Deep CFR enables end-to-end approximation of Nash equilibria in imperfect-information games(IIGs) but is sensitive to hyperparameters, making manual tuning inefficient. To address this, we propose PT-DCFR, which integrates Population-Based Training(PBT) with Deep CFR to dynamically optimize hyperparameters during training. Building upon this, we further introduce P2T-DCFR, which decouples parameter selection from model performance.

Accurately modeling opponent behaviors and integrating strategy are key challenges for multi-issue automated negotiation. Existing approaches often isolate preference learning or trend prediction and lack a unified cognitive structure with coordinated reasoning. This paper proposes a BDI (Belief-Desire-Intention)-based opponent modeling and strategy generation framework. The framework analyzes opponent responses (Belief), predicts preference weights and the utility function (Desire), and infers utilities of future offers (Intention). Building on these predictions, we design a responsive strategy, enabling gradual concessions and balanced outcomes. Our main contributions are: D-MBUE in the Desire module, I-DABI in the Intention module, and the BDI Negotiator on top of the modeling modules. Experiments on 45 standard negotiation domains and against 12 representative opponents demonstrate the effectiveness of our BDI framework.

With the evolving generative models, generated images are closer to reality, raising concerns about information authenticity and malicious misuse. Invisible watermarks offer a practical approach to detecting and tracing them. However, while image watermarking inevitably introduces quality degradation, most existing methods primarily focus on improving watermark robustness. To address this limitation, we propose AWMA-MoE, a framework that enhances the quality of generated images while preserving strong watermark robustness. Specifically, we design an attention-based adapter that adaptively embeds watermarks with spatially varying strengths across image regions. Building upon this, we introduce an MoE architecture that leverages diverse experts to further improve image quality while retaining watermark robustness. Experiments demonstrate that AWMA-MoE can reduce the distortion of generated images and exhibit competitive watermark performance, thus striking an improved balance for watermarking generated image tasks and better linking post-hoc and in-generation methods.

Retrieval-augmented generation (RAG) has become a powerful framework for enhancing large language models in knowledge-intensive and reasoning tasks. However, as reasoning chains deepen or search trees expand, RAG systems often face two persistent failures: evidence forgetting, where retrieved knowledge is not effectively used, and inefficiency, caused by uncontrolled query expansions and redundant retrieval. These issues reveal a critical gap between retrieval and evidence utilization in current RAG architectures. We propose PruneRAG, a confidence-guided query decomposition framework that builds a structured query decomposition tree to perform stable and efficient reasoning. PruneRAG introduces three key mechanisms: adaptive node expansion that regulates tree width and depth, confidence-guided decisions that accept reliable answers and prune uncertain branches, and fine-grained retrieval that extracts entity-level anchors to improve retrieval precision. Together, these components preserve salient evidence throughout multi-hop reasoning while significantly reducing retrieval overhead. To better analyze evidence misuse, we define the Evidence Forgetting Rate as a metric to quantify cases where golden evidence is retrieved but not correctly used. Extensive experiments across various multi-hop QA benchmarks show that PruneRAG achieves superior accuracy and efficiency over state-of-the-art baselines. The code is publicly available.

Multi-teacher knowledge distillation transfers knowledge from multiple large teacher models to a small student model and has performed well on many downstream tasks. However, when distilling knowledge from multiple teachers, it always suffers from the severe problems of being time-consuming and storage-extensive for multiple teacher models training and inference. We present MoE-KD, a simple but effective framework that produces supervision for training the student model from one single teacher model, which fixes the above problems and improves effectiveness. In the proposed MoE-KD, multiple trainable prompts are used to extract different views of samples from a single pre-trained language model and only a few parameters (prompts) need to be trained and stored. To guarantee the generated supervision signals with increased robustness and correctness, we introduce an uncertainty-based mechanism and a selector module, which routes the input instance to its corresponding teacher. We have also extended MoE KD to lifelong learning scenarios, proposing a lightweight solution for catastrophic forgetting. We conduct experiments on traditional KD scenarios and lifelong learning scenarios. MoE-KD yields improvements up to 1.1% and 140% in accuracy and efficiency in knowledge distillation and 2.8% improvements on average in lifelong learning, compared with the strong baseline methods.

Deep learning (DL) has achieved significant success across a wide range of domains, prompting the widespread deployment of GPU clusters equipped with specialized accelerators to support high-performance training workloads. To minimize operational costs while maximizing resource utilization, efficient job scheduling in these clusters is essential. Although recent schedulers have improved cluster efficiency through periodic reallocation or selection of GPU resources, they still face challenges such as preemption and migration overheads, along with the risk of degrading model accuracy. Despite these limitations, the potential of GPUsharing remains largely underexplored. Few existing studies have systematically examined GPU sharing as a strategy to enhance resource utilization and reduce job queuing delays in multi-tenant DL clusters. Motivated by these insights, we propose a job scheduling model that enables multiple jobs to share the same set of GPUs without modifying their original training configurations. We introduce Castor, a simple yet efficient scheduling system, to achieve intelligent GPU colocation for multiple DL jobs. Castor intelligently selects job pairs for GPU sharing and determines runtime parameters (sub-batch size and scheduling time point) to optimize overall system performance while preserving the accuracy of DL convergence through gradient accumulation. Through a combination of physical DL workloads and trace-driven simulations across various configurations, we demonstrate that Castor reduces average job completion time by 26–52% compared to state-of-the-art preemptive DL schedulers, despite operating under a preemption-free policy. Furthermore, Castor effectively identifies optimal resource-sharing configurations, outperforming the baseline first-fit sharing policy (SJF-FFS) by up to 20% on large-scale workload traces.

Ethereum 2.0 (ETH2) marks a pivotal shift in blockchain technology, transitioning from a Proof-of-Work (PoW) to a Proof-of-Stake (PoS) consensus mechanism, with Gasper at its core. While this evolution promises enhanced scalability and energy efficiency, the performance of its block proposal stage is highly sensitive to network latency and system parameters, such as slot length. This sensitivity introduces a critical trade-off between throughput and security, measured by the probability of blockchain forking. This paper reveals that network latency is not just a passive risk but an exploitable attack surface. We introduce the "adaptive latency-driven equivocation attack", a novel adversarial strategy where an attacker deliberately creates forks while mimicking the behavior of a high-latency node, thus achieving plausible deniability. To formally analyze and quantify the impact of this threat, we develop a comprehensive theoretical model by using Markov chains to analyze the fork probability and throughput of the Gasper's block proposal mechanism under both honest and adversarial conditions. Through extensive simulations, we validate the accuracy of our model in both normal and bursty traffic conditions. Our findings provide a systematic methodology for optimizing system parameters to achieve a robust balance between performance and security, offering a foundational guide for configuring ETH2 networks against sophisticated, latency-based threats.

Multi-view Multi-label Learning (MVML) aims to leverage multi-view information from input samples to achieve accurate predictions of multiple labels. Unfortunately, most existing MVML methods operate under the assumption of data completeness, which makes them ineffective in practical scenarios involving missing views or uncertain labels. Recent methods address incomplete data, but few approaches handle scenarios where both views and labels are missing. To address this challenge, we propose a Dual-view Feature-guided Fusion Learning (DFFL) framework. DFFL considers both view-specific unique features and inter-view consistent features. Specifically, DFFL constructs view-uniqueness contrastive learning to ensure that features within the same view maintain high semantic relevance under the condition of view missing, while the semantics between different views are different. Unlike previous methods, DFFL assumes that label relevance can be reversely mapped to high-dimensional features. By establishing View-consistency learning, the mutual information in the shared embedding space is maximized to achieve consistent feature alignment. In particular, DFFL minimizes the conditional entropy of the marginal distribution of multi-view features through dual prediction, thereby deriving the maximum joint distribution of feature fusion and combining the missing view index matrix to achieve feature fusion. This process can effectively alleviate the fusion feature suppression existing in previous methods. Finally, the missing label index matrix is combined with the fusion feature to complete the classification task. We validate the framework on five widely used datasets, and experimental results demonstrate that our approach achieves superior performance compared to state-of-the-art methods. Ablation studies further validated the effectiveness of each component in DFFL.

In recent years, reinforcement learning has been widely applied in the field of games. However, most studies focus on assisting agents to achieve victory, with less attention paid to whether the agents exhibit human-like characteristics. In order to build human-like agents with high performance, we propose a method for learning the strategies of human players in modern three-dimensional video games. Our method utilizes a hierarchical framework, learning basic behaviors and intentions of human players at the lower level through imitation learning, and generalized policies at the high level through reinforcement learning. Compared with other existing methods, our method demonstrates significant advantages in learning human-like strategies in complex environments.

Dataset distillation is an emerging dataset reduction method, which condenses large-scale datasets while maintaining task accuracy. Current parameterization methods achieve enhanced performance under extremely high compression ratio by optimizing determined synthetic dataset in informative feature domain. However, they limit themselves to a fixed optimization space for distillation, neglecting the diverse guidance across different informative latent spaces. To overcome this limitation, we propose a novel parameterization method dubbed Hierarchical Parameterization Distillation (H-PD), to systematically explore hierarchical feature within provided feature space (e.g., layers within pre-trained generative adversarial networks). We verify the correctness of our insights by applying the hierarchical optimization strategy on GAN-based parameterization method. In addition, we introduce a novel class-relevant feature distance metric to alleviate the computational burden associated with synthetic dataset evaluation, bridging the gap between synthetic and original datasets. Experimental results demonstrate that the proposed H-PD achieves a significant performance improvement under various settings with equivalent time consumption, and even surpasses current generative distillation using diffusion models under extreme compression ratios IPC=1 and IPC=10. Our code is available at https://github.com/ndhg1213/H-PD

Recent advances in Multimodal Large Language Models (MLLMs) have enhanced their versatility as they integrate a growing number of modalities.Considering the heavy cost of training MLLMs, it is efficient to reuse the existing ones and extend them to more modalities through Modality-incremental Continual Learning (MCL).The exploration of MCL is in its early stages.In this work, we dive into the causes of performance degradation in MCL.We uncover that it suffers not only from forgetting as in traditional continual learning, but also from misalignment between the modality-agnostic and modality-specific components.To this end, we propose an elegantly simple MCL paradigm called "MErge then ReAlign" (MERA) to address both forgetting and misalignment.MERA avoids introducing heavy model budgets or modifying model architectures, hence is easy to deploy and highly reusable in the MLLM community.Extensive experiments demonstrate the impressive performance of MERA, holding an average of 99.84% Backward Relative Gain when extending to four modalities, achieving nearly lossless MCL performance.Our findings underscore the misalignment issue in MCL.More broadly, our work showcases how to adjust different components of MLLMs during continual learning.

Microphone array techniques are widely used in sound source localization and smart city acoustic-based traffic monitoring, but these applications face significant challenges due to the scarcity of labeled real-world traffic audio data and the complexity and diversity of application scenarios. The DCASE Challenge’s Task 10 focuses on using multi-channel audio signals to count vehicles (cars or commercial vehicles) and identify their directions (left-to-right or vice versa). In this paper, we propose a graph-enhanced dual-stream feature fusion network (GEDF-Net) for acoustic traffic monitoring, which simultaneously considers vehicle type and direction to improve detection. We propose a graph-enhanced dual-stream feature fusion strategy which consists of a vehicle type feature extraction (VTFE) branch, a vehicle direction feature extraction (VDFE) branch, and a frame-level feature fusion module to combine the type and direction feature for enhanced performance. A pre-trained model (PANNs) is used in the VTFE branch to mitigate data scarcity and enhance the type features, followed by a graph attention mechanism to exploit temporal relationships and highlight important audio events within these features. The frame-level fusion of direction and type features enables fine-grained feature representation, resulting in better detection performance. Experiments demonstrate the effectiveness of our proposed method. GEDF-Net is our submission that achieved 1st place in the DCASE 2024 Challenge Task 10.

Multi-Agent Reinforcement Learning (MARL) has shown great success in many aspects. However, the cooperative policy trained by MARL is vulnerable to adversarial attacks towards agents' observations, which could cause immeasurable damage to the agent team. A few techniques have been developed to test the robustness of MARL, but the feasibility of real implementation is not carefully considered. In this work, we propose a two-step framework to conduct destructive and sparse attacks under realistic conditions. The first step contains two attack scenes: Ally Observation Attack (AOA) and Enemy Observation Attack (EOA), which have limitations on the use of agents' observations. The first step selects victim agent from the team and screens the attack time, while the second step generates adversarial perturbation and adds it to the observation of the chosen victim in a black-box environment. To the best of our knowledge, this is the first work to test the robustness with partial observation and conduct the test in a black-box environment. Experiments on SMAC environments demonstrate that our methods show great performance in reducing the win rate and the team reward of the agent team trained by QMIX algorithm with lower perturbation steps.

The shopping query suggestion offers personalized queries to users and plays a crucial role in search engines. However, existing shopping query suggestion methods suffer from poor task generalization and limited semantic comprehension problems. This paper presents a comprehensive framework for the shopping query suggestion that effectively addresses the shortcomings of existing approaches. Our proposed framework leverages a generative language model and fine-grained preference alignment to enhance semantic comprehension and improve the quality of generated queries. Our key contributions include the introduction of a personalized prompt set for diverse query suggestion tasks, the integration of interaction behavior time to capture user query interests, and the utilization of reinforcement learning techniques to align user preferences. Experimental results demonstrate enhancements in different scenarios. Our codes are available at https://github.com/1170300319/CIKM2024_SOUP.

Deep learning (DL) has demonstrated significant success across diverse fields, leading to the construction of dedicated GPU accelerators within GPU clusters for high-quality training services. Efficient scheduler designs for such clusters are vital to reduce operational costs and enhance resource utilization. While recent schedulers have shown impressive performance in optimizing DL job performance and cluster utilization through periodic reallocation or selection of GPU resources, they also encounter challenges such as preemption and migration overhead, along with potential DL accuracy degradation. Nonetheless, few explore the potential benefits of GPU sharing to improve resource utilization and reduce job queuing times.Motivated by these insights, we present a job scheduling model allowing multiple jobs to share the same set of GPUs without altering job training settings. We introduce SJF-BSBF (shortest job first with best sharing benefit first), a straightforward yet effective heuristic scheduling algorithm. SJF-BSBF intelligently selects job pairs for GPU resource sharing and runtime settings (sub-batch size and scheduling time point) to optimize overall performance while ensuring DL convergence accuracy through gradient accumulation. In experiments with both physical DL workloads and trace-driven simulations, even as a preemptionfree policy, SJF-BSBF reduces the average job completion time by 27-33% relative to the state-of-the-art preemptive DL schedulers. Moreover, SJF-BSBF can wisely determine the optimal resource sharing settings, such as the sharing time point and sub-batch size for gradient accumulation, outperforming the aggressive GPU sharing approach (baseline SJF-FFS policy) by up to 17% in large-scale traces.

Deep reinforcement learning (DRL) has garnered remarkable success across various domains, propelled by advancements in deep learning (DL) technologies. However, the opacity of DL presents significant challenges, limiting the application of DRL in critical systems. In response, decision tree (DT)-based methods, known for their transparent decision-making mechanisms, have shown promise in making interpretable policies for decision-making problems. Existing methods often employ differential DTs to model RL policies and discretize them to conventional DTs for higher interpretability. Yet, this method leads to discrepancies between the trained differential DTs and the discretized DTs. To address this issue, we introduce Generative Consistent Trees (GCTs), a novel solution that circumvents the information loss typically associated with the argmax operation in prior research. By implementing a reparameterization technique to approximate the categorical distribution, GCTs ensure the consistencies between trained GCTs and discretized counterparts. Moreover, we have developed an imitation-learning-based framework for interpretable reinforcement learning. This framework is designed to train GCTs by efficiently mimicking expert policies. Our extensive experiments across multiple environments have validated the effectiveness of this approach, highlighting the potential of GCTs in enhancing the interpretability and applicability of DRL.

Pre-trained models are frequently employed in multimodal learning. However, these models have too many parameters and need too much effort to fine-tune the downstream tasks. Knowledge distillation (KD) is a method to transfer knowledge using the soft label from this pre-trained teacher model to a smaller student, where the parameters of the teacher are fixed (or partially) during training. Recent studies show that this mode may cause difficulties in knowledge transfer due to the mismatched model capacities. To alleviate the mismatch problem, adjustment of temperature parameters, label smoothing and teacher-student joint training methods (online distillation) to smooth the soft label of a teacher network, have been proposed. But those methods rarely explain the effect of smoothed soft labels to enhance the KD performance. The main contributions of our work are the discovery, analysis, and validation of the effect of the smoothed soft label and a less time-consuming and adaptive transfer of the pre-trained teacher's knowledge method, namely PESF-KD by adaptive tuning soft labels of the teacher network. Technically, we first mathematically formulate the mismatch as the sharpness gap between teacher's and student's predictive distributions, where we show such a gap can be narrowed with the appropriate smoothness of the soft label. Then, we introduce an adapter module for the teacher and only update the adapter to obtain soft labels with appropriate smoothness. Experiments on various benchmarks including CV and NLP show that PESF-KD can significantly reduce the training cost while obtaining competitive results compared to advanced online distillation methods.

Counterfactual regret minimization (CFR) is a popular method for finding approximate Nash equilibrium in two-player zero-sum games with imperfect information. Solving large-scale games with CFR needs a combination of abstraction techniques and certain expert knowledge, which constrains its scalability. Recent neural-based CFR methods mitigate the need for abstraction and expert knowledge by training an efficient network to directly obtain counterfactual regret without abstraction. However, these methods only consider estimating regret values for individual actions, neglecting the evaluation of state values, which are significant for decision-making. In this article, we introduce deep dueling CFR (D2CFR), which emphasizes the state value estimation by employing a novel value network with a dueling structure. Moreover, a rectification module based on a time-shifted Monte Carlo simulation is designed to rectify the inaccurate state value estimation. Extensive experimental results are conducted to show that D2CFR converges faster and outperforms comparison methods on test games.

Modeling the interactive relationships of agents is critical to improving the collaborative capability of a multiagent system. Some methods model these by predefined rules. However, due to the nonstationary problem, the interactive relationship changes over time and cannot be well captured by rules. Other methods adopt a simple mechanism such as an attention network to select the neighbors the current agent should collaborate with. However, in large-scale multiagent systems, collaborative relationships are too complicated to be described by a simple attention network. We propose an adaptive and gated graph attention network (AGGAT), which models the interactive relationships between agents in a cascaded manner. In the AGGAT, we first propose a graph-based hard attention network that roughly filters irrelevant agents. Then, normal soft attention is adopted to decide the importance of each neighbor. Finally, gated attention further refines the collaborative relationship of agents. By using cascaded attention, the collaborative relationship of agents is precisely learned in a coarse-to-fine style. Extensive experiments are conducted on a variety of cooperative tasks. The results indicate that our proposed method outperforms state-of-the-art baselines.

Most existing chemical experiment teaching systems lack solid immersive experiences, making it difficult to engage students. To address these challenges, we propose a chemical simulation teaching system based on virtual reality and gesture interaction. The parameters of the models were obtained through actual investigation, whereby Blender and 3DS MAX were used to model and import these parameters into a physics engine. By establishing an interface for the physics engine, gesture interaction hardware, and virtual reality (VR) helmet, a highly realistic chemical experiment environment was created. Using code script logic, particle systems, as well as other systems, chemical phenomena were simulated. Furthermore, we created an online teaching platform using streaming media and databases to address the problems of distance teaching. The proposed system was evaluated against two mainstream products in the market. In the experiments, the proposed system outperformed the other products in terms of fidelity and practicality. The proposed system which offers realistic simulations and practicability, can help improve the high school chemistry experimental education.

Although the vision-and-language pretraining (VLP) equipped cross-modal image-text retrieval (ITR) has achieved remarkable progress in the past two years, it suffers from a major drawback: the ever-increasing size of VLP models restrict its deployment to real-world search scenarios (where the high latency is unacceptable). To alleviate this problem, we present a novel plug-in dynamic contrastive distillation (DCD) framework to compress the large VLP models for the ITR task. Technically, we face the following two challenges: 1) the typical uni-modal metric learning approach is difficult to directly apply to cross-modal task, due to the limited GPU memory to optimize too many negative samples during handling cross-modal fusion features. 2) it is inefficient to static optimize the student network from different hard samples, which have different effects on distillation learning and student network optimization. We try to overcome these challenges from two points. First, to achieve multi-modal contrastive learning, and balance the training costs and effects, we propose to use a teacher network to estimate the difficult samples for students, making the students absorb the powerful knowledge from pre-trained teachers, and master the knowledge from hard samples. Second, to dynamic learn from hard sample pairs, we propose dynamic distillation to dynamically learn samples of different difficulties, from the perspective of better balancing the difficulty of knowledge and students' self-learning ability. We successfully apply our proposed DCD strategy on two state-of-the-art vision-language pretrained models, i.e. ViLT and METER. Extensive experiments on MS-COCO and Flickr 30 K benchmarks show the effectiveness and efficiency of our DCD framework. Encouragingly, we can speed up the inference at least 129 × compared to the existing ITR models. We further provide in-depth analyses and discussions that explain where the performance improvement comes from. We hope our work can shed light on other tasks that require distillation and contrastive learning.

App reviews are crowdsourcing knowledge of user experience with the apps, providing valuable information for app release planning, such as major bugs to fix and important features to add. There exist prior explorations on app review mining for release planning; however, most of the studies strongly rely on predefined classes or manually annotated reviews. Also, the new review characteristic, i.e., the number of users who rated the review as helpful, which can help capture important reviews, has not been considered previously. In the article, we propose a novel framework, named SOLAR, aiming at accurately summarizing helpful user reviews to developers. The framework mainly contains three modules: the review helpfulness prediction module, topic-sentiment modeling module, and multifactor ranking module. The review helpfulness prediction module assesses the helpfulness of reviews, i.e., whether the review is useful for developers. The topic-sentiment modeling module groups the topics of the helpful reviews and also predicts the associated sentiment, and the multifactor ranking module aims at prioritizing semantically representative reviews for each topic as the review summary. Experiments on five popular apps indicate that SOLAR is effective for review summarization and promising for facilitating app release planning.

Due to the problem of severe deformation, occlusion, diversified scenarios, general object detection methods cannot achieve satisfactory results in underwater object detection tasks. In this paper, we propose a two-stage Underwater Small Target Detection (USTD) network. In the proposed USTD, the Deformable Convolutional Pyramid(DCP) is proposed to deal with the problems of deformation, occlusion, and various object sizes effectively. Besides, we also propose a strategy of domain generalization based on curriculum learning to improve generalization in multi-domain environments, which is named as Phased Learning. Afterward, we construct an underwater target detection set (UTDS) to evaluate the accuracy of our method in underwater target detection tasks. Our method shows superior detection performance in experiments and reaches state-of-the-art for underwater target detection. Finally, in the 2020 China Underwater Robot Professional Contest (URPC), our method reached third place in terms of accuracy.

This article aims to provide the information retrieval community with some reflections on recent advances in retrieval learning by analyzing the reproducibility of image-text retrieval models. Due to the increase of multimodal data over the last decade, image-text retrieval has steadily become a major research direction in the field of information retrieval. Numerous researchers train and evaluate image-text retrieval algorithms using benchmark datasets such as MS-COCO and Flickr30k. Research in the past has mostly focused on performance, with multiple state-of-the-art methodologies being suggested in a variety of ways. According to their assertions, these techniques provide improved modality interactions and hence more precise multimodal representations. In contrast to previous works, we focus on the reproducibility of the approaches and the examination of the elements that lead to improved performance by pretrained and nonpretrained models in retrieving images and text.

Astounding results from transformer models with Vision-and Language Pretraining (VLP) on joint vision-and-language downstream tasks have intrigued the multi-modal community. On the one hand, these models are usually so huge that make us more difficult to fine-tune and serve real-time online applications. On the other hand, the compression of the original transformer block will ignore the difference in information between modalities, which leads to the sharp decline of retrieval accuracy.

How to track the target robustly is a challenging task in the field of computer vision. Occlusion as one of the most difficult problems, occurs due to the information lost when three-dimensional subjects are projected in two-dimensional interface, therefore, the 2D or 3D tracking algorithms which adopted depth information that expects to rely on three-dimensional special structure to resolve these problems and made somewhat progress. The 2D tracking algorithm is not efficient in fully using depth information, and the 3D tracking method is not robust because of the lack of mature 3D feature extraction method, which fairly restricts the actual tracking effect. Responding to above questions, we propose an adoption of adaptive quantified depth information, establish an adaptive hierarchical structure according to various scenarios. Hierarchical structure can filter the foreground and background information to reduce the interference in tracking, at the same time simplify the use of the depth information. Combined with kernel correlation filter tracking method, we design the algorithm using 2D apparent model under the spatial structures, which is efficient to deal with the problems of occlusion and the change of target scale, and prove its effectiveness on Princeton Tracking Dataset.

While the artificial intelligence theory for solving the perfect-information games has been well developed in recent years, great challenges are still posed in dealing with the imperfect-information game due to the huge state space and hidden information involved in it. In this paper, we design an online strategy solving framework for six-player no-limit Texas hold'em poker. Based on hand isomorphism and hand strength evalution, the framework provides an efficient situation estimation method for six-player poker. Such method could greatly reduce the the state space in six-player poker as well as effectively evaluate the current hands. The poker agent based on our method won the third place in the 2018 AAAI-ACPC.

Due to the development of deep learning and Fully Convolutional Neural Network (FCN), the research on salient object detection has made great progress in recent years. However, such FCN based models are always affected by the scale-space problem, which reduces the saliency detection accuracy and leads to a blurred object boundary. In this paper, we propose a novel saliency object detection method, which predicts the saliency by incorporating both pixel-level and region-level predictions. First, in order to alleviate the scale-space problem, modified dilated convolution layers and short connections are integrated into the FCN model, and the pixel-level saliency maps is generated by a pixel-wise salient classifier. Then, we employ a superpixel based manifold learning algorithm to obtain a better boundary of salient object, by which a region-level saliency map with clearer object boundary is generated. At last, a simple fusion method is utilized to fuse the two saliency maps into a unified saliency map, followed by a DenseCRF post-refinement module to further optimize the final results. Experiments are conducted on two benchmark datasets to demonstrate the effectiveness of our method.

Learning to hash is widely used in approximate nearest-neighbor (ANN) search. However, traditional hash learning methods, which split the hashing into two parts: feature extraction and hash function learning, usually result in a low retrieval accuracy. Although existing deep learning based hashing methods can improve hashing quality by coupling feature learning and hash encoding, they are always affected by the positive-negative sample imbalance problem. It often deteriorates the performance of the generated hash code. In this paper, we propose an end-to-end deep hashing framework, in which a weighted pairwise loss function is employed to alleviate sample imbalance problem. The loss generated by the positive pairs and negative pairs are given different weights automatically. Moreover, we integrate a classification network into the hashing framework, which can preserve the semantic information by making sure the generated hash codes are also optimal for classification. Comparison experiments are conducted on two benchmark datasets to demonstrate the performance of our proposed approach.

In sharp contrast to the traditional category/subcategory level image retrieval, product image search aims to find the images containing the exact same product. This is a challenging problem because in addition to being robust under different imaging conditions such as varying viewpoints and illumination changes, the features should also be able to distinguish the specific product among many similar products. Consequently, it is important to utilize a large dataset, containing many product classes, to learn a strongly discriminative representation. Building such a dataset requires laborious manual annotation. Toward learning fine-grained, robust, discriminative features for product image search, we present a novel paradigm that can construct the required dataset without any human annotation. Unlike other fine-grained recognition works that rely on high-quality annotated datasets and are very narrowly focused on a specific object category, our method handles multiple object classes and requires minimum human effort. First, an ImageNet pretrained model is used to generate product clusters. As the original features from ImageNet are not discriminative, the clusters generated by this unsupervised procedure contain much noise. We alleviate noise by explicitly modeling noise distribution and automatically detecting errors during learning. The proposed paradigm is general, requires minimum human efforts, and is applicable to any deep learning task where fine-grained discriminative features are desired. Extensive experiments on the ALISC dataset have demonstrated that our approach is sound and effective, surpassing the baseline GoogleNet model by 15.09%.