Deep hash networks are widely used in tasks such as large-scale image retrieval due to high search efficiency and low storage costs through binary hash codes. With the growing demand for deploying deep hash networks on resource-constrained devices, it is crucial to perform network compression on them, in which automatic pruning constitutes a priority option owing to efficacy maintenance. However, existing pruning methods are mostly designed for image classification, while hashing networks must generate compact binary codes, making each channel more sensitive to retrieval objectives. As a result, their performance often degrades when applied to image retrieval tasks. In this paper, we propose a novel Automatic Channel Pruning framework by Searching with Structure Embedding (ACP-SSE). To the best of our knowledge, this is the first study to explore pruning techniques for deep hash networks and the first automatic pruning method by searching based on network topology structure. Specifically, we first design a structure encoding model by Graph Convolutional Networks (GCNs) whose graph is constructed by hash network and nodes' features are initialized by pruning strategies. The model is trained by contrastive learning loss efficiently without accuracy supervision by fine-tuning pruned models. In addition, we introduce a dynamic pruning search space in consideration of the resource constraints. By converting the automatic channel pruning task into searching the pruned structure with effect similar to the unpruned structure, it enables the method to adapt to various network architectures. Finally, the optimal networks are selected from the candidate set according to their performance in specific downstream tasks. Extensive experiments demonstrate that ACP-SSE indeed works in the automatic channel pruning area, outperforming state-of-the-art baselines in hashing-based image retrieval, while maintaining competitive accuracy in image classification.

Trajectory similarity computation plays a critical role in a wide range of trajectory-related applications, including transportation optimization and behavior study. Most studies aim at learning discriminative real-valued trajectory representations. However, these methods struggle to scale to large datasets due to their linear time complexity. To address this problem, only one hypergraph hash learning approach (HHL-Traj) has been proposed to realize efficient trajectory similarity computation by calculating Hamming distances among the trajectory hash codes. Nevertheless, it fails to effectively integrate the spatial and temporal information of trajectory data with semantic relevance. In this paper, we present a novel Trajectory Similarity Hash Learning method with Spatio-Temporal GRU (TrajH-ST), which fuses the spatial and temporal features through reset and update gates within the network architecture. Additionally, we design an alternating sampling strategy to generate two sub-trajectories for contrastive learning, which enhances both generalization and robustness compared to nonuniform sampling techniques. To optimize the proposed end-to-end model, we develop an objective function that incorporates InfoNCE loss, alignment loss, and quantization loss. Extensive experiments on two widely-used trajectory datasets demonstrate that the proposed model consistently outperforms state-of-the-art baselines, achieving accuracy improvements of up to 5.58% and 4.33% with real-valued and binary features, respectively. Our code is available athttps://github.com/caoyuan618/Traj-ST

Cross-modal hashing enables efficient retrieval across different modalities by mapping heterogeneous data into compact binary codes within a shared Hamming space. However, most existing methods assume that data from each class are evenly distributed, which contradicts the long-tailed nature of real-world data. Consequently, these approaches often exhibit suboptimal performance when handling imbalanced datasets. The only existing cross-modal hashing method that considers long-tailed data attempts to mine both the individuality and commonality across modalities, yet it relies on a negative log-likelihood pairwise loss that tends to bias the model toward head categories. To address this issue, we propose a novel Long-tailed Approaching Cross-modal Hashing (LACH) framework based on multi-expert collaborative learning. Specifically, LACH constructs a multi-expert architecture with a Graph Convolutional Network (GCN) backbone to facilitate knowledge transfer. Unlike conventional multi-expert models that either share identical data copies or employ entirely distinct data subsets, we introduce a partial data replication strategy that ensures each expert receives a balanced yet overlapping training set. Furthermore, we design a proxy-based pointwise loss to treat head and tail categories equitably, along with an inter-modal approaching loss to enhance the alignment of hash codes across modalities within each class. Extensive experiments demonstrate that LACH achieves accuracy improvements of up to 4.2% and 6.3% over state-of-the-art baselines on balanced and long-tailed datasets, respectively. Our code is available at https://github.com/caoyuan618/LACH.

Online hashing has attracted much research attention for large-scale image retrieval in a streaming way. The main challenge lies in keeping balance between high retrieval accuracy and low training time. Existing online hashing methods almost rely on shallow models rather than deep networks due to high training costs, because it is unacceptable to update hash functions on an order of hours. In addition, the multi-label supervision information is not fully utilized to guide the hash learning process and the affinity matrix is always fixed once constructed. In this paper, we propose a novel Deep Graph Online Hashing (DGOH) method, which for the first time introduces inductive graph neural networks (GNNs) to realize deep online hashing with acceptable training costs on an order of seconds. Furthermore, we mine the multi-label information of the images by constructing a label network and learn label-wise weights dynamically to help to update the affinity matrix. In addition, we provide a strategy to obtain examples from the old data to solve the catastrophic forgetting problem. An integrated objective function is designed to train the entire architecture. Extensive experiments on two common benchmarks demonstrate that the proposed method achieves up to 13.3% accuracy gains over state-of-the-art baselines and shows competitive performance on training time.

Large language models (LLMs) have shown remarkable progress in understanding and generating natural language across various applications. However, they often struggle with resolving ambiguities in real-world, enterprise-level interactions, where context and domain-specific knowledge play a crucial role. In this demonstration, we introduce ECLAIR (Enhanced CLArification for Interactive Responses), a multi-agent framework for interactive disambiguation. ECLAIR enhances ambiguous user query clarification through an interactive process where custom agents are defined, ambiguity reasoning is conducted by the agents, clarification questions are generated, and user feedback is leveraged to refine the final response. When tested on real-world customer data, ECLAIR demonstrates significant improvements in clarification question generation compared to standard few-shot methods.

We present ECLAIR (Enhanced CLArification for Interactive Responses), a novel unified and end-to-end framework for interactive disambiguation in enterprise AI assistants. ECLAIR generates clarification questions for ambiguous user queries and resolves ambiguity based on the user's response. We introduce a generalized architecture capable of integrating ambiguity information from multiple downstream agents, enhancing context-awareness in resolving ambiguities and allowing enterprise specific definition of agents. We further define agents within our system that provide domain-specific grounding information. We conduct experiments comparing ECLAIR to few-shot prompting techniques and demonstrate ECLAIR's superior performance in clarification question generation and ambiguity resolution.

Reinforcement learning (RL) has proven to be well-performed and versatile in inventory control (IC). However, further improvement of RL algorithms in the IC domain is impeded by two limitations of online experience. First, online experience is expensive to acquire in real-world applications. With the low sample efficiency nature of RL algorithms, it would take extensive time to collect enough data and train the RL policy to convergence. Second, online experience may not reflect the true demand due to the lost-sales phenomenon typical in IC, which makes the learning process more challenging. To address the above challenges, we propose a training framework that combines reinforcement learning with feedback graph (RLFG) and intrinsically motivated exploration (IME) to boost sample efficiency. In particular, we first leverage the MDP structure inherent in lost-sales IC problems and design the feedback graph (FG) tailored to lost-sales IC problems to generate abundant side experiences aiding in RL updates. Then we conduct a rigorous theoretical analysis of how the designed FG reduces the sample complexity of RL methods. Guided by these insights, we design an intrinsic reward to direct the RL agent to explore to the state-action space with more side experiences, further exploiting FG’s capability. Experimental results on single-item, multi-item, and multi-echelon environments demonstrate that our method greatly improves the sample efficiency of applying RL in IC. Our code is available at \url{https://github.com/Ziffer-byakuya/RLIMFG4IC}

Safe offline reinforcement learning aims to develop policies that maximize cumulative rewards while satisfying safety constraints without the need for risky online interaction. However, existing methods often struggle with the out-of-distribution (OOD) problem, leading to potentially unsafe and suboptimal policies. To address this issue, we first propose Constrained Implicit Q-learning (CIQL), a novel algorithm designed to avoid the OOD problem. In particular, CIQL expands the implicit update of reward value functions to constrained settings and then estimates cost value functions under the same implicit policy. Despite its advantages, the further performance improvement of CIQL is still hindered by the inaccurate discounted approximations of constraints. Thus, we further propose Constraint-Conditioned Implicit Q-learning (C2IQL). Building upon CIQL, C2IQL employs a cost reconstruction model to derive non-discounted cumulative costs from discounted values and incorporates a flexible, constraint-conditioned mechanism to accommodate dynamic safety constraints. Experiment results on DSRL benchmarks demonstrate the superiority of C2IQL compared to baseline methods in achieving higher rewards while guaranteeing safety constraints under different threshold conditions.

With the rapid expansion of online information, cross-modal retrieval has emerged as a crucial and dynamic research focus. Deep hashing has gained significant traction in this field due to its efficiency in storage and retrieval speed, making it particularly valuable for remote sensing multi-modal retrieval. However, existing deep cross-modal hashing techniques often rely on parallel network structures for processing different modalities, overlooking a unified representation that captures cross-modal visual information. To address this limitation, we introduce a novel unsupervised cross-modal hashing framework that incorporates two modality-specific encoders and a fusion module. This fusion module facilitates modality interaction, enabling the extraction of meaningful semantic relationships across different data types. To ensure comprehensive similarity preservation, we design an integrated objective function that incorporates inter-modal and intra-modal constraints, joint consistency, and binary alignment losses. Furthermore, instead of conventional convolutional networks, we adopt the Swin Transformer as the backbone to enhance the discriminative power of image features. Our approach achieves an average 2.3% improvement in mAP on remote sensing cross-modal retrieval tasks compared to existing methods. The implementation is available at https://github.com/sellaner/TUCH.

In multi-agent reinforcement learning (MARL), effective exploration is critical, especially in sparse reward environments. Although introducing global intrinsic rewards can foster exploration in such settings, it often complicates credit assignment among agents. To address this difficulty, we propose Individual Contributions as intrinsic Exploration Scaffolds (ICES), a novel approach to motivate exploration by assessing each agent’s contribution from a global view. In particular, ICES constructs exploration scaffolds with Bayesian surprise, leveraging global transition information during centralized training. These scaffolds, used only in training, help to guide individual agents towards actions that significantly impact the global latent state transitions. Additionally, ICES separates exploration policies from exploitation policies, enabling the former to utilize privileged global information during training. Extensive experiments on cooperative benchmark tasks with sparse rewards, including Google Research Football (GRF) and StarCraft Multi-agent Challenge (SMAC), demonstrate that ICES exhibits superior exploration capabilities compared with baselines. The code is publicly available at https://github.com/LXXXXR/ICES.

Finetuning foundation models for specific tasks is an emerging paradigm in modern machine learning. The efficacy of task-specific finetuning largely depends on the selection of appropriate training data. We present TSDS (Task-Specific Data Selection), a framework to select data for task-specific model finetuning, guided by a small but representative set of examples from the target task. To do so, we formulate data selection for task-specific finetuning as an optimization problem with a distribution alignment loss based on optimal transport to capture the discrepancy between the selected data and the target distribution. In addition, we add a regularizer to encourage the diversity of the selected data and incorporate kernel density estimation into the regularizer to reduce the negative effects of near-duplicates among the candidate data. We connect our optimization problem to nearest neighbor search and design efficient algorithms to compute the optimal solution based on approximate nearest neighbor search techniques. We evaluate our method on data selection for both continued pretraining and instruction tuning of language models. We show that instruction tuning using data selected by our method with a 1\% selection ratio often outperforms using the full dataset and beats the baseline selection methods by 1.5 points in F1 score on average.

Denial Constraint (DC) is a well-established formalism that captures a wide range of integrity constraints commonly encountered, including candidate keys, functional dependencies, and ordering constraints, among others. Given their significance, there has been considerable research interest in achieving fast detection of DC violations, especially to support activities related to data exploration and preparation. Despite the significant advancements in the field, prior work exhibits notable limitations when confronted with large-scale datasets: the current state-of-the-art algorithm demonstrates a quadratic (worst-case) time and space complexity relative to the dataset's number of rows. In this paper, we establish a connection between orthogonal range search and DC violation detection. We then introduce Rapidash, a novel algorithm that demonstrates near-linear time and space complexity, representing a theoretical improvement over prior work. To validate the effectiveness of our algorithm, we conduct comprehensive evaluations on both open-source and real-world production datasets, with our production datasets notably being an order of magnitude larger than the datasets employed in prior studies. Our results reveal that Rapidash achieves up to 84× faster performance compared to state-of-the-art approaches while also exhibiting superior scalability.

Dynamic changes in network topology bring challenges to the management of mega low earth orbit (mega-LEO) systems. End-to-end network topology prediction is one of the key technologies to meet the challenges. At present, the graph theory-based prediction method can predict periodic changing links such as inter-satellite links (ISL) and satellite-ground links (GSL), but it cannot support the prediction of aperiodic user links. Moreover, when the scale of network nodes grows, the memory consumption and calculation time also increase rapidly, and not applicable in LEO mega-constellation networks with more than 10,000 nodes, such as Starlink satellite networks. To address these problems, we propose a prediction method based on graph convolutional neural network (GCN) and gated recursive unit (GRU). The key point of our method is to predict the end-to-end link changes of LEO mega-constellation, while reducing memory consumption and computing time. Simulation results show that the proposed method can achieve the topology prediction accuracy of more than 85% and reduce the memory consumption and computation time by more than 25% and 18.1%, respectively.

High throughput satellites (HTS), which can provide wide coverage, are recognized as a promising extension and supplement to achieve global coverage for the space-air-ground integrated network (SAGIN). Different to current literature on seamless coverage of HTS which considers that the shapes of multi-spot beam coverage regions are regular, this paper considers a practical HTS system, in which the multi-spot beams with irregular coverage regions are discussed, and the limitations of spot beam resource and coverage overlapping are taken into account. In order to achieve efficient seamless coverage of HTS, a grid division based multi-spot beam footprint planning (GD-MBFP) scheme is proposed with a given maximum coverage overlap ratio, aiming to minimize the number of spot beams. Meanwhile, for GD-MBFP scheme, a grid division based singlespot beam coverage calculation (GD-SBCC) method is proposed to obtain the irregular coverage regions of spot beams accurately. Numerical results validate the effectiveness of the proposed GD-MBFP scheme. Given a maximum coverage overlap ratio of ηmax=30%, the number of spot beams achieving seamless coverage with the proposed GD-MBFP scheme is only 37.9% of that achieving seamless coverage with cellular beam footprint planning (CBFP) scheme. Meanwhile, the average SINR of UTs with the proposed GD-MBFP scheme is 1.58 times that with CBFP scheme. The average throughput of spot beam with the proposed GD-MBFP scheme is improved by 30.4% than that with CBFP scheme.

We study the problem of robust mean estimation and introduce a novel Hamming distance-based measure of distribution shift for coordinate-level corruptions. We show that this measure yields adversary models that capture more realistic corruptions than those used in prior works, and present an information-theoretic analysis of robust mean estimation in these settings. We show that for structured distributions, methods that leverage the structure yield information theoretically more accurate mean estimation. We also focus on practical algorithms for robust mean estimation and study when data cleaning-inspired approaches that first fix corruptions in the input data and then perform robust mean estimation can match the information theoretic bounds of our analysis. We finally demonstrate experimentally that this two-step approach outperforms structure-agnostic robust estimation and provides accurate mean estimation even for high-magnitude corruption.

Finding a small object (e.g., earbuds, keys or a wallet) in an indoor environment (e.g., in a house or an office) can be frustrating. In this paper, we propose an innovative system, calledHyperEar, to localize such an object using only a single smartphone, based on enhanced time-difference-of-arrival (TDoA) measurements over acoustic signals issued from the object. One major challenge is the hardware limitations of a Commercial-Off-The-Shelf (COTS) phone with a short separation between the two microphones and the low sampling rate of such microphones. HyperEar enhances the accuracy of TDoA measurements by virtually increasing distances between microphones through sliding the phone in the air. HyperEar requires no communication for synchronization between the phone and the object and is a low-cost and easy-to-use system. We evaluate the performance of HyperEar via extensive experiments in various indoor conditions and the results demonstrate that, for an object of 7 m away, HyperEar can achieve a mean localization accuracy of about 15 cm when the object in normal indoor environments.

Finding a small object (e.g., keys or a wallet) in an indoor environment (e.g., in a house or an office) can be frustrating. In this paper, we propose an innovative system, called HyperEar, to localize such an object using only one single smartphone, based on enhanced time-difference-of-arrival (TDoA) measurements over acoustic signals issued from the object. One major challenge is the hardware limitations of a Commercial-Off-The-Shelf (COTS) phone with a short separation between the two microphones and the low sampling rate of such microphones. HyperEar enhances the accuracy of TDoA measurements by virtually increasing distances between microphones through sliding the phone in the air. HyperEar requires no communication for synchronization between the phone and the object and is a low-cost and easy-to-use system. We evaluate the performance of HyperEar via extensive experiments in various indoor conditions and the results demonstrate that, for an object of 7m away, HyperEar can achieve a mean localization accuracy of about 15cm when the object in normal indoor environments.

Smartphones with dual cameras are increasingly popular due to the need of supporting 3D vision. The depth information is critical for 3D vision. However, the two cameras on a smartphone are too close to accurately estimate the depth information especially for objects beyond two meters. In this paper, we propose an innovative system, called HyperSight, to estimate the depth information of objects using a dual camera smartphone. HyperSight realizes a virtual longbaseline stereo vision rig by having a user to move the phone in the air. The phone movement is continuously tracked and estimated using the short-baseline dual camera seeing nearby objects. We implement HyperSight as software on a Commercial-Off-The-Shelf (COTS) smartphone and conduct real-world experiments. The results show that when measuring feature-rich objects at a distance of five meters, HyperSight achieves a mean depth error of 6cm, which is up to 10× and 18× improvement in the accuracy compared with the stereo vision system using the native dual cameras and the Measure app based on ARKit 1 on mobile devices, respectively.

The eigenvalue equation intervenes in models of infectious disease propagation and could be used as an ally of vaccination campaigns in the actions carried out by health care organizations. The stochastic model based on Page Rank allows to simulate the epidemic spread, where a Like-like infection vector is calculated to help establish efficient vaccination strategy. In the context of epidemic spread, generally the damping factor is high. This is because the probability that an infected individual contaminates any other individual through some unusual contact is low. One consequence of this results is that the second largest eigenvalue of Page Rank matrix could be very close to its dominant eigenvalue. Another difficulty arises from the growing size of real networks. Handling very big graph becomes a challenge for computing Page Rank. Furthermore, the high damping factor makes many existing algorithms less efficient. In this paper, we explore the computation methods of Page Rank to address these issues. Specifically, we study the implicitly restarted Arnoldi method (IRAM) and discuss some possible improvements over it. We also present a parallel implementation for IRAM, targeting big data and sparse matrices representing scale-free networks (also known as power law networks). The algorithm is tested on a nation wide cluster of clusters Grid5000. Experiments on very large networks such as twitter, yahoo (over 1 billion nodes) are conducted.

In this paper, we will present an effective improved parallel hybrid asynchronous preconditioned GMRES method implemented on a nation wide grid environment. The classic restarted GMRES method is used widely to solve the large sparse linear systems. In order to accelerate the convergence, we use Arnoldi method to compute Ritz elements in parallel to optimize the computation of a polynomial. These two methods are asynchronously interconnected and we distributed them on a large nation wide cluster of clusters. From the numeric results for solving the real and complex system, we will present the performance of this hybrid method on such nation wide grids.