Federated domain generalization (FDG) for time-series classification (TSC) poses a critical challenge for modern intelligent web services, which rely on edge-collected time-series signals from diverse mobile applications and web devices (e.g., wearables sensors) to support decision-making. The source heterogeneity and temporal dynamics give rise to out-of-distribution (OOD) patterns, which hinder the model's ability to generalize to previously unseen users and devices. In this work, we propose Federated Generalization via Diversity Generation (FedDiG), a diffusion-based FDG framework that captures intra-client distribution shifts from a frequency-domain perspective and employs cross-frequency sampling to synthesize time-series data with diverse spectral patterns. Specifically, FedDiG first performs frequency-proxy representation learning on clients to serve as diffusion conditions. The server then aggregates client-side frequency proxies to construct a global proxy pool and applies class-wise mixup to create novel frequency features. These features guide a global diffusion model to produce diverse data, enabling the simulation of previously unseen patterns and thereby enhancing model training. Extensive experiments on four cross-domain time-series benchmarks demonstrate that FedDiG significantly outperforms state-of-the-art federated learning and FDG baselines, particularly under small-data regimes and large-scale client scenarios, achieving robust generalization to unseen domains in federated settings. This work bridges distribution-diversity synthesis and FDG for time-series to support robust, scalable web applications fed by edge-collected signals, delivering web-scale generalization across heterogeneous web, mobile, and IoT clients.

As neural network models grow larger and more complex, federated learning (FL) faces challenges in terms of communication and computation efficiency. To address these issues, layer-wise learning has been proposed. Existing approaches did not leverage useful properties of layer-wise learning including update locking and variations in convergence rates, thereby resulting in sub-par model performance. To bridge this gap, we propose theFederatedMeanBlockDifference-based global model aggregation approach withPatience-based local training (FedMBDP). We automatically partition the neural network model into uncoupled blocks and progressively train them. Determining which blocks to train and aggregate becomes a critical task. To improve computation efficiency, we propose a patience-based local training algorithm to adaptively select training blocks, reducing computation latency. To improve communication efficiency, we introduce a mean block difference-based global model aggregation algorithm to dynamically select blocks for aggregation to minimize communication latency. We provide the convergence analysis of FedMBDP. Extensive experiments on three widely adopted benchmark datasets show that FedMBDP achieves superior performance compared to six state-of-the-art approaches. It reduces FL training latency by 26.37% compared to the best baseline, while achieving similar test accuracy.

With growing client diversity, model-heterogeneous personalized federated learning (MHPFL) supports collaboration over structure-heterogeneous client models. However, existing MHPFL methods only achieve client-level personalization but ignore inherent discrepancies within each client's different data samples, leading to limited model performance. To this end, we propose a novel model-heterogeneouspersonalizedFederated learning withMixtureofExperts (pFedMoE) to achieve a fine-grained data-level personalization. As the first work that incorporates MoE in MHPFL, it introduces three innovations: (1) Different clients hold heterogeneous local models, we add a small proxy global homogeneous feature extractor shared by clients for knowledge exchange. (2) To achieve a fine-grained data-level personalization, we construct a personalized local MoE for each client: a local expert (local heterogeneous client model's feature extractor), a global expert (global proxy homogeneous feature extractor), and a local personalized gating network, which dynamically balances the generalization and personalization of the local model at the data sample level. (3) We customize a lightweight linear gating network to capture the generalized and personalized data characteristics of each local data sample. We theoretically prove its$\mathcal {O}(1/T)$convergence rate. Experiments on 3 benchmark image datasets, 1 real-world image dataset and 1 real-world text dataset against 9 baselines demonstrate its state-of-the-art model accuracy with up to 2.79% accuracy improvement while saving up to 43.12% computational overheads and keeping satisfactory communication costs.

Federated learning (FL), an emerging distributed machine learning paradigm, utilizes edge decentralized data from multiple edge nodes (clients) to train a shared model under preserved data privacy. Furthermore, model-heterogeneous personalized federated learning (MHPFL) enables FL clients to train structurally different personalized models on non-independent and identically distributed (non-lID) local data. Existing MHPFL methods focus on data distribution differences among clients, and they propose various client-level personalization approaches to alleviate non-lID issues. However, different data samples in one client may also have different features, which are often ignored, resulting in constrained model performances. To bridge this gap, we propose a novel model-heterogeneous personalized Federated learning approach with Adaptive Feature Mixture (pFedAFM) to achieve data-level personalization while maintaining efficient communication and computation. It consists of three innovative designs: 1) We add a homogeneous small feature extractor alongside each client's local heterogeneous model, and the server aggregates these homogeneous small feature extractors for cross-client knowledge fusion. 2) We design an iterative training strategy to alternately train the global homogeneous small feature extractor and the local heterogeneous client model, for effective bidirectional exchange between global generalized knowledge and local personalized knowledge. 3) During model training, we devise a trainable weight vector to adaptively mix the features (representation) extracted by the global homogeneous and local heterogeneous models for different data samples, i.e., fulfilling data-level personalized feature mixture. Theoretical analysis proves that pFedAFM converges over time. Extensive experiments on 3 computer vision (CV) and 1 natural lan-guage processing (NLP) benchmark datasets demonstrate that pFedAFM significantly outperforms 8 state-of-the-art MHPFL methods, achieving up to 7.93% accuracy improvement while incurring low communication and computation costs.

The widespread adoption of location-based services has led to the generation of vast amounts of mobility data, providing significant opportunities to model user movement dynamics within urban environments. Recent advancements have focused on adapting Large Language Models (LLMs) for mobility analytics. However, existing methods face two primary limitations: inadequate semantic representation of locations (i.e., discrete IDs) and insufficient modeling of mobility signals within LLMs (i.e., single templated instruction fine-tuning). To address these issues, we propose QT-Mob, a novel framework that significantly enhances LLMs for mobility analytics. QT-Mob introduces a location tokenization module that learns compact, semantically rich tokens to represent locations, preserving contextual information while ensuring compatibility with LLMs. Furthermore, QT-Mob incorporates a series of complementary fine-tuning objectives that align the learned tokens with the internal representations in LLMs, improving the model's comprehension of sequential movement patterns and location semantics. The proposed QT-Mob framework not only enhances LLMs' ability to interpret mobility data but also provides a more generalizable approach for various mobility analytics tasks. Experiments on three real-world dataset demonstrate the superior performance in both next-location prediction and mobility recovery tasks, outperforming existing deep learning and LLM-based methods.

Single-domain generalization (SDG) in time series classification (TSC) poses significant challenges for current time-series domain generalization methods due to the extremely limited data available from only one source domain. In this study, we propose Segment-dErived Expansion of Domains (SEED), a diffusion-based method that effectively expands domain diversity for SDG. We reveal that individual instances exhibit intrinsic temporal shifts over time, which provides a principled foundation for creating multiple pseudo domains by segmenting each instance into distinct parts. To do so, SEED extracts two complementary representations from each time-series segment: 1) a segment-specific representation that captures diverse distributional variations, and 2) a segment-invariant representation that preserves class semantics. SEED formulates these representations as pseudo-domain prompts to guide a diffusion model in generating diverse yet semantically consistent time-series data. Additionally, SEED introduces a novel prompt-fused sampling for diffusion, enabling flexible recombination of segment-specific features to continuously expand the pseudo-domain space. We provide both theoretical analysis and extensive empirical evaluations on four widely used TSC benchmarks to validate its ability in reducing generalization error and improving model's performances in SDG. In our experiments, SEED significantly improves classification accuracy by 7.68% on average compared to the strong baselines.

Federated Foundation Models (FedFMs) represent a distributed learning paradigm that fuses general competences of foundation models as well as privacy-preserving capabilities of federated learning. This combination allows the large foundation models and the small local domain models at the remote clients to learn from each other in a teacher-student learning setting. This paper provides a comprehensive summary of the ten challenging problems inherent in FedFMs, encompassing foundational theory, utilization of private data, continual learning, unlearning, Non-IID and graph data, bidirectional knowledge transfer, incentive mechanism design, game mechanism design, model watermarking, and efficiency. The ten challenging problems manifest in five pivotal aspects: “Foundational Theory,” which aims to establish a coherent and unifying theoretical framework for FedFMs. “Data,” addressing the difficulties in leveraging domain-specific knowledge from private data while maintaining privacy; “Heterogeneity,” examining variations in data, model, and computational resources across clients; “Security and Privacy,” focusing on defenses against malicious attacks and model theft; and “Efficiency,” highlighting the need for improvements in training, communication, and parameter efficiency. For each problem, we offer a clear mathematical definition on the objective function, analyze existing methods, and discuss the key challenges and potential solutions. This in-depth exploration aims to advance the theoretical foundations of FedFMs, guide practical implementations, and inspire future research to overcome these obstacles, thereby enabling the robust, efficient, and privacy-preserving FedFMs in various real-world applications.

Cross-domain recommendation (CDR) aims to alleviate the data sparsity problem by leveraging the benefits of modeling two domains. However, existing research often focuses on the recommendation performance while ignores the privacy leakage issue. We find that an attacker can infer user attribute information from the knowledge (e.g., user preferences) transferred between the source and target domains. For example, in our experiments, the average inference accuracies of attack models on gender and age attributes are 0.8323 and 0.3897. The best-performing attack model achieves accuracies of 0.8847 and 0.4634, exceeding a random inference by 25.10% and 64.04%. We can see that the leakage of user attribute information may significantly exceed what would be expected from random inference. In this paper, we propose a novel recommendation framework named CVGAE (short for camouflaged variational graph autoencoder), which effectively models user behaviors and mitigates the risk of user attribute information leakage at the same time. Specifically, our CVGAE combines the strengths of VAEs in capturing latent features and variability with the ability of GCNs in exploiting high-order relational information. Moreover, to ensure against attribute inference attacks without sacrificing the recommendation performance, we design a user attribute protection module that fuses user attribute-camouflaged information with knowledge transfer during cross-domain processes. We then conduct extensive experiments on three real-world datasets, and find our CVGAE is able to achieve strong privacy protection while making little sacrifices in recommendation accuracy.

In recent years, recommender systems have advanced rapidly, where embedding learning for users and items plays a critical role. A standard method learns a unique embedding vector for each user and item. However, such a method has two important limitations in real-world applications: (1) it is hard to learn embeddings that generalize well for users and items that have rare interactions, and (2) it may incur unbearably high memory costs when the number of users and items scales up. Existing approaches either can only address one of the limitations or have flawed overall performances. In this article, we propose Clustered Embedding Learning (CEL) as an integrated solution to these two problems. CEL is a plug-and-play embedding learning framework that can be combined with any differentiable feature interaction model. It is capable of achieving improved performance, especially for cold users and items, with reduced memory cost. CEL enables automatic and dynamic clustering of users and items in a top-down fashion, where clustered entities could jointly learn a shared embedding. The accelerated version of CEL has an optimal time complexity, which supports efficient online updates. Theoretically, we prove the identifiability and the existence of a unique optimal number of clusters for CEL in the context of nonnegative matrix factorization. Empirically, we validate the effectiveness of CEL on three public datasets and one business dataset, showing its consistently superior performance against state-of-the-art methods. In particular, when incorporating CEL into the business model, it brings an improvement of \(+0.6\%\) in AUC, which translates into a significant revenue gain; meanwhile, the size of the embedding table gets 2,650 times smaller. Additionally, we demonstrate that if there is enough memory, learning a personalized embedding for each user and item around their clustering centers is feasible and can further boost performance. In this article, we enhance and extend the personalization technique we initially proposed in our earlier work [ 4 ], which introduced an offset regularization to prevent personalized embeddings from drifting too far away from the central (cluster) embedding, thereby mitigating overfitting. However, in [ 4 ], we simply applied a uniform regularization weight across all embeddings, which, given the considerable variation in the number of their associated interactions, is suboptimal. To address this, we investigate in this article the strategies for non-uniform offset regularization that adjusts regularization weights according to the number of associated interactions, which leads to significant improvements compared with uniform offset regularization. Furthermore, we extend CEL into Meta-CEL, factoring in future personalization during cluster optimization, which leads to additional enhancements in personalization performance.

Since the development of artificial intelligence technology, E-Commerce has gradually become one of the world's largest commercial markets. Within this domain, sales events, which are based on sociological mechanisms, play a significant role. E-Commerce platforms frequently offer sales and promotions to encourage users to purchase items, leading to significant changes in live environments. Learning-To-Rank (LTR) is a crucial component of E-Commerce search and recommendations, and substantial efforts have been devoted to this area. However, existing methods often assume an independent and identically distributed data setting, which does not account for the evolving distribution of online systems beyond online finetuning strategies. This limitation can lead to inaccurate predictions of user behaviors during sales events, resulting in significant loss of revenue. In addition, models must readjust themselves once sales have concluded in order to eliminate any effects caused by the sales events, leading to further regret. To address these limitations, we introduce a long-term E-Commerce search data set specifically designed to incubate LTR algorithms during such sales events, with the objective of advancing the capabilities of E-Commerce search engines. Our investigation focuses on typical industry practices and aims to identify potential solutions to address these challenges.

Graph convolutional networks (GCNs) are promising for graph learning tasks. For privacy-preserving graph learning tasks involving distributed graph datasets, federated learning (FL)-based GCN (FedGCN) training is required. An important open challenge for FedGCN is scaling to large graphs, which typically incurs 1) high computation overhead for handling the explosively-increasing number of neighbors, and 2) high communication overhead of training GCNs involving multiple FL clients. Thus, neighbor sampling is being studied to enhance the scalability of FedGCNs. Existing FedGCN training techniques with neighbor sampling often produce extremely large communication and computation overhead and inaccurate node embeddings, leading to poor model performance. To bridge this gap, we propose the Federated Adaptive Attention-based Sampling (FedAAS) approach. It achieves substantial cost savings by efficiently leveraging historical embedding estimators and focusing the limited communication resources on transmitting the most influential neighbor node embeddings across FL clients. We further design an adaptive embedding synchronization scheme to optimize the efficiency and accuracy of FedAAS on large-scale datasets. Theoretical analysis shows that the approximation error induced by the staleness of historical embedding is upper bounded, and the model is guaranteed to converge in an efficient manner. Extensive experimental evaluation against four state-of-the-art baselines on six real-world graph datasets show that FedAAS achieves up to 5.12% higher test accuracy, while saving communication and computation costs by 95.11% and 94.76%, respectively.

As auxiliary collaborative signals, the entity connectivity and relation semanticity beneath knowledge graph (KG) triples can alleviate the data sparsity and cold-start issues of recommendation tasks. Thus many works consider obtaining user and item representations via information aggregation on graph-structured data within Euclidean space. However, the scale-free graphs (e.g., KGs) inherently exhibit non-Euclidean geometric topologies, such as tree-like and circle-like structures. The existing recommendation models built in a single type of embedding space do not have enough capacity to embrace various geometric patterns, consequently, resulting in suboptimal performance. To address this limitation, we propose a KG-aware recommendation model with mixed-curvature manifolds interaction learning, namely CurvRec. On the one hand, it aims to preserve various global geometric structures in KG with mixed-curvature manifold spaces as the backbone. On the other hand, we integrate Ricci curvature into graph convolutional networks (GCNs) to capture local geometric structural properties when aggregating neighbor nodes. Besides, to exploit the expressive spatial features in KG, we incorporate interaction learning to ensure the geometric message passing between curved manifolds. Specifically, we adopt curvature-aware geodesic distance metrics to maximize the mutual information between Euclidean space and non-Euclidean spaces. Through extensive experiments, we demonstrate that the proposed CurvRec outperforms state-of-the-art baselines.

In recent years, recommender systems have advanced rapidly, where embedding learning for users and items plays a critical role. A standard method learns a unique embedding vector for each user and item. However, such a method has two important limitations in real-world applications: 1) it is hard to learn embeddings that generalize well for users and items with rare interactions; and 2) it may incur unbearably high memory costs when the number of users and items scales up. Existing approaches either can only address one of the limitations or have flawed overall performances. In this paper, we propose Clustered Embedding Learning (CEL) as an integrated solution to these two problems. CEL is a plug-and-play embedding learning framework that can be combined with any differentiable feature interaction model. It is capable of achieving improved performance, especially for cold users and items, with reduced memory cost. CEL enables automatic and dynamic clustering of users and items in a top-down fashion, where clustered entities jointly learn a shared embedding. The accelerated version of CEL has an optimal time complexity, which supports efficient online updates. Theoretically, we prove the identifiability and the existence of a unique optimal number of clusters for CEL in the context of nonnegative matrix factorization. Empirically, we validate the effectiveness of CEL on three public datasets and one business dataset, showing its consistently superior performance against current state-of-the-art methods. In particular, when incorporating CEL into the business model, it brings an improvement of in AUC, which translates into a significant revenue gain; meanwhile, the size of the embedding table gets 2650 times smaller.1

Federated Learning (FL) enables a large number of data owners (a.k.a. FL clients) to jointly train a machine learning model without disclosing private local data. The importance of local data samples to the FL model vary widely. This is exacerbated by the presence of noisy data, which exhibit large losses similar to important (hard) samples. Currently, there lacks an FL approach that can effectively distinguish hard samples (which are beneficial) from noisy samples (which are harmful). To bridge this gap, we propose the Federated Client and Sample Selection (FedCSS) approach. It is a bilevel optimization approach for FL client-and-sample selection to achieve hard sample-aware noise-robust learning in a privacy preserving manner. It performs meta-learning based online approximation to iteratively update global FL models, select the most positively influential samples and deal with training data noise. Theoretical analysis shows that it is guaranteed to converge in an efficient manner. Experimental comparison against six state-of-the-art baselines on five real-world datasets in the presence of data noise and heterogeneity shows that it achieves up to 26.4% higher test accuracy, while saving communication and computation costs by at least 41.5% and 1.2%, respectively.

Most existing LTR approaches follow a supervised learning paradigm from offline data collected from the online system. However, it has been noticed that previous LTR models can have good performances over offline validation data but have poor online performances, which implies a possible large inconsistency between the offline and online evaluation. We investigate and confirm in this paper that such inconsistency exists and can have a significant impact on AliExpress Search. Reasons for the inconsistency include the ignorance of item context. Therefore, this paper proposes an evaluator-generator framework for LTR with item context. The framework consists of an evaluator that generalizes to evaluate recommendations involving the context, and a generator that maximizes the evaluator score by reinforcement learning, and a discriminator that ensures the generalization of the evaluator. Extensive experiments in simulation environments and AliExpress Search online system show that, firstly, the classic data-based metrics on the offline dataset can show significant inconsistency with online performance. Secondly, the proposed evaluator score is significantly more consistent with the online performance than common ranking metrics. Finally, as the consequence, our method achieves a significant improvement in terms of Conversion Rate over the industrial-level fine-tuned model in online A/B tests.

As a revolutionary auction mechanism for display advertising, real-time bidding (RTB) allows advertisers to purchase individual ad impressions through real-time auctions. In RTB, the demand-side platform (DSP) acts as advertisers' bidding agent and aims at developing appropriate bidding strategies to maximize their specific key performance indicators (KPIs). Existing bidding strategies perform well for optimizing profits when the ad budget severely limited. However, when there is sufficient budget, their performance deteriorates. This results in added complexity for advertisers when applying these approaches in practice, hindering wider adoption. To address this challenging limitation, we propose the Adaptive ROI-Aware Bidding (ARAB) approach. It intelligently analyzes the budget setting and auction market conditions, and adjusts the bidding function accordingly to optimize profits. Different from previous studies that only bid based on the ad revenue, our proposed ROI-aware bidding function also takes into account the ad cost at impression-level. By doing so, ARAB dynamically allocates the budget on more cost-effective impressions to increase profits. Through extensive offline experiments on two real-world public datasets, we demonstrate that the proposed ARAB has achieved significant improvements in terms of both profit and ROI compared to state-of-the-art approaches.

Unsupervised anomaly detection aims to build models to effectively detect unseen anomalies by only training on the normal data. Although previous reconstruction-based methods have made fruitful progress, their generalization ability is limited due to two critical challenges. First, the training dataset only contains normal patterns, which limits the model generalization ability. Second, the feature representations learned by existing models often lack representativeness which hampers the ability to preserve the diversity of normal patterns. In this paper, we propose a novel approach called Adaptive Memory Network with Self-supervised Learning (AMSL) to address these challenges and enhance the generalization ability in unsupervised anomaly detection. Based on the convolutional autoencoder structure, AMSL incorporates a self-supervised learning module to learn general normal patterns and an adaptive memory fusion module to learn rich feature representations. Experiments on four public multivariate time series datasets demonstrate that AMSL significantly improves the performance compared to other state-of-the-art methods. Specifically, on the largest CAP sleep stage detection dataset with 900 million samples, AMSL outperforms the second-best baseline by4%+ in both accuracy and F1 score. Apart from the enhanced generalization ability, AMSL is also more robust against input noise.

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.

Federated Learning (FL) bridges the gap between collaborative machine learning and preserving data privacy. To sustain the long-term operation of an FL ecosystem, it is important to attract high-quality data owners with appropriate incentive schemes. As an important building block of such incentive schemes, it is essential to fairly evaluate participants’ contribution to the performance of the final FL model without exposing their private data. Shapley Value (SV)–based techniques have been widely adopted to provide a fair evaluation of FL participant contributions. However, existing approaches incur significant computation costs, making them difficult to apply in practice. In this article, we propose the Guided Truncation Gradient Shapley (GTG-Shapley) approach to address this challenge. It reconstructs FL models from gradient updates for SV calculation instead of repeatedly training with different combinations of FL participants. In addition, we design a guided Monte Carlo sampling approach combined with within-round and between-round truncation to further reduce the number of model reconstructions and evaluations required. This is accomplished through extensive experiments under diverse realistic data distribution settings. The results demonstrate that GTG-Shapley can closely approximate actual Shapley values while significantly increasing computational efficiency compared with the state-of-the-art, especially under non-i.i.d. settings.

The rapid rise of real-time bidding-based online advertising has brought significant economic benefits and attracted extensive research attention. From the perspective of an advertiser, it is crucial to perform accurate utility estimation and cost estimation for each individual auction in order to achieve cost-effective advertising. These problems are known as the click through rate (CTR) prediction task and the market price modeling task, respectively. However, existing approaches treat CTR prediction and market price modeling as two independent tasks to be optimized without regard to each other, thus resulting in suboptimal performance. Moreover, they do not make full use of unlabeled data from the losing bids during estimations, which makes them suffer from the sample selection bias issue. To address these limitations, we propose Multi-task Advertising Estimator (MTAE), an end-to-end joint optimization framework which performs both CTR prediction and market price modeling simultaneously. Through multi-task learning, both estimation tasks can take advantage of knowledge transfer to achieve improved feature representation and generalization abilities. In addition, we leverage the abundant bid price signals in the full-volume bid request data and introduce an auxiliary task of predicting the winning probability into the framework for unbiased learning. Through extensive experiments on two large-scale real-world public datasets, we demonstrate that our proposed approach has achieved significant improvements over the state-of-the-art models under various performance metrics.

The huge number of machine learning (ML) methods has resulted in significant information overload. Faced with an overwhelming number of ML methods, it is challenging to select appropriate ones for the given dataset and task. In general, the names of ML methods or datasets are rather condensed, thus lacking specific explanations, while the rich latent relationships between ML entities are not fully explored. In this paper, we propose a description-enhanced machine learning knowledge graph-based approach - DEKR - to help recommend appropriate ML methods for given ML datasets. The proposed knowledge graph (KG) not only includes the connections between entities but also contains the descriptions of the dataset and method entities. DEKR fuses the structural information with the description information of entities in the knowledge graph. It is a deep hybrid recommendation framework, which incorporates the knowledge graph-based and text-based methods, overcoming the limitations of previous knowledge graph-based recommendation systems that ignore the description information. There are two key components of DEKR: 1) a graph neural network aggregating information from multi-order neighbors with attention to enrich the seed (i.e. dataset or method) node's own representation, and 2) a deep collaborative filtering network based on the description text to obtain the linear and nonlinear interactions of description features. Through extensive experiments, we demonstrated the efficiency of DEKR, which outperforms the current state-of-the-art baselines by a large margin.

Transfer learning aims to learn robust classifiers for the target domain by leveraging knowledge from a source domain. Since the source and the target domains are usually from different distributions, existing methods mainly focus on adapting the cross-domain marginal or conditional distributions. However, in real applications, the marginal and conditional distributions usually have different contributions to the domain discrepancy. Existing methods fail to quantitatively evaluate the different importance of these two distributions, which will result in unsatisfactory transfer performance. In this article, we propose a novel concept called Dynamic Distribution Adaptation (DDA), which is capable of quantitatively evaluating the relative importance of each distribution. DDA can be easily incorporated into the framework of structural risk minimization to solve transfer learning problems. On the basis of DDA, we propose two novel learning algorithms: (1) Manifold Dynamic Distribution Adaptation (MDDA) for traditional transfer learning, and (2) Dynamic Distribution Adaptation Network (DDAN) for deep transfer learning. Extensive experiments demonstrate that MDDA and DDAN significantly improve the transfer learning performance and set up a strong baseline over the latest deep and adversarial methods on digits recognition, sentiment analysis, and image classification. More importantly, it is shown that marginal and conditional distributions have different contributions to the domain divergence, and our DDA is able to provide good quantitative evaluation of their relative importance, which leads to better performance. We believe this observation can be helpful for future research in transfer learning.

Federated learning (FL) allows multiple parties to collaboratively train a machine learning model without sharing raw data. However, existing approaches are mainly designed for homogeneous feature spaces and fail to tackle covariate shift and feature heterogeneity without privacy leakage. In this paper, we propose a transfer learning approach to tackle the covariate shift of the overlapped homogeneous feature spaces, and bridge different data owners' heterogeneous feature spaces with stringent privacy preservation in FL. We propose an end-to-end privacy-preserving multi-party learning approach with two variants based on homomorphic encryption and secret sharing techniques, respectively, to build a heterogeneous federated transfer learning (HFTL) framework. Finally, we not only demonstrate experimentally that the HFTL is secure, effective and highly scalable on five benchmark datasets, but also apply it into a real application of in-hospital mortality prediction from MIMIC-III dataset, where privacy is of significant concern.

Video contents have become a critical tool for promoting products in E-commerce. However, the lack of automatic promotional video generation solutions makes large-scale video-based promotion campaigns infeasible. The first step of automatically producing promotional videos is to generate visual storylines, which is to select the building block footage and place them in an appropriate order. This task is related to the subjective viewing experience. It is hitherto performed by human experts and thus, hard to scale. To address this problem, we propose WundtBackpack, an algorithmic approach to generate storylines based on available visual materials, which can be video clips or images. It consists of two main parts, 1) the Learnable Wundt Curve to evaluate the perceived persuasiveness based on the stimulus intensity of a sequence of visual materials, which only requires a small volume of data to train; and 2) a clustering-based backpacking algorithm to generate persuasive sequences of visual materials while considering video length constraints. In this way, the proposed approach provides a dynamic structure to empower artificial intelligence (AI) to organize video footage in order to construct a sequence of visual stimuli with persuasive power. Extensive real-world experiments show that our approach achieves close to 10% higher perceived persuasiveness scores by human testers, and 12.5% higher expected revenue compared to the best performing state-of-the-art approach.

With the rapid development of artificial intelligence (AI), ethical issues surrounding AI have attracted increasing attention. In particular, autonomous vehicles may face moral dilemmas in accident scenarios, such as staying the course resulting in hurting pedestrians or swerving leading to hurting passengers. To investigate such ethical dilemmas, recent studies have adopted preference aggregation, in which each voter expresses her/his preferences over decisions for the possible ethical dilemma scenarios, and a centralized system aggregates these preferences to obtain the winning decision. Although a useful methodology for building ethical AI systems, such an approach can potentially violate the privacy of voters since moral preferences are sensitive information and their disclosure can be exploited by malicious parties resulting in negative consequences. In this paper, we report a first-of-its-kind privacy-preserving crowd-guided AI decision-making approach in ethical dilemmas. We adopt the formal and popular notion of differential privacy to quantify privacy, and consider four granularities of privacy protection by taking voter-/record-level privacy protection and centralized/distributed perturbation into account, resulting in four approaches VLCP, RLCP, VLDP, and RLDP, respectively. Moreover, we propose different algorithms to achieve these privacy protection granularities, while retaining the accuracy of the learned moral preference model. Specifically, VLCP and RLCP are implemented with the data aggregator setting a universal privacy parameter and perturbing the averaged moral preference to protect the privacy of voters' data. VLDP and RLDP are implemented in such a way that each voter perturbs her/his local moral preference with a personalized privacy parameter. Extensive experiments based on both synthetic data and real-world data of voters' moral decisions demonstrate that the proposed approaches achieve high accuracy of preference aggregation while protecting individual voter's privacy.

As short videos become an increasingly popular form of storytelling, there is a growing demand for video summarization to convey information concisely with a subset of video frames. Some criteria such as interestingness and diversity are used by existing efforts to pick appropriate segments of content. However, there lacks a mechanism to infuse insights from cinematography and persuasion into this process. As a result, the results of the video summarization sometimes deviate from the original. In addition, the exploration of the vast design space to create customized video summaries is costly for video producer. To address these challenges, we propose a domain specific and idiom adaptive video summarization approach. Specifically, our approach first segments the input video and extracts high-level information from each segment. Such labels are used to represent a collection of idioms and summarization metrics as submodular components which users can combine to create personalized summary styles in a variety of ways. In order to identify the importance of the idioms and metrics in different domains, we leverage max margin learning. Experimental results have validated the effectiveness of our approach. We also plan to release a dataset containing over 600 videos with expert annotations which can benefit further research in this area.

Most machine-learning methods focus on classifying instances whose classes have already been seen in training. In practice, many applications require classifying instances whose classes have not been seen previously. Zero-shot learning is a powerful and promising learning paradigm, in which the classes covered by training instances and the classes we aim to classify are disjoint. In this paper, we provide a comprehensive survey of zero-shot learning. First of all, we provide an overview of zero-shot learning. According to the data utilized in model optimization, we classify zero-shot learning into three learning settings. Second, we describe different semantic spaces adopted in existing zero-shot learning works. Third, we categorize existing zero-shot learning methods and introduce representative methods under each category. Fourth, we discuss different applications of zero-shot learning. Finally, we highlight promising future research directions of zero-shot learning.

Recognizing activities of daily living is an important research topic for health monitoring and elderly care. However, most existing activity recognition models only work with static and pre-defined sensor configurations. Enabling an existing activity recognition model to adapt to the emergence of new sensors in a dynamic environment is a significant challenge. In this paper, we propose a novel feature incremental learning method, namely the Feature Incremental Random Forest (FIRF), to improve the performance of an existing model with a small amount of data on newly appeared features. It consists of two important components - 1) a mutual information based diversity generation strategy (MIDGS) and 2) a feature incremental tree growing mechanism (FITGM). MIDGS enhances the internal diversity of random forests, while FITGM improves the accuracy of individual decision trees. To evaluate the performance of FIRF, we conduct extensive experiments on three well-known public datasets for activity recognition. Experimental results demonstrate that FIRF is significantly more accurate and efficient compared with other state-of-the-art methods. It has the potential to allow the dynamic exploitation of new sensors in changing environments.

Changes in the motor pattern have been shown to be useful advanced indicators of cognitive disorders, such as Parkinson's disease (PD) and cerebral small vessel disease (SVD). It would be highly advantageous to tap into data containing people's motor patterns from motion sensing devices to analyze subtle changes in cognitive abilities, thereby providing personalized interventions before the actual onset of such conditions. However, this goal is very challenging due to two main technical problems: 1) the size of data labeled by doctors is small, and 2) the available data tends to be highly imbalanced (the vast majority tend to be from normal subjects with only a small fraction from subjects with cognitive disorder). In order to effectively deal with these challenges to infer cognitive wellness from motor patterns with high accuracy, we propose the MOtor-Cognitive Analytics (MOCA) framework. The proposed MOCA first uses the random oversampling iterative random forest based feature selection method to reduce the feature space dimensionality and avoid overfitting, and then adds a bias in the optimization problem of weighted extreme learning machine to achieve good generalization ability in handling imbalanced small-sampling dataset. Experimental results on two real-world datasets including SVD and stroke patients show that MOCA can effectively reduce the rate of misdiagnosis and significantly outperform state-of-the-art methods in inferring people's cognitive capabilities. This work opens up opportunities for population-level pre-screening using motion sensing devices and can inform current discussions on reforming the health-care infrastructure.