Federated learning (FL) is a distributed learning paradigm that allows multiple clients to collaboratively train a shared model without disclosing local data. However, temporal multimodal FL remains difficult to implement when clients have missing modalities and heterogeneous data distributions. Existing methods often fail to preserve modality-specific temporal dynamics under incomplete observations or do not address the training instability caused by degraded local representations. We propose FedGMM, a robust framework that stabilizes local representations before global aggregation by preserving intra-modal temporal structures and suppressing unreliable updates. Specifically, the framework employs a lightweight graph-formulated temporal fusion strategy to construct modality-specific temporal graphs, using masked adjacency matrices to prevent invalid feature propagation from missing observations. It also incorporates modality presence embeddings to explicitly encode availability patterns. To improve global collaboration, a momentum meta-margin aggregation strategy tracks the historical stability of client updates to reduce the influence of severely degraded updates caused by missing or corrupted multimodal observations. Extensive experiments on three public datasets show that FedGMM consistently improves over standard attention-based fusion. The model remains robust even at a 50% missing-modality rate, demonstrating its effectiveness in handling temporal incompleteness and client heterogeneity in federated settings.

WiFi-based human activity recognition (HAR) holds significant application potential across various fields. To handle dynamic environments where new activities are continuously introduced, WiFi-based HAR systems must adapt by learning new concepts without forgetting previously learned ones. Furthermore, retaining knowledge from old activities by storing historical exemplar is impractical for WiFi-based HAR due to privacy concerns and limited storage capacity of edge devices. In this work, we propose ConSense, a lightweight and fast-adapted exemplar-free class incremental learning framework for WiFi-based HAR. The framework leverages the transformer architecture and involves dynamic model expansion and selective retraining to preserve previously learned knowledge while integrating new information. Specifically, during incremental sessions, small-scale trainable parameters that are trained specifically on the data of each task are added in the multi-head self-attention layer. In addition, a selective retraining strategy that dynamically adjusts the weights in multilayer perceptron based on the performance stability of neurons across tasks is used. Rather than training the entire model, the proposed strategies of dynamic model expansion and selective retraining reduce the overall computational load while balancing stability on previous tasks and plasticity on new tasks. Evaluation results on three public WiFi datasets demonstrate that ConSense not only outperforms several competitive approaches but also requires fewer parameters, highlighting its practical utility in class-incremental scenarios for HAR.

Personalized federated learning is a decentralized approach that enables clients to collaboratively train a shared model while customizing it to their unique data and requirements. However, existing methods are not suited for dynamic scenarios where clients continuously learn from new classes of data, as they struggle with the increased data heterogeneity caused by this influx. To address this, we propose a Personalized Federated Class Incremental Learning (PFCIL) framework based on critical parameter transfer, called CPT-PFCIL, which each local client can continuously learn new classes, and before uploading the local model to the server, the client separates parameters of the local model into global federated parameters and task-specific parameters. Through communication with the server, each client selectively receives knowledge from other clients, achieving continuous learning and updating of the model while providing personalized model optimization. Our method has been validated on three datasets, demonstrating better performance compared to existing personalized federated learning methods.

Online Class-Incremental Continual Learning (CICL) methods often rely on data replay, where a fixed-size memory buffer stores a small subset of previous data to mitigate catastrophic forgetting. However, as incremental tasks progress, the fixed buffer size reduces the number of exemplars stored per class, weakening their representativeness. We argue that preserving classification centers is more effective than maintaining decision boundaries with limited exemplars. This highlights the dual challenges of selecting high-quality exemplars for storage and addressing the imbalance between old and new classes. To tackle these issues, we propose the Representative Exploration Replay (RER) framework. Our approach evaluates exemplar representativeness using a novel metric based on the model’s classification performance, ensuring that more representative exemplars are prioritized for storage. Additionally, it mitigates class imbalance through mutual information gradient masking and knowledge distillation. Comprehensive experiments on three datasets demonstrate that RER achieves an average performance improvement of 9% over 13 state-of-the-art methods.

Federated Class-Incremental Learning (FCIL) aims to design privacy-preserving collaborative training methods to continuously learn new classes from distributed datasets. In these scenarios, federated clients face the challenge of encountering new classes while being constrained by limited memory capacity, which can lead to catastrophic forgetting in the resulting global model. Existing FCIL approaches tend to overlook the challenges posed by the heterogeneity of dataset label distribution among clients, thereby constraining the generalization capacity of the global model they learn. Some of these methods also suffer from excessive computational burdens when addressing catastrophic forgetting problems. Furthermore, certain approaches are constrained to handling only straightforward data, posing significant difficulties in managing complex datasets and tackling more intricate scenarios. In this article, we propose the Weighted Aggregation and Distillation-based FCIL (WAD-FCIL) method to address these limitations. To address data heterogeneity arising from class imbalance, we first introduce a task-aware client clustering method to identify clients with extreme class deviations before global model aggregation to eliminate potential impact on the global model. Then, we propose a multisampling weighted aggregation approach during the global FL model update that integrates knowledge from different clients and dynamically adjusts the weight of each client model to facilitate model update. To mitigate catastrophic forgetting, we propose a multimodel distillation strategy that involves selecting multiple teacher models for knowledge distillation. Extensive experiments comparingWAD-FCILwith ten state-of-the-art methods demonstrate that it significantly outperforms the baselines by 0.8%–3.2% in terms of average test accuracy on three representative benchmark datasets. The code of this work is available athttps://github.com/wufeng10010/WAD-FCIL.

Federated learning (FL) is the prevailing paradigm in privacy-preserving machine learning. Despite recent advances yielding state-of-the-art outcomes, FL systems face challenges in adapting to dynamic real-world scenarios, where the local data distributions of clients may shift over time. This limitation stems from the common assumption of data stationarity in existing FL methods. In this paper, we propose the Federated Learning framework for CLass Incremental Pretrained models (FL-CLIP) to bridge this gap, by enabling continuous adaptation to changes in underlying FL client data distributions. The focus is on the incremental adaptation of existing FL models, trained on a substantial number of base classes, to newly arriving classes. To address the issue of catastrophic forgetting, the concepts of plasticity and stability are integrated into two distinct stages of FL-CLIP. Additionally, a task-weighted auxiliary loss is designed to tackle the class imbalance problem, and we propose a lightweight distillation method to enable FL-CLIP to adapt to resource-constrained environments. Extensive experiments demonstrate that FL-CLIP significantly outperforms 7 state-of-theart baselines in terms of average task accuracy, achieving a balanced performance across base and novel classes with a smaller memory footprint.

Federated class-incremental learning (FCIL) aims to allow federated learning (FL) systems to consistently learn new tasks with classes that change dynamically, without forgetting knowledge from previous classes. In FCIL scenarios, both heterogeneity in both label and data distribution across clients and catastrophic forgetting caused by continual emergence of new classes can significantly affect the performance of a FL system. Existing FCIL methods assume only changes in class distribution over time for each single client while ignoring class-specific domain distribution. Furthermore, these methods often rely on storing old class exemplars to mitigate catastrophic forgetting, potentially raising privacy concerns and computational burdens. In this article, we propose a FCIL framework called generative federated class-incremental learning (GenFCIL) that effectively addresses the aforementioned challenges. First, we introduce a lightweight generator that promotes knowledge sharing among clients and preserves the accumulated knowledge from all clients. By collecting classes and their associated data from each client, the generator effectively tackles data heterogeneity, facilitating information transfer across clients, and mitigating catastrophic forgetting in a replay-free manner. Importantly, the lightweight nature of the generator ensures that it does not impose excessive memory and computation requirements. Second, to tackle challenges from shifts in both class distribution and class-specific domain distribution in general FCIL scenarios, which may exacerbate catastrophic forgetting, we incorporate and update multiple logit scores from clients focusing on their old and new overlapping classes to incorporate more intraclass information. Experimental results show that GenFCIL effectively alleviates the impact of catastrophic forgetting and heterogeneity.

We study the problem of allocating many mobile robots for the execution of a pre-defined sweep schedule in a known two-dimensional environment, with applications toward search and rescue, coverage, surveillance, monitoring, pursuit-evasion, and so on. The mobile robots (or agents) are assumed to have one-dimensional sensing capability with probabilistic guarantees that deteriorate as the sensing distance increases. In solving such tasks, a time-parameterized distribution of robots along the sweep frontier must be computed, to minimize the number of robots used to achieve some desired coverage quality guarantee or to maximize the probabilistic guarantee for a given the number of robots. We propose a max-flow-based algorithm for solving the allocation task, which builds on a decomposition technique of the workspace as a generalization of the well-known boustrophedon decomposition. Our proposed algorithm has a very low polynomial running time and completes in under two seconds for polygonal environments with over 105vertices. Simulation experiments are carried out on three realistic use cases with randomly generated obstacles of varying shapes, sizes, and spatial distributions, demonstrating our proposed method's applicability and scalability. Introduction video: https://youtu.be/8taX92rzC5k.

Object rearrangement is a fundamental sub-task in accomplishing a great many physical tasks. As such, effectively executing rearrangement is an important skill for intelligent robots to master. In this study, we conduct the first algorithmic study on optimally solving the problem of Multi-layer Object Rearrangement on a Tabletop (MORT), in which one object may be relocated at a time, and an object can only be moved if other objects do not block its top surface. In addition, any intermediate structure during the reconfiguration process must be physically stable, i.e., it should stand without external support. To tackle the dual challenges of untangling the dependencies between objects and ensuring structural stability, we develop an algorithm that interleaves the computation of the optimal rearrangement plan and structural stability checking. Using a carefully constructed integer linear programming (ILP) model, our algorithm, Stability-Aware Rearrangement Programming (SARP), readily scales to optimally solve complex rearrangement problems of 3D structures with over 60 building blocks, with solution quality significantly outperforming natural greedy best-first approaches. Upon the publication of the manuscript source code and data will be available at https//github.com/arc-1/mort/.

In this work, we carry out structural and al-gorithmic studies of a problem of barrier forming: selecting the minimum number of straight line segments (barriers) that separate several sets of mutually disjoint objects in the plane. The problem models the optimal placement of line sensors (e.g., infrared laser beams) for isolating many types of regions in a pair- wise manner for practical purposes (e.g., guarding against intrusions). The problem is NP-hard even if we want to find the minimum number of lines to separate two sets of points in the plane. Under the umbrella problem of barrier forming with minimum number of line segments, three settings are examined: barrier forming for point sets, point sets with polygonal obstacles, polygonal sets with polygonal obstacles. We describe methods for computing the optimal solution for the first two settings with the assistance of mathematical programming, and provide a 2-OPT solution for the third. We demonstrate the effectiveness of our methods through extensive simulations.

For enabling efficient, large-scale coordination of unmanned aerial vehicles (UAV s) under the labeled setting, in this work, we develop the first polynomial time algorithm for the reconfiguration of many moving bodies in three-dimensional spaces, with provable 1.$x$asymptotic makespan optimality guarantee under high robot density. More precisely, on an$m_{1} \times m_{2} \times m_{3}$grid,$m_{1}\geq m_{2}\geq m_{3}$, our method computes solutions for routing up to$\displaystyle \frac{m_{1}m_{2}m_{3}}{3}$uniquely labeled robots with uniformly randomly distributed start and goal configurations within a makespan of$m_{1}+2m_{2}+2m_{3}+o(m_{1})$, with high probability. Because the makespan lower bound for such instances is$m_{1}+m_{2}+m_{3}-o(m_{1})$, also with high probability, as$m_{1}\displaystyle \rightarrow\infty, \frac{m_{1}+2m_{2}+2m_{3}}{m_{1}+m_{2}+m_{3}}$optimality guarantee is achieved.$\displaystyle \frac{m_{1}+2 m_{2}+2m_{3}}{m_{1}+m_{2}+m_{3}}\in\left(1, \displaystyle \frac{5}{3}\right]$, yielding 1.$x$optimality. In contrast, it is well-known that multi-robot path planning is NP-hard to optimally solve. In numerical evaluations, our method readily scales to support the motion planning of over 100, 000 robots in 3D while simultaneously achieving 1.$x$optimality. We demonstrate the application of our method in coordinating many quadcopters in both simulation and hardware experiments.

Computational understanding of humor is an important topic under creative language understanding and modeling. It can play a key role in complex human-AI interactions. The challenge here is that human perception of humorous content is highly subjective. The same joke may receive different funniness ratings from different readers. This makes it highly challenging for humor recognition models to achieve personalization in practical scenarios. Existing approaches are generally designed based on the assumption that users have a consensus on whether a given text is humorous or not. Thus, they cannot handle diverse humor preferences well. In this article, we propose the FedHumor approach for the recognition of humorous content in a personalized manner through Federated Learning (FL). Extending a pre-trained language model, FedHumor guides the fine-tuning process by considering diverse distributions of humor preferences from individuals. It incorporates a diversity adaptation strategy into the FL paradigm to train a personalized humor recognition model. To the best of our knowledge, FedHumor is the first text-based personalized humor recognition model through federated learning. Extensive experiments demonstrate the advantage of FedHumor in recognizing humorous texts compared to nine state-of-the-art humor recognition approaches with superior capability for handling the diversity in humor labels produced by users with diverse preferences.

We carry out a structural and algorithmic study of a mobile sensor coverage optimization problem targeting 2D surfaces embedded in a 3D workspace. The investigated settings model multiple important applications including camera net- work deployment for surveillance, geological monitoring/survey of 3D terrains, and UVC-based surface disinfection for the prevention of the spread of disease agents (e.g., SARS-CoV-2). Under a unified general "sensor coverage" problem, three concrete formulations are examined, focusing on optimizing visibility, single-best coverage quality, and cumulative quality, respectively. After demonstrating the computational intractability of all these formulations, we describe approximation schemes and mathematical programming models for near-optimally solving them. The effectiveness of our methods is thoroughly evaluated under realistic and practical scenarios.

High spectral resolution brings hyperspectral images with large amounts of information, which makes these images more useful in many applications than images obtained from traditional multispectral scanners with low spectral resolution. However, the high data dimensionality of hyperspectral images increases the burden on data computation, storage, and transmission; fortunately, the high redundancy in the spectral domain allows for significant dimensionality reduction. Band selection provides a simple dimensionality reduction scheme by discarding bands that are highly redundant, thereby preserving the structure of the data set. This paper proposes a new criterion for pointwise-ranking-based band selection that uses a nonhomogeneous hidden Markov chain (NHMC) model for redundant wavelet coefficients of each hyperspectral signature. The model provides a binary multiscale label that encodes semantic features that are useful to discriminate spectral types. A band ranking score considers the average correlation among the average NHMC labels for each band. We also test richer discrete-valued label vectors that provide a more finely grained quantization of spectral fluctuations. In addition, since band selection methods based on band ranking often ignore correlations in selected bands, we study the effect of redundancy elimination, applied on the selected features, on the performance of an example classification problem. Our experimental results also include an optional redundancy elimination step and test their effect on classification performance that is based on the selected bands. The experimental results also include a comparison with several relevant supervised band selection techniques.

This paper proposes a new hyperspectral unmixing method for nonlinearly mixed hyperspectral data using a semantic representation in a semisupervised fashion, assuming the availability of a spectral reference library. Existing semisupervised unmixing algorithms select members from an endmember library that are present at each of the pixels; most such methods assume a linear mixing model. However, those methods will fail in the presence of nonlinear mixing among the observed spectra. To address this issue, we develop an endmember selection method using a recently proposed semantic spectral representation obtained via nonhomogeneous hidden Markov chain model for a wavelet transform of the spectra. The semantic representation can encode spectrally discriminative features for any observed spectrum, and therefore, our proposed method can perform endmember selection without any assumption on the mixing model. The experimental results show that in the presence of sufficiently nonlinear mixing, our proposed method outperforms dictionary-based sparse unmixing approaches based on linear models.

We consider the application of non-homogeneous hidden Markov chain (NHMC) models to the problem of hyperspectral signature classification. It has been previously shown that the NHMC model enables the detection of several semantic structural features of hyperspectral signatures. However, there are some aspects of the spectral data that are not fully captured by the proposed NHMC models such as the relatively smooth but fluctuating regions and the fluctuation orientations. In order to address these limitations, we propose an improved NHMC model based on Daubechies-1 wavelets in conjunction with an increased the model complexity. Experimental results show that the revised approach outperforms existing approaches relevant in classification tasks.