Qingyang Li

Xingxing Xie, Gong Cheng, Qingyang Li et al.

Current mainstream object detection methods for large aerial images usually divide large images into patches and then exhaustively detect the objects of interest on all patches, no matter whether there exist objects or not. This paradigm, although effective, is inefficient because the detectors have to go through all patches, severely hindering the inference speed. This paper presents an Objectness Activation Network (OAN) to help detectors focus on fewer patches but achieve more efficient inference and more accurate results, enabling a simple and effective solution to object detection in large images. In brief, OAN is a light fully-convolutional network for judging whether each patch contains objects or not, which can be easily integrated into many object detectors and jointly trained with them end-to-end. We extensively evaluate our OAN with five advanced detectors. Using OAN, all five detectors acquire more than 30.0% speed-up on three large-scale aerial image datasets, meanwhile with consistent accuracy improvements. On extremely large Gaofen-2 images (29200$\times$27620 pixels), our OAN improves the detection speed by 70.5%. Moreover, we extend our OAN to driving-scene object detection and 4K video object detection, boosting the detection speed by 112.1% and 75.0%, respectively, without sacrificing the accuracy. Code is available at https://github.com/Ranchosky/OAN.

Lei Lin, Jiayi Fu, Pengli Liu et al.

Although chain-of-thought (CoT) prompting combined with language models has achieved encouraging results on complex reasoning tasks, the naive greedy decoding used in CoT prompting usually causes the repetitiveness and local optimality. To address this shortcoming, ensemble-optimization tries to obtain multiple reasoning paths to get the final answer assembly. However, current ensemble-optimization methods either simply employ rule-based post-processing such as \textit{self-consistency}, or train an additional model based on several task-related human annotations to select the best one among multiple reasoning paths, yet fail to generalize to realistic settings where the type of input questions is unknown or the answer format of reasoning paths is unknown. To avoid their limitations, we propose \textbf{Self-Agreement}, a generalizable ensemble-optimization method applying in almost all scenarios where the type of input questions and the answer format of reasoning paths may be known or unknown. Self-agreement firstly samples from language model's decoder to generate a \textit{diverse} set of reasoning paths, and subsequently prompts the language model \textit{one more time} to determine the optimal answer by selecting the most \textit{agreed} answer among the sampled reasoning paths. Self-agreement simultaneously achieves remarkable performance on six public reasoning benchmarks and superior generalization capabilities.

Yanqiu Wu, Qingyang Li, Zhiwei Qin

A fundamental question in any peer-to-peer ride-sharing system is how to, both effectively and efficiently, meet the request of passengers to balance the supply and demand in real time. On the passenger side, traditional approaches focus on pricing strategies by increasing the probability of users' call to adjust the distribution of demand. However, previous methods do not take into account the impact of changes in strategy on future supply and demand changes, which means drivers are repositioned to different destinations due to passengers' calls, which will affect the driver's income for a period of time in the future. Motivated by this observation, we make an attempt to optimize the distribution of demand to handle this problem by learning the long-term spatio-temporal values as a guideline for pricing strategy. In this study, we propose an offline deep reinforcement learning based method focusing on the demand side to improve the utilization of transportation resources and customer satisfaction. We adopt a spatio-temporal learning method to learn the value of different time and location, then incentivize the ride requests of passengers to adjust the distribution of demand to balance the supply and demand in the system. In particular, we model the problem as a Markov Decision Process (MDP).

DeepSeek-AI, Aixin Liu, Aoxue Mei et al.

We introduce DeepSeek-V3.2, a model that harmonizes high computational efficiency with superior reasoning and agent performance. The key technical breakthroughs of DeepSeek-V3.2 are as follows: (1) DeepSeek Sparse Attention (DSA): We introduce DSA, an efficient attention mechanism that substantially reduces computational complexity while preserving model performance in long-context scenarios. (2) Scalable Reinforcement Learning Framework: By implementing a robust reinforcement learning protocol and scaling post-training compute, DeepSeek-V3.2 performs comparably to GPT-5. Notably, our high-compute variant, DeepSeek-V3.2-Speciale, surpasses GPT-5 and exhibits reasoning proficiency on par with Gemini-3.0-Pro, achieving gold-medal performance in both the 2025 International Mathematical Olympiad (IMO) and the International Olympiad in Informatics (IOI). (3) Large-Scale Agentic Task Synthesis Pipeline: To integrate reasoning into tool-use scenarios, we developed a novel synthesis pipeline that systematically generates training data at scale. This methodology facilitates scalable agentic post-training, yielding substantial improvements in generalization and instruction-following robustness within complex, interactive environments.

Qingyang Li, Yihang Zhang, Zhidong Jia et al.

It is an interesting question Can and How Large Language Models (LLMs) understand non-language network data, and help us detect unknown malicious flows. This paper takes Carpet Bombing as a case study and shows how to exploit LLMs' powerful capability in the networking area. Carpet Bombing is a new DDoS attack that has dramatically increased in recent years, significantly threatening network infrastructures. It targets multiple victim IPs within subnets, causing congestion on access links and disrupting network services for a vast number of users. Characterized by low-rates, multi-vectors, these attacks challenge traditional DDoS defenses. We propose DoLLM, a DDoS detection model utilizes open-source LLMs as backbone. By reorganizing non-contextual network flows into Flow-Sequences and projecting them into LLMs semantic space as token embeddings, DoLLM leverages LLMs' contextual understanding to extract flow representations in overall network context. The representations are used to improve the DDoS detection performance. We evaluate DoLLM with public datasets CIC-DDoS2019 and real NetFlow trace from Top-3 countrywide ISP. The tests have proven that DoLLM possesses strong detection capabilities. Its F1 score increased by up to 33.3% in zero-shot scenarios and by at least 20.6% in real ISP traces.

Hao Yi, Qingyang Li, Yulan Hu et al.

Recently, enhancing the numerical and logical reasoning capability of Large Language Models (LLMs) has emerged as a research hotspot. Existing methods face several limitations: inference-phase techniques (e.g., Chain of Thoughts) rely on prompt selection and the pretrained knowledge; sentence-level Supervised Fine-Tuning (SFT) and Direct Preference Optimization (DPO) struggle with step-wise mathematical correctness and depend on stronger models distillation or human annotations; while Reinforcement Learning (RL) approaches incur high GPU memory costs and unstable training. To address these, we propose \textbf{S}elf-training framework integrating \textbf{P}rocess \textbf{P}reference learning using \textbf{D}ynamic value margin (SPPD). SPPD leverages a process-based Markov Decision Process (MDP) and Bellman optimality equation to derive \textbf{dynamic value margin} on step-level preference optimization, which employs tree-based self-sampling on model responses \textbf{without any distillation} from other models. Furthermore, we theoretically prove that SPPD is \textbf{equivalent to on-policy policy gradient methods} under reward constraints. Experiments on 7B-scale models demonstrate superior performance across in-domain and out-domain mathematical benchmarks. We open-source our code at \href{https://anonymous.4open.science/r/SSDPO-D-DCDD}{https://anonymous.4open.science/r/SPPD-DCDD}.

Kaihui Chen, Hao Yi, Qingyang Li et al.

Enhancing the conformity of large language models (LLMs) to human preferences remains an ongoing research challenge. Recently, offline approaches such as Direct Preference Optimization (DPO) have gained prominence as attractive options due to offering effective improvement in simple, efficient, and stable without interactions with reward models. However, these offline preference optimization methods highly rely on the quality of pairwise preference samples. Meanwhile, numerous iterative methods require additional training of reward models to select positive and negative samples from the model's own generated responses for preference learning. Furthermore, as LLMs' capabilities advance, it is quite challenging to continuously construct high-quality positive and negative preference instances from the model's outputs due to the lack of diversity. To tackle these challenges, we propose TSO, or Self-Training with Scaled Preference Optimization, a framework for preference optimization that conducts self-training preference learning without training an additional reward model. TSO enhances the diversity of responses by constructing a model matrix and incorporating human preference responses. Furthermore, TSO introduces corrections for model preference errors through human and AI feedback. Finally, TSO adopts iterative and dual clip reward strategies to update the reference model and its responses, adaptively adjusting preference data and balancing the optimization process. Experimental results demonstrate that TSO outperforms existing mainstream methods on various alignment evaluation benchmarks, providing practical insight into preference data construction and model training strategies in the alignment domain.

Qingyang Li

The exponential growth of user-generated movie reviews on digital platforms has made accurate text sentiment classification a cornerstone task in natural language processing. Traditional models, including standard BERT and recurrent architectures, frequently struggle to capture long-distance semantic dependencies and resolve ambiguous emotional expressions in lengthy review texts. This paper proposes a novel hybrid framework that seamlessly integrates dynamic adaptive multi-head attention with supervised contrastive learning into a BERT-based Transformer encoder. The dynamic adaptive attention module employs a global context pooling vector to dynamically regulate the contribution of each attention head, thereby focusing on critical sentiment-bearing tokens while suppressing noise. Simultaneously, the supervised contrastive learning branch enforces tighter intra-class compactness and larger inter-class separation in the embedding space. Extensive experiments on the IMDB dataset demonstrate that the proposed model achieves competitive performance with an accuracy of 94.67\%, outperforming strong baselines by 1.5--2.5 percentage points. The framework is lightweight, efficient, and readily extensible to other text classification tasks.

Yunpeng Tan, Qingyang Li, Mingxin Yang et al.

Encryption has been commonly used in network traffic to secure transmission, but it also brings challenges for malicious traffic detection, due to the invisibility of the packet payload. Graph-based methods are emerging as promising solutions by leveraging multi-host interactions to promote detection accuracy. But most of them face a critical problem: Graph Drift, where the flow statistics or topological information of a graph change over time. To overcome these drawbacks, we propose a graph-assisted encrypted traffic detection system, MalMoE, which applies Mixture of Experts (MoE) to select the best expert model for drift-aware classification. Particularly, we design 1-hop-GNN-like expert models that handle different graph drifts by analyzing graphs with different features. Then, the redesigned gate model conducts expert selection according to the actual drift. MalMoE is trained with a stable two-stage training strategy with data augmentation, which effectively guides the gate on how to perform routing. Experiments on open-source, synthetic, and real-world datasets show that MalMoE can perform precise and real-time detection.

Hongyi Duan, Qingyang Li, Yuchen Li et al.

As global attention on renewable and clean energy grows, the research and implementation of microgrids become paramount. This paper delves into the methodology of exploring the relationship between the operational and environmental costs of microgrids through multi-objective optimization models. By integrating various optimization algorithms like Genetic Algorithm, Simulated Annealing, Ant Colony Optimization, and Particle Swarm Optimization, we propose an integrated approach for microgrid optimization. Simulation results depict that these algorithms provide different dispatch results under economic and environmental dispatch, revealing distinct roles of diesel generators and micro gas turbines in microgrids. Overall, this study offers in-depth insights and practical guidance for microgrid design and operation.

Yunpeng Tan, Junlin Hao, Jiangkai Wu et al.

Neural Radiance Field (NeRF) is widely known for high-fidelity novel view synthesis. However, even the state-of-the-art NeRF model, Gaussian Splatting, requires minutes for training, far from the real-time performance required by multimedia scenarios like telemedicine. One of the obstacles is its inefficient sampling, which is only partially addressed by existing works. Existing point-sampling algorithms uniformly sample simple-texture regions (easy to fit) and complex-texture regions (hard to fit), while existing ray-sampling algorithms sample these regions all in the finest granularity (i.e. the pixel level), both wasting GPU training resources. Actually, regions with different texture intensities require different sampling granularities. To this end, we propose a novel dynamic-resolution ray-sampling algorithm, MCBlock, which employs Monte Carlo Tree Search (MCTS) to partition each training image into pixel blocks with different sizes for active block-wise training. Specifically, the trees are initialized according to the texture of training images to boost the initialization speed, and an expansion/pruning module dynamically optimizes the block partition. MCBlock is implemented in Nerfstudio, an open-source toolset, and achieves a training acceleration of up to 2.33x, surpassing other ray-sampling algorithms. We believe MCBlock can apply to any cone-tracing NeRF model and contribute to the multimedia community.

Qingyang Li, Yuchen Li, Hongyi Duan et al.

In this paper, the limitations of YOLOv5s model on small target detection task are deeply studied and improved. The performance of the model is successfully enhanced by introducing GhostNet-based convolutional module, RepGFPN-based Neck module optimization, CA and Transformer's attention mechanism, and loss function improvement using NWD. The experimental results validate the positive impact of these improvement strategies on model precision, recall and mAP. In particular, the improved model shows significant superiority in dealing with complex backgrounds and tiny targets in real-world application tests. This study provides an effective optimization strategy for the YOLOv5s model on small target detection, and lays a solid foundation for future related research and applications.

Yelaman Abdullin, Diego Molla-Aliod, Bahadorreza Ofoghi et al.

Linear programming (LP) problems are pervasive in real-life applications. However, despite their apparent simplicity, an untrained user may find it difficult to determine the linear model of their specific problem. We envisage the creation of a goal-oriented conversational agent that will engage in conversation with the user to elicit all information required so that a subsequent agent can generate the linear model. In this paper, we present an approach for the generation of sample dialogues that can be used to develop and train such a conversational agent. Using prompt engineering, we develop two agents that "talk" to each other, one acting as the conversational agent, and the other acting as the user. Using a set of text descriptions of linear problems from NL4Opt available to the user only, the agent and the user engage in conversation until the agent has retrieved all key information from the original problem description. We also propose an extrinsic evaluation of the dialogues by assessing how well the summaries generated by the dialogues match the original problem descriptions. We conduct human and automatic evaluations, including an evaluation approach that uses GPT-4 to mimic the human evaluation metrics. The evaluation results show an overall good quality of the dialogues, though research is still needed to improve the quality of the GPT-4 evaluation metrics. The resulting dialogues, including the human annotations of a subset, are available to the research community. The conversational agent used for the generation of the dialogues can be used as a baseline.

Qingyang Li

With the increasing diversity of 5G service types and the intensifying dynamic fluctuations of network load, achieve differentiated quality of service assurance in a network slicing environment has become a key issue in resource management. To address this problem, this paper proposes a deep reinforcement learning mechanism for 5G network slicing quality of service assurance based on the traditional proximal policy optimization actor-critic framework. First, the slicing resource allocation is modeled as a constrained Markov decision process, jointly considering the collaborative optimization of bandwidth, computing, and wireless resources. Meanwhile, a graph attention network and bidirectional long short-term memory are introduced to extract topological correlations and temporal service features, combined with an adaptive Lagrangian penalty and dynamic reward shaping mechanism, to comprehensively optimize delay, throughput, reliability, fairness, and slice isolation performance. Experimental results show that the proposed method outperforms existing baseline models in terms of quality of service satisfaction rate, delay control, resource utilization, and convergence stability.

Yiyi Wang, Jian'an Zhang, Hongyi Duan et al.

State Space Models (SSMs), particularly recent selective variants like Mamba, have emerged as a leading architecture for sequence modeling, challenging the dominance of Transformers. However, the success of these state-of-the-art models largely relies on heuristically designed selective mechanisms, which lack a rigorous first-principle derivation. This theoretical gap raises questions about their optimality and robustness against spurious correlations. To address this, we introduce the Principle of Predictive Sufficiency, a novel information-theoretic criterion stipulating that an ideal hidden state should be a minimal sufficient statistic of the past for predicting the future. Based on this principle, we propose the Minimal Predictive Sufficiency State Space Model (MPS-SSM), a new framework where the selective mechanism is guided by optimizing an objective function derived from our principle. This approach encourages the model to maximally compress historical information without losing predictive power, thereby learning to ignore non-causal noise and spurious patterns. Extensive experiments on a wide range of benchmark datasets demonstrate that MPS-SSM not only achieves state-of-the-art performance, significantly outperforming existing models in long-term forecasting and noisy scenarios, but also exhibits superior robustness. Furthermore, we show that the MPS principle can be extended as a general regularization framework to enhance other popular architectures, highlighting its broad potential.

Sheng Ouyang, Yulan Hu, Ge Chen et al.

Rewards serve as proxies for human preferences and play a crucial role in Reinforcement Learning from Human Feedback (RLHF). However, if these rewards are inherently imperfect, exhibiting various biases, they can adversely affect the alignment of large language models (LLMs). In this paper, we collectively define the various biases present in rewards as the problem of reward unfairness. We propose a bias-agnostic method to address the issue of reward fairness from a resource allocation perspective, without specifically designing for each type of bias, yet effectively mitigating them. Specifically, we model preference learning as a resource allocation problem, treating rewards as resources to be allocated while considering the trade-off between utility and fairness in their distribution. We propose two methods, Fairness Regularization and Fairness Coefficient, to achieve fairness in rewards. We apply our methods in both verification and reinforcement learning scenarios to obtain a fairness reward model and a policy model, respectively. Experiments conducted in these scenarios demonstrate that our approach aligns LLMs with human preferences in a more fair manner.

Nicholas Miklaucic, Lai Wei, Rongzhi Dong et al.

Computational materials discovery is limited by the high cost of first-principles calculations. Machine learning (ML) potentials that predict energies from crystal structures are promising, but existing methods face computational bottlenecks. Steerable graph neural networks (GNNs) encode geometry with spherical harmonics, respecting atomic symmetries -- permutation, rotation, and translation -- for physically realistic predictions. Yet maintaining equivariance is difficult: activation functions must be modified, and each layer must handle multiple data types for different harmonic orders. We present Facet, a GNN architecture for efficient ML potentials, developed through systematic analysis of steerable GNNs. Our innovations include replacing expensive multi-layer perceptrons (MLPs) for interatomic distances with splines, which match performance while cutting computational and memory demands. We also introduce a general-purpose equivariant layer that mixes node information via spherical grid projection followed by standard MLPs -- faster than tensor products and more expressive than linear or gate layers. On the MPTrj dataset, Facet matches leading models with far fewer parameters and under 10% of their training compute. On a crystal relaxation task, it runs twice as fast as MACE models. We further show SevenNet-0's parameters can be reduced by over 25% with no accuracy loss. These techniques enable more than 10x faster training of large-scale foundation models for ML potentials, potentially reshaping computational materials discovery.

Xinggong Zhang, Haotian Meng, Qingyang Li et al.

Malicious traffic detection is a pivotal technology for network security to identify abnormal network traffic and detect network attacks. Large Language Models (LLMs) are trained on a vast corpus of text, have amassed remarkable capabilities of context-understanding and commonsense knowledge. This has opened up a new door for network attacks detection. Researchers have already initiated discussions regarding the application of LLMs on specific cyber-security tasks. Unfortunately, there remains a lack of comprehensive analysis on harnessing LLMs for traffic detection, as well as the opportunities and challenges. In this paper, we focus on unleashing the full potential of Large Language Models (LLMs) in malicious traffic detection. We present a holistic view of the architecture of LLM-powered malicious traffic detection, including the procedures of Pre-training, Fine-tuning, and Detection. Especially, by exploring the knowledge and capabilities of LLM, we identify three distinct roles LLM can act in traffic classification: Classifier, Encoder, and Predictor. For each of them, the modeling paradigm, opportunities and challenges are elaborated. Finally, we present our design on LLM-powered DDoS detection as a case study. The proposed framework attains accurate detection on carpet bombing DDoS by exploiting LLMs' capabilities in contextual mining. The evaluation shows its efficacy, exhibiting a nearly 35% improvement compared to existing systems.

Hao Yi, Qingyang Li, Yulan Hu et al.

High-quality video-text preference data is crucial for Multimodal Large Language Models (MLLMs) alignment. However, existing preference data is very scarce. Obtaining VQA preference data for preference training is costly, and manually annotating responses is highly unreliable, which could result in low-quality pairs. Meanwhile, AI-generated responses controlled by temperature adjustment lack diversity. To address these issues, we propose a high-quality VQA preference dataset, called \textit{\textbf{M}ultiple \textbf{M}ultimodal \textbf{A}rtificial \textbf{I}ntelligence \textbf{P}reference Datasets in \textbf{V}QA} (\textbf{MMAIP-V}), which is constructed by sampling from the response distribution set and using an external scoring function for response evaluation. Furthermore, to fully leverage the preference knowledge in MMAIP-V and ensure sufficient optimization, we propose \textit{\textbf{Iter}ative \textbf{W}eak-to-\textbf{S}trong \textbf{R}einforcement \textbf{L}earning from \textbf{AI} \textbf{F}eedback for video MLLMs} (\textbf{Iter-W2S-RLAIF}), a framework that gradually enhances MLLMs' alignment capabilities by iteratively updating the reference model and performing parameter extrapolation. Finally, we propose an unbiased and information-complete evaluation scheme in VQA evaluation. Experiments demonstrate that MMAIP-V is beneficial for MLLMs in preference learning and Iter-W2S-RLAIF fully exploits the alignment information in MMAIP-V. We believe that the proposed automatic VQA preference data generation pipeline based on AI feedback can greatly promote future work in the MLLMs alignment. \textbf{Code and dataset are available} \href{https://anonymous.4open.science/r/MMAIP-V_Iter-W2S-RLAIF-702F}{MMAIP-V\_Iter-W2S-RLAIF-702F}.

Yulan Hu, Qingyang Li, Sheng Ouyang et al.

Reinforcement Learning from Human Feedback (RLHF) facilitates the alignment of large language models (LLMs) with human preferences, thereby enhancing the quality of responses generated. A critical component of RLHF is the reward model, which is trained on preference data and outputs a scalar reward during the inference stage. However, the collection of preference data still lacks thorough investigation. Recent studies indicate that preference data is collected either by AI or humans, where chosen and rejected instances are identified among pairwise responses. We question whether this process effectively filters out noise and ensures sufficient diversity in collected data. To address these concerns, for the first time, we propose a comprehensive framework for preference data collection, decomposing the process into four incremental steps: Prompt Generation, Response Generation, Response Filtering, and Human Labeling. This structured approach ensures the collection of high-quality preferences while reducing reliance on human labor. We conducted comprehensive experiments based on the data collected at different stages, demonstrating the effectiveness of the proposed data collection method.

Xiong-Hui Chen, Bowei He, Yang Yu et al.

Long-term user engagement (LTE) optimization in sequential recommender systems (SRS) is shown to be suited by reinforcement learning (RL) which finds a policy to maximize long-term rewards. Meanwhile, RL has its shortcomings, particularly requiring a large number of online samples for exploration, which is risky in real-world applications. One of the appealing ways to avoid the risk is to build a simulator and learn the optimal recommendation policy in the simulator. In LTE optimization, the simulator is to simulate multiple users' daily feedback for given recommendations. However, building a user simulator with no reality-gap, i.e., can predict user's feedback exactly, is unrealistic because the users' reaction patterns are complex and historical logs for each user are limited, which might mislead the simulator-based recommendation policy. In this paper, we present a practical simulator-based recommender policy training approach, Simulation-to-Recommendation (Sim2Rec) to handle the reality-gap problem for LTE optimization. Specifically, Sim2Rec introduces a simulator set to generate various possibilities of user behavior patterns, then trains an environment-parameter extractor to recognize users' behavior patterns in the simulators. Finally, a context-aware policy is trained to make the optimal decisions on all of the variants of the users based on the inferred environment-parameters. The policy is transferable to unseen environments (e.g., the real world) directly as it has learned to recognize all various user behavior patterns and to make the correct decisions based on the inferred environment-parameters. Experiments are conducted in synthetic environments and a real-world large-scale ride-hailing platform, DidiChuxing. The results show that Sim2Rec achieves significant performance improvement, and produces robust recommendations in unseen environments.

Qingyang Li, Zhiwen Yu, Lina Yao et al.

Identity recognition plays an important role in ensuring security in our daily life. Biometric-based (especially activity-based) approaches are favored due to their fidelity, universality, and resilience. However, most existing machine learning-based approaches rely on a traditional workflow where models are usually trained once for all, with limited involvement from end-users in the process and neglecting the dynamic nature of the learning process. This makes the models static and can not be updated in time, which usually leads to high false positive or false negative. Thus, in practice, an expert is desired to assist with providing high-quality observations and interpretation of model outputs. It is expedient to combine both advantages of human experts and the computational capability of computers to create a tight-coupling incremental learning process for better performance. In this study, we develop RLTIR, an interactive identity recognition approach based on reinforcement learning, to adjust the identification model by human guidance. We first build a base tree-structured identity recognition model. And an expert is introduced in the model for giving feedback upon model outputs. Then, the model is updated according to strategies that are automatically learned under a designated reinforcement learning framework. To the best of our knowledge, it is the very first attempt to combine human expert knowledge with model learning in the area of identity recognition. The experimental results show that the reinforced interactive identity recognition framework outperforms baseline methods with regard to recognition accuracy and robustness.

Mengyue Yang, Qingyang Li, Zhiwei Qin et al.

Contextual multi-armed bandit (MAB) achieves cutting-edge performance on a variety of problems. When it comes to real-world scenarios such as recommendation system and online advertising, however, it is essential to consider the resource consumption of exploration. In practice, there is typically non-zero cost associated with executing a recommendation (arm) in the environment, and hence, the policy should be learned with a fixed exploration cost constraint. It is challenging to learn a global optimal policy directly, since it is a NP-hard problem and significantly complicates the exploration and exploitation trade-off of bandit algorithms. Existing approaches focus on solving the problems by adopting the greedy policy which estimates the expected rewards and costs and uses a greedy selection based on each arm's expected reward/cost ratio using historical observation until the exploration resource is exhausted. However, existing methods are hard to extend to infinite time horizon, since the learning process will be terminated when there is no more resource. In this paper, we propose a hierarchical adaptive contextual bandit method (HATCH) to conduct the policy learning of contextual bandits with a budget constraint. HATCH adopts an adaptive method to allocate the exploration resource based on the remaining resource/time and the estimation of reward distribution among different user contexts. In addition, we utilize full of contextual feature information to find the best personalized recommendation. Finally, in order to prove the theoretical guarantee, we present a regret bound analysis and prove that HATCH achieves a regret bound as low as $O(\sqrt{T})$. The experimental results demonstrate the effectiveness and efficiency of the proposed method on both synthetic data sets and the real-world applications.

Qingyang Li, Guoqiang Zhong, Cui Xie

The ocean front has an important impact in many areas, it is meaningful to obtain accurate ocean front positioning, therefore, ocean front detection is a very important task. However, the traditional edge detection algorithm does not detect the weak edge information of the ocean front very well. In response to this problem, we collected relevant ocean front gradient images and found relevant experts to calibrate the ocean front data to obtain groundtruth, and proposed a weak edge identification nets(WEIN) for ocean front detection. Whether it is qualitative or quantitative, our methods perform best. The method uses a welltrained deep learning model to accurately extract the ocean front from the ocean front gradient image. The detection network is divided into multiple stages, and the final output is a multi-stage output image fusion. The method uses the stochastic gradient descent and the correlation loss function to obtain a good ocean front image output.

Wenjie Shang, Yang Yu, Qingyang Li et al.

Reinforcement learning aims at searching the best policy model for decision making, and has been shown powerful for sequential recommendations. The training of the policy by reinforcement learning, however, is placed in an environment. In many real-world applications, however, the policy training in the real environment can cause an unbearable cost, due to the exploration in the environment. Environment reconstruction from the past data is thus an appealing way to release the power of reinforcement learning in these applications. The reconstruction of the environment is, basically, to extract the casual effect model from the data. However, real-world applications are often too complex to offer fully observable environment information. Therefore, quite possibly there are unobserved confounding variables lying behind the data. The hidden confounder can obstruct an effective reconstruction of the environment. In this paper, by treating the hidden confounder as a hidden policy, we propose a deconfounded multi-agent environment reconstruction (DEMER) approach in order to learn the environment together with the hidden confounder. DEMER adopts a multi-agent generative adversarial imitation learning framework. It proposes to introduce the confounder embedded policy, and use the compatible discriminator for training the policies. We then apply DEMER in an application of driver program recommendation. We firstly use an artificial driver program recommendation environment, abstracted from the real application, to verify and analyze the effectiveness of DEMER. We then test DEMER in the real application of Didi Chuxing. Experiment results show that DEMER can effectively reconstruct the hidden confounder, and thus can build the environment better. DEMER also derives a recommendation policy with a significantly improved performance in the test phase of the real application.

Guoqiang Zhong, Xin Lin, Kang Chen et al.

Attention is an important cognition process of humans, which helps humans concentrate on critical information during their perception and learning. However, although many machine learning models can remember information of data, they have no the attention mechanism. For example, the long short-term memory (LSTM) network is able to remember sequential information, but it cannot pay special attention to part of the sequences. In this paper, we present a novel model called long short-term attention (LSTA), which seamlessly integrates the attention mechanism into the inner cell of LSTM. More than processing long short term dependencies, LSTA can focus on important information of the sequences with the attention mechanism. Extensive experiments demonstrate that LSTA outperforms LSTM and related models on the sequence learning tasks.

Jie Zhang, Qingyang Li, Richard J. Caselli et al.

Algorithmic image-based diagnosis and prognosis of neurodegenerative diseases on longitudinal data has drawn great interest from computer vision researchers. The current state-of-the-art models for many image classification tasks are based on the Convolutional Neural Networks (CNN). However, a key challenge in applying CNN to biological problems is that the available labeled training samples are very limited. Another issue for CNN to be applied in computer aided diagnosis applications is that to achieve better diagnosis and prognosis accuracy, one usually has to deal with the longitudinal dataset, i.e., the dataset of images scanned at different time points. Here we argue that an enhanced CNN model with transfer learning for the joint analysis of tasks from multiple time points or regions of interests may have a potential to improve the accuracy of computer aided diagnosis. To reach this goal, we innovate a CNN based deep learning multi-task dictionary learning framework to address the above challenges. Firstly, we pre-train CNN on the ImageNet dataset and transfer the knowledge from the pre-trained model to the medical imaging progression representation, generating the features for different tasks. Then, we propose a novel unsupervised learning method, termed Multi-task Stochastic Coordinate Coding (MSCC), for learning different tasks by using shared and individual dictionaries and generating the sparse features required to predict the future cognitive clinical scores. We apply our new model in a publicly available neuroimaging cohort to predict clinical measures with two different feature sets and compare them with seven other state-of-the-art methods. The experimental results show our proposed method achieved superior results.

Qingyang Li, Dajiang Zhu, Jie Zhang et al.

Genome-wide association studies (GWAS) have achieved great success in the genetic study of Alzheimer's disease (AD). Collaborative imaging genetics studies across different research institutions show the effectiveness of detecting genetic risk factors. However, the high dimensionality of GWAS data poses significant challenges in detecting risk SNPs for AD. Selecting relevant features is crucial in predicting the response variable. In this study, we propose a novel Distributed Feature Selection Framework (DFSF) to conduct the large-scale imaging genetics studies across multiple institutions. To speed up the learning process, we propose a family of distributed group Lasso screening rules to identify irrelevant features and remove them from the optimization. Then we select the relevant group features by performing the group Lasso feature selection process in a sequence of parameters. Finally, we employ the stability selection to rank the top risk SNPs that might help detect the early stage of AD. To the best of our knowledge, this is the first distributed feature selection model integrated with group Lasso feature selection as well as detecting the risk genetic factors across multiple research institutions system. Empirical studies are conducted on 809 subjects with 5.9 million SNPs which are distributed across several individual institutions, demonstrating the efficiency and effectiveness of the proposed method.

Qingyang Li, Tao Yang, Liang Zhan et al.

Genome-wide association studies (GWAS) offer new opportunities to identify genetic risk factors for Alzheimer's disease (AD). Recently, collaborative efforts across different institutions emerged that enhance the power of many existing techniques on individual institution data. However, a major barrier to collaborative studies of GWAS is that many institutions need to preserve individual data privacy. To address this challenge, we propose a novel distributed framework, termed Local Query Model (LQM) to detect risk SNPs for AD across multiple research institutions. To accelerate the learning process, we propose a Distributed Enhanced Dual Polytope Projection (D-EDPP) screening rule to identify irrelevant features and remove them from the optimization. To the best of our knowledge, this is the first successful run of the computationally intensive model selection procedure to learn a consistent model across different institutions without compromising their privacy while ranking the SNPs that may collectively affect AD. Empirical studies are conducted on 809 subjects with 5.9 million SNP features which are distributed across three individual institutions. D-EDPP achieved a 66-fold speed-up by effectively identifying irrelevant features.

Binbin Lin, Qingyang Li, Qian Sun et al.

\textit{Drosophila melanogaster} has been established as a model organism for investigating the fundamental principles of developmental gene interactions. The gene expression patterns of \textit{Drosophila melanogaster} can be documented as digital images, which are annotated with anatomical ontology terms to facilitate pattern discovery and comparison. The automated annotation of gene expression pattern images has received increasing attention due to the recent expansion of the image database. The effectiveness of gene expression pattern annotation relies on the quality of feature representation. Previous studies have demonstrated that sparse coding is effective for extracting features from gene expression images. However, solving sparse coding remains a computationally challenging problem, especially when dealing with large-scale data sets and learning large size dictionaries. In this paper, we propose a novel algorithm to solve the sparse coding problem, called Stochastic Coordinate Coding (SCC). The proposed algorithm alternatively updates the sparse codes via just a few steps of coordinate descent and updates the dictionary via second order stochastic gradient descent. The computational cost is further reduced by focusing on the non-zero components of the sparse codes and the corresponding columns of the dictionary only in the updating procedure. Thus, the proposed algorithm significantly improves the efficiency and the scalability, making sparse coding applicable for large-scale data sets and large dictionary sizes. Our experiments on Drosophila gene expression data sets demonstrate the efficiency and the effectiveness of the proposed algorithm.