the fixation-based method (about 1 s). Conclusion: This work introduces a novel TAGMM for modeling gaze movement and a variety of practical features for recognizing grasp intentions. Experiments confirm the effectiveness of our approach. Significance: The proposed TAGMM is capable of modeling gaze movements and can be utilized to process gaze data, and the proposed features can reveal the user's intentions. These results contribute to the development of gaze-based human-robot interaction.

Details DOI

AAAI Conference 2023 Conference Paper

Tracking and Reconstructing Hand Object Interactions from Point Cloud Sequences in the Wild

Jiayi Chen
Mi Yan
Jiazhao Zhang
Yinzhen Xu
Xiaolong Li
Yijia Weng
Li Yi
Shuran Song

In this work, we tackle the challenging task of jointly tracking hand object poses and reconstructing their shapes from depth point cloud sequences in the wild, given the initial poses at frame 0. We for the first time propose a point cloud-based hand joint tracking network, HandTrackNet, to estimate the inter-frame hand joint motion. Our HandTrackNet proposes a novel hand pose canonicalization module to ease the tracking task, yielding accurate and robust hand joint tracking. Our pipeline then reconstructs the full hand via converting the predicted hand joints into a MANO hand. For object tracking, we devise a simple yet effective module that estimates the object SDF from the first frame and performs optimization-based tracking. Finally, a joint optimization step is adopted to perform joint hand and object reasoning, which alleviates the occlusion-induced ambiguity and further refines the hand pose. During training, the whole pipeline only sees purely synthetic data, which are synthesized with sufficient variations and by depth simulation for the ease of generalization. The whole pipeline is pertinent to the generalization gaps and thus directly transferable to real in-the-wild data. We evaluate our method on two real hand object interaction datasets, e.g. HO3D and DexYCB, without any fine-tuning. Our experiments demonstrate that the proposed method significantly outperforms the previous state-of-the-art depth-based hand and object pose estimation and tracking methods, running at a frame rate of 9 FPS. We have released our code on https://github.com/PKU-EPIC/HOTrack.

PDF Details DOI

IROS Conference 2022 Conference Paper

Metabolic Efficiency Improvement of Human Walking by Shoulder Stress Reduction through Load Transfer Backpack

Yu Cao 0008
Jian Huang 0001
Xiaolong Li
Mengshi Zhang
Caihua Xiong
Samer Mohammed
Yaonan Zhu
Yasuhisa Hasegawa

The dynamic load attached to the load gravity imposes an excessive burden to human shoulders during load carriage, resulting in possible muscle injuries and additional physical exertion. This paper proposes an active suspension backpack, capable of transferring partial load from human shoulders to pelvis and alleviating the dynamic load through separated panels and motor actuation, to reduce pressure on human shoulders and improve walking metabolic efficiency. Based on the human body motion in the vertical direction, the dynamical model of the human-backpack system with shoulder interaction force measured by a soft ballonet with an embedded air pressure sensor is introduced, and an impedance controller has been implemented to maintain a relatively small and constant pressure on the shoulder. In an experimental case study, we presents preliminary results of three healthy subjects performing a treadmill walking with a 20kg load in ACTIVE configuration where the shoulder pressure shows a decrease by 30% along with a reduction of the metabolic energy consumption by 16. 4%, compared with the load LOCKED case.

Details

NeurIPS Conference 2021 Conference Paper

Leveraging SE(3) Equivariance for Self-supervised Category-Level Object Pose Estimation from Point Clouds

Xiaolong Li
Yijia Weng
Li Yi
Leonidas J. Guibas
A. Abbott
Shuran Song
He Wang

Category-level object pose estimation aims to find 6D object poses of previously unseen object instances from known categories without access to object CAD models. To reduce the huge amount of pose annotations needed for category-level learning, we propose for the first time a self-supervised learning framework to estimate category-level 6D object pose from single 3D point clouds. During training, our method assumes no ground-truth pose annotations, no CAD models, and no multi-view supervision. The key to our method is to disentangle shape and pose through an invariant shape reconstruction module and an equivariant pose estimation module, empowered by SE(3) equivariant point cloud networks. The invariant shape reconstruction module learns to perform aligned reconstructions, yielding a category-level reference frame without using any annotations. In addition, the equivariant pose estimation module achieves category-level pose estimation accuracy that is comparable to some fully supervised methods. Extensive experiments demonstrate the effectiveness of our approach on both complete and partial depth point clouds from the ModelNet40 benchmark, and on real depth point clouds from the NOCS-REAL 275 dataset. The project page with code and visualizations can be found at: dragonlong. github. io/equi-pose.

PDF Details

AAAI Conference 2021 Conference Paper

LRC-BERT: Latent-representation Contrastive Knowledge Distillation for Natural Language Understanding

Hao Fu
Shaojun Zhou
Qihong Yang
Junjie Tang
Guiquan Liu
Kaikui Liu
Xiaolong Li

The pre-training models such as BERT have achieved great results in various natural language processing problems. However, a large number of parameters need significant amounts of memory and the consumption of inference time, which makes it difficult to deploy them on edge devices. In this work, we propose a knowledge distillation method LRC- BERT based on contrastive learning to fit the output of the intermediate layer from the angular distance aspect, which is not considered by the existing distillation methods. Furthermore, we introduce a gradient perturbation-based training architecture in the training phase to increase the robustness of LRC-BERT, which is the first attempt in knowledge distillation. Additionally, in order to better capture the distribution characteristics of the intermediate layer, we design a twostage training method for the total distillation loss. Finally, by verifying 9 datasets on the General Language Understanding Evaluation (GLUE) benchmark, the performance of the proposed LRC-BERT exceeds the existing state-of-the-art methods, which proves the effectiveness of our method.

PDF Details

AAAI Conference 2020 Conference Paper

Long Short-Term Sample Distillation

Liang Jiang
Zujie Wen
Zhongping Liang
Yafang Wang
Gerard de Melo
Zhe Li
Liangzhuang Ma
Jiaxing Zhang

In the past decade, there has been substantial progress at training increasingly deep neural networks. Recent advances within the teacher–student training paradigm have established that information about past training updates show promise as a source of guidance during subsequent training steps. Based on this notion, in this paper, we propose Long Short- Term Sample Distillation, a novel training policy that simultaneously leverages multiple phases of the previous training process to guide the later training updates to a neural network, while efﬁciently proceeding in just one single generation pass. With Long Short-Term Sample Distillation, the supervision signal for each sample is decomposed into two parts: a long-term signal and a short-term one. The long-term teacher draws on snapshots from several epochs ago in order to provide steadfast guidance and to guarantee teacher– student differences, while the short-term one yields more upto-date cues with the goal of enabling higher-quality updates. Moreover, the teachers for each sample are unique, such that, overall, the model learns from a very diverse set of teachers. Comprehensive experimental results across a range of vision and NLP tasks demonstrate the effectiveness of this new training method.

PDF Details

AAAI Conference 2020 Conference Paper

Span-Based Neural Buffer: Towards Efficient and Effective Utilization of Long-Distance Context for Neural Sequence Models

Yangming Li
Kaisheng Yao
Libo Qin
Shuang Peng
Yijia Liu
Xiaolong Li

Neural sequence model, though widely used for modeling sequential data such as the language model, has sequential recency bias (Kuncoro et al. 2018) to the local context, limiting its full potential to capture long-distance context. To address this problem, this paper proposes augmenting sequence models with a span-based neural buffer that efﬁciently represents long-distance context, allowing a gate policy network to make interpolated predictions from both the neural buffer and the underlying sequence model. Training this policy network to utilize long-distance context is however challenging due to the simple sentence dominance problem (Marvin and Linzen 2018). To alleviate this problem, we propose a novel training algorithm that combines an annealed maximum likelihood estimation with an intrinsic reward-driven reinforcement learning. Sequence models with the proposed span-based neural buffer signiﬁcantly improve the state-of-the-art perplexities on the benchmark Penn Treebank and WikiText-2 datasets to 43. 9 and 35. 2 respectively. We conduct extensive analysis and conﬁrm that the proposed architecture and the training algorithm both contribute to the improvements.

PDF Details

IJCAI Conference 2020 Conference Paper

Two-stage Behavior Cloning for Spoken Dialogue System in Debt Collection

Zihao Wang
Jia Liu
Hengbin Cui
Chunxiang Jin
Minghui Yang
Yafang Wang
Xiaolong Li
Renxin Mao

With the rapid growth of internet finance and the booming of financial lending, the intelligent calling for debt collection in FinTech companies has driven increasing attention. Nowadays, the widely used intelligent calling system is based on dialogue flow, namely configuring the interaction flow with the finite-state machine. In our scenario of debt collection, the completed dialogue flow contains more than one thousand interactive paths. All the dialogue procedures are artificially specified, with extremely high maintenance costs and error-prone. To solve this problem, we propose the behavior-cloning-based collection robot framework without any dialogue flow configuration, called two-stage behavior cloning (TSBC). In the first stage, we use multi-label classification model to obtain policies that may be able to cope with the current situation according to the dialogue state; in the second stage, we score several scripts under each obtained policy to select the script with the highest score as the reply for the current state. This framework makes full use of the massive manual collection records without labeling and fully absorbs artificial wisdom and experience. We have conducted extensive experiments in both single-round and multi-round scenarios and showed the effectiveness of the proposed system. The accuracy of a single round of dialogue can be improved by 5%, and the accuracy of multiple rounds of dialogue can be increased by 3. 1%.

PDF Details DOI

IJCAI Conference 2019 Conference Paper

AntProphet: an Intention Mining System behind Alipay's Intelligent Customer Service Bot

Cen Chen
Xiaolu Zhang
Sheng Ju
Chilin Fu
Caizhi Tang
Jun Zhou
Xiaolong Li

We create an intention mining system, named AntProphet, for Alipay's intelligent customer service bot, to alleviate the burden of customer service. Whenever users have any questions, AntProphet is the first stop to help users to answer their questions. Our system gathers users' profile and their historical behavioral trajectories, together with contextual information to predict users' intention, i. e. , the potential questions that users want to resolve. AntProphet takes care of more than 90% of the customer service demands in the Alipay APP and resolves most of the users' problems on the spot, thus significantly reduces the burden of manpower. With the help of it, the overall satisfaction rate of our customer service bot exceeds 85%.

PDF Details

AAAI Conference 2019 Conference Paper

Cash-Out User Detection Based on Attributed Heterogeneous Information Network with a Hierarchical Attention Mechanism

Binbin Hu
Zhiqiang Zhang
Chuan Shi
Jun Zhou
Xiaolong Li
Yuan Qi

As one of the major frauds in financial services, cash-out fraud is that users pursue cash gains with illegal or insincere means. Conventional solutions for the cash-out user detection are to perform subtle feature engineering for each user and then apply a classifier, such as GDBT and Neural Network. However, users in financial services have rich interaction relations, which are seldom fully exploited by conventional solutions. In this paper, with the real datasets in Ant Credit Pay of Ant Financial Services Group, we first study the cashout user detection problem and propose a novel hierarchical attention mechanism based cash-out user detection model, called HACUD. Specifically, we model different types of objects and their rich attributes and interaction relations in the scenario of credit payment service with an Attributed Heterogeneous Information Network (AHIN). The HACUD model enhances feature representation of objects through meta-path based neighbors exploiting different aspects of structure information in AHIN. Furthermore, a hierarchical attention mechanism is elaborately designed to model user’s preferences towards attributes and meta-paths. Experimental results on two real datasets show that the HACUD outperforms the state-of-the-art methods.

PDF Details

TIST Journal 2019 Journal Article

Distributed Deep Forest and its Application to Automatic Detection of Cash-Out Fraud

Ya-Lin Zhang
Jun Zhou
Wenhao Zheng
Ji Feng
Longfei Li
Ziqi Liu
Ming Li
Zhiqiang Zhang

Internet companies are facing the need for handling large-scale machine learning applications on a daily basis and distributed implementation of machine learning algorithms which can handle extra-large-scale tasks with great performance is widely needed. Deep forest is a recently proposed deep learning framework which uses tree ensembles as its building blocks and it has achieved highly competitive results on various domains of tasks. However, it has not been tested on extremely large-scale tasks. In this work, based on our parameter server system, we developed the distributed version of deep forest. To meet the need for real-world tasks, many improvements are introduced to the original deep forest model, including MART (Multiple Additive Regression Tree) as base learners for efficiency and effectiveness consideration, the cost-based method for handling prevalent class-imbalanced data, MART based feature selection for high dimension data, and different evaluation metrics for automatically determining the cascade level. We tested the deep forest model on an extra-large-scale task, i.e., automatic detection of cash-out fraud, with more than 100 million training samples. Experimental results showed that the deep forest model has the best performance according to the evaluation metrics from different perspectives even with very little effort for parameter tuning. This model can block fraud transactions in a large amount of money each day. Even compared with the best-deployed model, the deep forest model can additionally bring a significant decrease in economic loss each day.

Details DOI

AAAI Conference 2019 Conference Paper

GeniePath: Graph Neural Networks with Adaptive Receptive Paths

Ziqi Liu
Chaochao Chen
Longfei Li
Jun Zhou
Xiaolong Li
Le Song
Yuan Qi

In this paper, we propose a new online feature selection algorithm for streaming data. We aim to focus on the following two problems which remain unaddressed in literature. First, most existing online feature selection algorithms merely utilize the first-order information of the data streams, regardless of the fact that second-order information explores the correlations between features and significantly improves the performance. Second, most online feature selection algorithms are based on the balanced data presumption, which is not true in many real-world applications. For example, in fraud detection, the number of positive examples are much less than negative examples because most cases are not fraud. The balanced assumption will make the selected features biased towards the majority class and fail to detect the fraud cases. We propose an Adaptive Sparse Confidence-Weighted (ASCW) algorithm to solve the aforementioned two problems. We first introduce an `0-norm constraint into the second-order confidence-weighted (CW) learning for feature selection. Then the original loss is substituted with a cost-sensitive loss function to address the imbalanced data issue. Furthermore, our algorithm maintains multiple sparse CW learner with the corresponding cost vector to dynamically select an optimal cost. We theoretically enhance the theory of sparse CW learning and analyze the performance behavior in F-measure. Empirical studies show the superior performance over the stateof-the-art online learning methods in the online-batch setting.

PDF Details

AAAI Conference 2018 Conference Paper

cw2vec: Learning Chinese Word Embeddings with Stroke n-gram Information

Shaosheng Cao
Wei Lu
Jun Zhou
Xiaolong Li

We propose cw2vec, a novel method for learning Chinese word embeddings. It is based on our observation that exploiting stroke-level information is crucial for improving the learning of Chinese word embeddings. Specifically, we design a minimalist approach to exploit such features, by using stroke n-grams, which capture semantic and morphological level information of Chinese words. Through qualitative analysis, we demonstrate that our model is able to extract semantic information that cannot be captured by existing methods. Empirical results on the word similarity, word analogy, text classification and named entity recognition tasks show that the proposed approach consistently outperforms state-of-the-art approaches such as word-based word2vec and GloVe, character-based CWE, component-based JWE and pixel-based GWE.

PDF Details

AAAI Conference 2018 Conference Paper

Privacy Preserving Point-of-Interest Recommendation Using Decentralized Matrix Factorization

Chaochao Chen
Ziqi Liu
Peilin Zhao
Jun Zhou
Xiaolong Li

Points of interest (POI) recommendation has been drawn much attention recently due to the increasing popularity of location-based networks, e. g. , Foursquare and Yelp. Among the existing approaches to POI recommendation, Matrix Factorization (MF) based techniques have proven to be effective. However, existing MF approaches suffer from two major problems: (1) Expensive computations and storages due to the centralized model training mechanism: the centralized learners have to maintain the whole user-item rating matrix, and potentially huge low rank matrices. (2) Privacy issues: the users’ preferences are at risk of leaking to malicious attackers via the centralized learner. To solve these, we present a Decentralized MF (DMF) framework for POI recommendation. Speciﬁcally, instead of maintaining all the low rank matrices and sensitive rating data for training, we propose a random walk based decentralized training technique to train MF models on each user’s end, e. g. , cell phone and Pad. By doing so, the ratings of each user are still kept on one’s own hand, and moreover, decentralized learning can be taken as distributed learning with multi-learners (users), and thus alleviates the computation and storage issue. Experimental results on two real-world datasets demonstrate that, comparing with the classic and state-of-the-art latent factor models, DMF signiﬁcantly improvements the recommendation performance in terms of precision and recall.

PDF Details