We study the well-motivated problem of online distribution shift in which the data arrive in batches and the distribution of each batch can change arbitrarily over time. Since the shifts can be large or small, abrupt or gradual, the length of the relevant historical data to learn from may vary over time, which poses a major challenge in designing algorithms that can automatically adapt to the best "attention span” while remaining computationally efficient. We propose a meta-algorithm that takes any network architecture and any Online Learner (OL) algorithm as input and produces a new algorithm which provably enhances the performance of the given OL under non-stationarity. Our algorithm is efficient (it requires maintaining only $O(\log T)$ OL instances) and adaptive (it automatically chooses OL instances with the ideal "attention” length at every timestamp). Experiments on various real-world datasets across text and image modalities show that our method consistently improves the accuracy of user specified OL algorithms for classification tasks. Key novel algorithmic ingredients include a multi-resolution instance design inspired by wavelet theory and a cross-validation-through-time technique. Both could be of independent interest.

Graph Anomaly Detection (GAD) identifies deviant patterns including anomalous nodes, edges, and subgraphs in graph data, with significant applications in social networks, cybersecurity, and financial risk control. While spectral methods have proven effective for homogeneous graph anomaly detection, their application to heterogeneous graphs remains challenging due to structural complexity and semantic richness. Existing heterogeneous graph anomaly detection methods either rely on manually designed meta-paths or decompose the graph into homogeneous subgraphs, leading to limited flexibility or loss of structural integrity. To address these limitations, we propose the Graph Wavelet Transformer (GWT), a novel spectral-based approach that integrates global graph properties and spectral analysis without requiring meta-path information. GWT employs a three-stage process: heterogeneous-to-homogeneous graph conversion, global dependency modeling via graph transformers, and spectral-aware feature enhancement focusing on frequency band components. Extensive experiments on multiple benchmarks demonstrate that GWT significantly outperforms ten baseline methods, providing a new paradigm for heterogeneous graph anomaly detection that preserves structural completeness while achieving computational efficiency.

Anomaly detection in continuous-time event sequences is a crucial task in safety-critical applications. While existing methods primarily focus on developing a superior test statistic, they fail to provide guarantees regarding the false positive rate (FPR), which undermines their reliability in practical deployments. In this paper, we propose CADES (Conformal Anomaly Detection in Event Sequences), a novel test procedure based on conformal inference for the studied task with finite-sample FPR control. Specifically, by using the time-rescaling theorem, we design two powerful non-conformity scores tailored to event sequences, which exhibit complementary sensitivities to different abnormal patterns. CADES combines these scores with Bonferroni correction to leverage their respective strengths and addresses non-identifiability issues of existing methods. Theoretically, we prove the validity of CADES and further provide strong guarantees on calibration-conditional FPR control. Experimental results on synthetic and real-world datasets, covering various types of anomalies, demonstrate that CADES outperforms state-of-the-art methods while maintaining FPR control.

Graph Transformers (GTs) have emerged as powerful tools for handling graph-structured data through global attention mechanisms. While GTs can effectively capture long-range dependencies, they introduce difficulties in optimization due to their complex, non-differentiable operators, which cannot be directly handled by standard gradient-based optimizers (such as Adam or AdamW). To investigate the above issues, this work adopts the line of Zeroth-Order Optimization (ZOO) technique. However, direct integration of ZOO incurs considerable challenges due to the sharp loss landscape and steep gradients within the GT parameter space. Under the above observations, we propose a Sharpness-aware Zeroth-order Optimizer (SZO) that combines Sharpness-Aware Minimization (SAM) technique facilitating convergence within a flatter neighborhood, and leverages parallel computing for efficient gradient estimation. Theoretically, we provide a comprehensive analysis of the optimizer from both convergence and generalization perspectives. Empirically, we conduct extensive experiments on various classical GTs across a wide range of benchmark datasets, which underscore the superior performance of SZO over the state-of-the-art optimizers.

Retrieval-augmented generation (RAG) can enhance the generation quality of large language models (LLMs) by incorporating external token databases. However, retrievals from large databases can constitute a substantial portion of the overall generation time, particularly when retrievals are periodically performed to align the retrieved content with the latest states of generation. In this paper, we introduce PipeRAG, a novel algorithm-system co-design approach to reduce generation latency and enhance generation quality. PipeRAG integrates (1) pipeline parallelism to enable concurrent retrieval and generation processes, (2) flexible retrieval intervals to maximize the efficiency of pipeline parallelism, and (3) a performance model to automatically balance retrieval quality and latency based on the generation states and underlying hardware. Our evaluation shows that, by combining the three aforementioned methods, PipeRAG achieves up to 2.6× speedup in end-to-end generation latency while improving generation quality. These promising results showcase the effectiveness of co-designing algorithms with underlying systems, paving the way for the adoption of PipeRAG in future RAG systems.

Since the seminal work of TabPFN, research on tabular foundation models (TFMs) based on in-context learning (ICL) has challenged long-standing paradigms in machine learning. Without seeing any real-world data, models pretrained on purely synthetic datasets generalize remarkably well across diverse datasets, often using only a moderate number of in-context examples. This shifts the focus in tabular machine learning from model architecture design to the design of synthetic datasets, or, more precisely, to the prior distributions that generate them. Yet the guiding principles for prior design remain poorly understood. This work marks the first attempt to address the gap. We systematically investigate and identify key properties of synthetic priors that allow pretrained TFMs to generalize well. Based on these insights, we introduce Mitra, a TFM trained on a curated mixture of synthetic priors selected for their diversity, distinctiveness, and performance on real-world tabular data. Mitra consistently outperforms state-of-the-art TFMs, such as TabPFNv2 and TabICL, across both classification and regression benchmarks, with better sample efficiency.

Recommender systems powered by generative models (Gen-RecSys) extend beyond classical item-ranking by producing open-ended content, which simultaneously unlocks richer user experiences and introduces new risks. On one hand, these systems can enhance personalization and appeal through dynamic explanations and multi-turn dialogues. On the other hand, they might venture into unknown territory-hallucinating nonexistent items, amplifying bias, or leaking private information. Traditional accuracy metrics cannot fully capture these challenges, as they fail to measure factual correctness, content safety, or alignment with user intent.

We present GENNEXT, a workshop dedicated to exploring the integration of language agents, generative models, and conversational AI within information retrieval (IR) and recommender systems (RS). Building on the success of our recent RecSys'24 workshop, GENNEXT aims to advance discussions on the applications of language agents powered by Large Language Models (LLMs). The workshop will focus on enhancing interactivity between users and systems through multi-turn dialogues, improving creative content generation, advancing personalization, and enabling multifaceted, context-aware decision-making. For example, a language agent could respond to a query like ''Suggest an eco-friendly food tour for a weekend in my city'' by using a recommendation API to identify eateries specializing in sustainable or organic cuisine and a pollution API to ensure the selected routes have low air pollution levels.

Graph-level out-of-distribution (OOD) detection, which attempts to identify OOD graphs originated from an unknown distribution, is a vital building block for safety-critical applications in Web and society. Current approaches concentrate on how to learn better graph representations, but fail to provide any statistically guarantee on detection results, therefore impeding their deployments in the scenario where detection errors would result in serious consequences. To overcome this critical issue, we propose the Conformal Graph-level Out-of-distribution Detection (CGOD), extending the theory of conformal prediction to graph-level OOD detection with a rigorous control over the false positive rate. In CGOD, we develop a new aggregated non-conformity score function based on the proposed adaptive data augmentation. Through the guidance from two designed metrics, i.e., score consistency and representation diversity, our augmentation strategy can generate multiple non-conformity scores, and aggregating these generated non-conformity scores together is robust to the misleading information. Meanwhile, our score function can perceive the subsequent process of conformal inference, enabling the aggregated non-conformity score to be adaptive to different input graphs and deriving a more accurate score estimation. We conduct experiments on multiple real-world datasets with different empirical settings. Extensive results and model analyses demonstrate the superior performance of our approach over several competitive baselines.

In this work, we introduce Context-Aware MultiModal Learner (CaMML), for tuning large multimodal models (LMMs).CaMML, a lightweight module, is crafted to seamlessly integrate multimodal contextual samples into large models, thereby empowering the model to derive knowledge from analogous, domain-specific, up-to-date information and make grounded inferences.Importantly, CaMML is highly scalable and can efficiently handle lengthy multimodal context examples owing to its hierarchical design.Based on CaMML, we have developed two multimodal models, CaMML-7B and CaMML-13B, that have shown exceptional performance across an array of benchmark datasets for multimodal tasks.Remarkably, CaMML-13B achieves the state-of-the-art performance on over ten widely recognized multimodal benchmark datasets, surpassing LLaVA-1.5 (13B) with a noticeable margin, without integration of any external resources.Moreover, we have conducted extensive ablative studies to inspect the inner workings of CaMML and performed qualitative analyses to showcase its effectiveness in handling real-world challenging cases.Code and models are available at: https://github. com/amazon-science/camml.

In the realm of geospatial analysis, the diversity of remote sensors, encompassing both optical and microwave technologies, offers a wealth of distinct observational capabilities. Recognizing this, we present msGFM, a multisensor geospatial foundation model that effectively unifies data from four key sensor modalities. This integration spans an expansive dataset of two million multisensor images. ms-GFM is uniquely adept at handling both paired and unpaired sensor data. For data originating from identical geolocations, our model employs an innovative cross-sensor pretraining approach in masked image modeling, enabling the synthesis of joint representations from diverse sensors. ms-GFM, incorporating four remote sensors, upholds strong performance, forming a comprehensive model adaptable to various sensor types. msGFM has demonstrated enhanced proficiency in a range of both single-sensor and multisensor downstream tasks. These include scene classification, segmentation, cloud removal, and pan-sharpening. A key discovery of our research is that representations derived from natural images are not always compatible with the distinct characteristics of geospatial remote sensors, under-scoring the limitations of existing representations in this field. Our work can serve as a guide for developing multisensor geospatial pretraining models, paving the way for more advanced geospatial capabilities. Code can be found at https://github.com/boranhan/Geospatial_Foundation_Models

Methods based on graph neural networks for solving combinatorial optimization (CO) problems have exhibited promising results in tackling a range of NP-hard problems, eliminating the necessity for reliance on manually created domain knowledge. Existing models including reinforcement learning (RL) framework assume that combinatorial instances in the training set contribute equally during training. Nevertheless, there is considerable variation in the quality of training instances, and the performance of models may suffer from the inclusion of low-quality training instances. This paper expands the current scope of neural solvers for CO problems through the incorporation of curriculum learning (CL). To alleviate the adverse impact of low-quality training instances, we propose CL4CO which utilizes CL strategy, a selective training method, to train models based on the rank of instances' quality in neural Combinatorial Optimization framework. Also, we introduce several candidate topology-aware metrics based on heterophily ratio and evaluation of clustering for the training scheduler. Furthermore, it is noteworthy to emphasize that it has potential to enhance the generalization capacity of RL-based baselines and we give a experimental validation. This enhancement plugin from the fact that CL empowers the acquired RL-based solver to effectively leverage commonly shared features within the same class of CO. Empirically, we conduct a case study on MaxCut, a classical discrete Oil-vector CO, to verify our findings and our results demonstrate that CL4CO is efficient and superiority with good generalization ability.

The question-answering (QA) capabilities of foundation models are highly sensitive to prompt variations, rendering their performance susceptible to superficial, non-meaning-altering changes. This vulnerability often stems from the model's preference or bias towards specific input characteristics, such as option position or superficial image features in multi-modal settings. We propose to rectify this bias directly in the model's internal representation. Our approach, SteerFair, finds the bias direction in the model's representation space and steers activation values away from it during inference. Specifically, we exploit the observation that bias often adheres to simple association rules, such as the spurious association between the first option and correctness likelihood. Next, we construct demonstrations of these rules from unlabeled samples and use them to identify the bias directions. We empirically show that SteerFair significantly reduces instruction-tuned model performance variance across prompt modifications on three benchmark tasks. Remarkably, our approach surpasses a supervised baseline with 100 labels by an average of 10.86% accuracy points and 12.95 score points and matches the performance with 500 labels.

How do we transfer the relevant knowledge from ever larger foundation models into small, task-specific downstream models that can run at much lower costs? Standard transfer learning using pre-trained weights as the initialization transfers limited information and commits us to often massive pre-trained architectures. This procedure also precludes combining multiple pre-trained models that learn complementary information. To address these shortcomings, we introduce Adaptive Feature Transfer (AFT). Instead of transferring weights, AFT operates purely on features, thereby decoupling the choice of the pre-trained model from the smaller downstream model. Rather than indiscriminately compressing all pre-trained features, AFT adaptively transfers pre-trained features that are most useful for performing the downstream task, using a simple regularization that adds minimal overhead. Across multiple vision, language, and multi-modal datasets, AFT achieves significantly better downstream performance compared to alternatives with a similar computational cost. Furthermore, AFT reliably translates improvement in pre-trained models into improvement in downstream performance, even if the downstream model is over $50\times$ smaller, and can effectively transfer complementary information learned by multiple pre-trained models.

We present a novel perspective on temporal point processes (TPPs) by reformulating their intensity processes as solutions to stochastic differential equations (SDEs). In particular, we first prove the equivalent SDE formulations of several classical TPPs, including Poisson processes, Hawkes processes, and self-correcting processes. Based on these proofs, we introduce a unified TPP framework called Neural Jump-Diffusion Temporal Point Process (NJDTPP), whose intensity process is governed by a neural jump-diffusion SDE (NJDSDE) where the drift, diffusion, and jump coefficient functions are parameterized by neural networks. Compared to previous works, NJDTPP exhibits model flexibility in capturing intensity dynamics without relying on any specific functional form, and provides theoretical guarantees regarding the existence and uniqueness of the solution to the proposed NJDSDE. Experiments on both synthetic and real-world datasets demonstrate that NJDTPP is capable of capturing the dynamics of intensity processes in different scenarios and significantly outperforms the state-of-the-art TPP models in prediction tasks.

While the Transformer architecture has achieved remarkable success across various domains, a thorough theoretical foundation explaining its optimization dynamics is yet to be fully developed. In this study, we aim to bridge this understanding gap by answering the following two core questions: (1) Which types of Transformer architectures allow Gradient Descent (GD) to achieve guaranteed convergence? and (2) Under what initial conditions and architectural specifics does the Transformer achieve rapid convergence during training? By analyzing the loss landscape of a single Transformer layer using Softmax and Gaussian attention kernels, our work provides concrete answers to these questions. Our findings demonstrate that, with appropriate weight initialization, GD can train a Transformer model (with either kernel type) to achieve a global optimal solution, especially when the input embedding dimension is large. Nonetheless, certain scenarios highlight potential pitfalls: training a Transformer using the Softmax attention kernel may sometimes lead to suboptimal local solutions. In contrast, the Gaussian attention kernel exhibits a much favorable behavior. Our empirical study further validate the theoretical findings.

We present an overview of a workshop focused on the exploration of generative models within recommender systems (RS). It highlights the dual nature of these technologies: on the one hand, they offer groundbreaking opportunities for enhancing RS through improved personalization, innovative content creation, and interactive user experiences; on the other hand, they introduce a range of challenges, including bias, misinformation, privacy concerns, and environmental impact.

Data augmentation is vital for object detection tasks that require expensive bounding box annotations. Recent successes in diffusion models have inspired the use of diffusion-based synthetic images for data augmentation. However, existing works have primarily focused on image classification, and their applicability to boost object detection’s performance remains unclear. To address this gap, we propose a data augmentation pipeline based on controllable diffusion models and CLIP. Our approach involves generating appropriate visual priors to control the generation of synthetic data and implementing post-filtering techniques using category-calibrated CLIP scores. The evaluation of our approach is conducted under few-shot settings in MSCOCO, full PASCAL VOC dataset, and selected downstream datasets. We observe the performance increase using our augmentation pipeline. Specifically, the mAP improvement is +18.0%/+15.6%/+15.9% for COCO 5/10/30-shot, +2.9% on full PASCAL VOC dataset, and +12.4% on average for selected downstream datasets.

Decision-focused learning (DFL) combines both machine learning and combinatorial optimization so as to enhance the quality of decision-making. In general, DFL adds an optimization layer after the neural network and solves a focused combinatorial optimization problem. The optimization layer is usually based on KKT conditions or surrogate functions. However, the optimizer associated with the DFL is inflexible and requires a large amount of expert knowledge. Furthermore, it has been shown that the optimizer exhibits a lack of robustness and differentiability when confronted with complex tasks, especially in the graph domain. To solve this problem, we study a more generic situation by taking the optimization layer as a black-box operator on graphs. Then, we design a mixed zeroth-order optimization to differentiate the layer. Specifically, we combine DFL with graph neural networks and present a decision-focused graph neural network named ZO4Graph based on a mixed zeroth-order optimizer. Empirically, we conduct extensive experiments compared with two-stage models on community detection. Numerical results demonstrate that our proposed framework outperforms its peers.

Anomaly detection in multi-type event sequences is a crucial and challenging problem with important applications in various domains, including cybersecurity, finance and healthcare. Temporal point process has emerged as a powerful technique for modeling event sequences and has gained considerable attention in the field of anomaly detection. However, existing temporal point process approaches are either inapplicable to multi-type event sequence data or incur the loss of valuable information in subsequences associated with specific event types. To this end, we propose a novel Multiple Hypothesis Testing based Anomaly Detection method (MultiAD) to detect anomalous multi-type event sequences. The basic idea of MultiAD is to capture the underlying distribution of normal sequences using a neural multivariate point process, based on which the original hypothesis testing problem can be converted into a multiple hypothesis testing using the multivariate time rescaling theorem. By conducting multiple hypothesis tests on the time-rescaled subsequences, MultiAD makes full use of the valuable information contained within individual subsequences. Moreover, we claim that the existing test statistic ignores the sequential information of inter-event time intervals and propose new statistics to address this shortcoming. Finally, we employ the kernel method to obtain a smooth estimator of the distribution of the proposed statistics under the null hypothesis. This ensures a more accurate and reliable computation of the p-value, providing robust statistical inference. Extensive experimental results demonstrate that MultiAD significantly outperforms the state-of-the-art methods on both synthetic and real-world data.

Geometric representation learning (e.g., hyperbolic and spherical geometry) has proven to be efficacious in solving many intricate machine learning tasks. The fundamental challenge of geometric representation learning lies in aligning the inherent geometric bias with the underlying structure of the data, which is a rarely explored topic in the literature. Existing methods heavily rely on heuristic assumptions on the data structure to decide the type of geometry to be adopted, which often leads to suboptimal performance. This work aims to automate the alignment process via a data-informed strategy such that we optimize model performance with minimal overhead. Specifically, a sparse gating mechanism is employed to enable each input data point $\mathit{p}$ to select $K$ geometric spaces from a given candidate geometric space pool with $N$ ($K<N$) spaces of different geometry. The selected $K$ spaces are then tightly integrated to formulate a Cartesian product space, which is leveraged to process this input data $\mathit{p}$. In doing so, each input data is processed by the spaces it selected with maximum specialization. We empirically show that this method can effectively align data and spaces without human interventions and further boost performance on real-world tasks, demonstrating its potential in eliciting the expressive power of geometric representations and practical usability.

Vector search has emerged as the foundation for large-scale information retrieval and machine learning systems, with search engines like Google and Bing processing tens of thousands of queries per second on petabyte-scale document datasets by evaluating vector similarities between encoded query texts and web documents. As performance demands for vector search systems surge, accelerated hardware offers a promising solution in the post-Moore's Law era. We introduce FANNS, an end-to-end and scalable vector search framework on FPGAs. Given a user-provided recall requirement on a dataset and a hardware resource budget, FANNS automatically co-designs hardware and algorithm, subsequently generating the corresponding accelerator. The framework also supports scale-out by incorporating a hardware TCP/IP stack in the accelerator. FANNS attains up to 23.0× and 37.2× speedup compared to FPGA and CPU baselines, respectively, and demonstrates superior scalability to GPUs, achieving 5.5× and 7.6× speedup in median and 95th percentile (P95) latency within an eight-accelerator configuration. The remarkable performance of FANNS lays a robust groundwork for future FPGA integration in data centers and AI supercomputers.

Tree-based models and deep neural networks are two schools of effective classification methods in machine learning. While tree-based models are robust irrespective of data domain, deep neural networks have advantages in handling high-dimensional data. Adding a differentiable neural decision forest to the neural network can generally help exploit the benefits of both models. Therefore, traditional decision trees diverge into a bagging version (i.e., random forest) and a boosting version (i.e., gradient boost decision tree). In this work, we aim to harness the advantages of both bagging and boosting by applying gradient boost to a neural decision forest. We propose a gradient boost that can learn the residual using a neural decision forest, considering the residual as a part of the final prediction. Besides, we design a structure for learning the parameters of neural decision forest and gradient boost module in contiguous steps, which is extendable to incorporate multiple gradient-boosting modules in an end-to-end manner. Our extensive experiments on several public datasets demonstrate the competitive performance and efficiency of our model against a series of baseline methods in solving various machine learning tasks.

Malicious activities on the Internet are one of the most dangerous threats to users and organizations. Because of the flexibility a nd accessibility of domains, cyber criminals often utilize them to launch cyber attacks such as phishing or malware. Most of traditional malicious domain detection methods rely on feature engineering to learn the patterns of malicious domains. However, these methods can be easily evaded by some sophisticated evasion techniques such as Domain-flux or Fast-flux. Some recent studies utilized graph-based models to infer malicious domains and achieved better performance, yet without a fine-grained modeling of DNS scenarios. In this paper, we propose an attributed heterogeneous graph neural network model, GAMD, to detect malicious domains in a semi-supervised learning paradigm. Concretely, we utilize attributed heterogeneous information network to model the DNS scenarios with different types of nodes including domain, host, resolved-IP and different types of relation, such as request and resolution relations. We then design a fine-grained node type-aware feature transformation and edge type-aware aggregation mechanism to fuse the node attributes and structure information simultaneously and complete the inference over DNS graphs. In the experiments, we evaluate the performance of our model on a large-scale realworld passive DNS data and show that the proposed method outperforms the state-of-the-art in most evaluation metrics.

We present FleetRec, a high-performance and scalable recommendation inference system within tight latency constraints. FleetRec takes advantage of heterogeneous hardware including GPUs and the latest FPGAs equipped with high-bandwidth memory. By disaggregating computation and memory to different types of hardware and bridging their connections by high-speed network, FleetRec gains the best of both worlds, and can naturally scale out by adding nodes to the cluster. Experiments on three production models up to 114 GB show that FleetRec outperforms optimized CPU baseline by more than one order of magnitude in terms of throughput while achieving significantly lower latency.

Mining from graph-structured data is an integral component of graph data management. A recent trending technique, graph convolutional network (GCN), has gained momentum in the graph mining field, and plays an essential part in numerous graph-related tasks. Although the emerging GCN optimization techniques bring improvements to specific scenarios, they perform diversely in different applications and introduce many trial-and-error costs for practitioners. Moreover, existing GCN models often suffer from oversmoothing problem. Besides, the entanglement of various graph patterns could lead to non-robustness and harm the final performance of GCNs. In this work, we propose a simple yet efficient graph decomposition approach to improve the performance of general graph neural networks. We first empirically study existing graph decomposition methods and propose an automatic connectivity-ware graph decomposition algorithm, DeGNN. To provide a theoretical explanation, we then characterize GCN from the information-theoretic perspective and show that under certain conditions, the mutual information between the output after l layers and the input of GCN converges to 0 exponentially with respect to l. On the other hand, we show that graph decomposition can potentially weaken the condition of such convergence rate, alleviating the information loss when GCN becomes deeper. Extensive experiments on various academic benchmarks and real-world production datasets demonstrate that graph decomposition generally boosts the performance of GNN models. Moreover, our proposed solution DeGNN achieves state-of-the-art performances on almost all these tasks.

The design of expressive representations of entities and relations in a knowledge graph is an important endeavor.While many of the existing approaches have primarily focused on learning from relational patterns and structural information, the intrinsic complexity of KG entities has been more or less overlooked.More concretely, we hypothesize KG entities may be more complex than we think, i.e., an entity may wear many hats and relational triplets may form due to more than a single reason.To this end, this paper proposes to learn disentangled representations of KG entities -a new method that disentangles the inner latent properties of KG entities.Our disentangled process operates at the graph level and a neighborhood mechanism is leveraged to disentangle the hidden properties of each entity.This disentangled representation learning approach is model agnostic and compatible with canonical KG embedding approaches.We conduct extensive experiments on several benchmark datasets, equipping a variety of models (DistMult, SimplE, and QuatE) with our proposed disentangling mechanism.Experimental results demonstrate that our proposed approach substantially improves performance on key metrics.

While standard recurrent neural networks explicitly impose a chain structure on different forms of data, they do not have an explicit bias towards recursive self-instantiation where the extent of recursion is dynamic. Given diverse and even growing data modalities (e.g., logic, algorithmic input and output, music, code, images, and language) that can be expressed in sequences and may benefit from more architectural flexibility, we propose the self-instantiated recurrent unit (Self-IRU) with a novel inductive bias towards dynamic soft recursion. On one hand, theSelf-IRU is characterized by recursive self-instantiation via its gating functions, i.e., gating mechanisms of the Self-IRU are controlled by instances of the Self-IRU itself, which are repeatedly invoked in a recursive fashion. On the other hand, the extent of the Self-IRU recursion is controlled by gates whose values are between 0 and 1 and may vary across the temporal dimension of sequences, enabling dynamic soft recursion depth at each time step. The architectural flexibility and effectiveness of our proposed approach are demonstrated across multiple data modalities. For example, the Self-IRU achieves state-of-the-art performance on the logical inference dataset [Bowman et al., 2014] even when comparing with competitive models that have access to ground-truth syntactic information.

Modeling user interests is crucial in real-world recommender systems. In this paper, we present a new user interest representation model for personalized recommendation. Specifically, the key novelty behind our model is that it explicitly models user interests as a hypercuboid instead of a point in the space. In our approach, the recommendation score is learned by calculating a compositional distance between the user hypercuboid and the item. This helps to alleviate the potential geometric inflexibility of existing collaborative filtering approaches, enabling a greater extent of modeling capability. Furthermore, we present two variants of hypercuboids to enhance the capability in capturing the diversities of user interests. A neural architecture is also proposed to facilitate user hypercuboid learning by capturing the activity sequences (e.g., buy and rate) of users. We demonstrate the effectiveness of our proposed model via extensive experiments on both public and commercial datasets. Empirical results show that our approach achieves very promising results, outperforming existing state-of-the-art.

At online retail platforms, it is crucial to actively detect the risks of transactions to improve customer experience and minimize financial loss. In this work, we propose xFraud, an explainable fraud transaction prediction framework which is mainly composed of a detector and an explainer. The xFraud detector can effectively and efficiently predict the legitimacy of incoming transactions. Specifically, it utilizes a heterogeneous graph neural network to learn expressive representations from the informative heterogeneously typed entities in the transaction logs. The explainer in xFraud can generate meaningful and human-understandable explanations from graphs to facilitate further processes in the business unit. In our experiments with xFraud on real transaction networks with up to 1.1 billion nodes and 3.7 billion edges, xFraud is able to outperform various baseline models in many evaluation metrics while remaining scalable in distributed settings. In addition, we show that xFraud explainer can generate reasonable explanations to significantly assist the business analysis via both quantitative and qualitative evaluations.

This paper investigates the notion of learning user and item representations in non-Euclidean space. Specifically, we study the connection between metric learning in hyperbolic space and collaborative filtering by exploring Mobius gyrovector spaces where the formalism of the spaces could be utilized to generalize the most common Euclidean vector operations. Overall, this work aims to bridge the gap between Euclidean and hyperbolic geometry in recommender systems through metric learning approach. We propose HyperML (Hyperbolic Metric Learning), a conceptually simple but highly effective model for boosting the performance. Via a series of extensive experiments, we show that our proposed HyperML not only outperforms their Euclidean counterparts, but also achieves state-of-the-art performance on multiple benchmark datasets, demonstrating the effectiveness of personalized recommendation in hyperbolic geometry.

Unlike all prior work, in this article, we investigate the notion of “unraveling metric vector spaces,” i.e., deriving meaning and low-rank structure from distance or metric space. Our new model bridges two commonly adopted paradigms for recommendations-metric learning approaches and factorization-based models, distinguishing itself accordingly. More concretely, we show that factorizing a metric vector space can be surprisingly efficacious. All in all, our proposed method, factorized metric learning, is highly effective for two classic recommendation tasks, possessing the potential of displacing many popular choices as an extremely strong baseline. We have done experiments on a number of real-world datasets, which show that our model performs better than recent state of the art largely on the rating prediction and item ranking tasks.

Factorization Machines (FMs) are a class of popular algorithms that have been widely adopted for collaborative filtering and recommendation tasks. FMs are characterized by its usage of the inner product of factorized parameters to model pairwise feature interactions, making it highly expressive and powerful. This paper proposes Holographic Factorization Machines (HFM), a new novel method of enhancing the representation capability of FMs without increasing its parameter size. Our approach replaces the inner product in FMs with holographic reduced representations (HRRs), which are theoretically motivated by associative retrieval and compressed outer products. Empirically, we found that this leads to consistent improvements over vanilla FMs by up to 4% improvement in terms of mean squared error, with improvements larger at smaller parameterization. Additionally, we propose a neural adaptation of HFM which enhances its capability to handle nonlinear structures. We conduct extensive experiments on nine publicly available datasets for collaborative filtering with explicit feedback. HFM achieves state-of-theart performance on all nine, outperforming strong competitors such as Attentional Factorization Machines (AFM) and Neural Matrix Factorization (NeuMF).

Many state-of-the-art neural models for NLP are heavily parameterized and thus memory inefficient.This paper proposes a series of lightweight and memory efficient neural architectures for a potpourri of natural language processing (NLP) tasks.To this end, our models exploit computation using Quaternion algebra and hypercomplex spaces, enabling not only expressive inter-component interactions but also significantly (75%) reduced parameter size due to lesser degrees of freedom in the Hamilton product.We propose Quaternion variants of models, giving rise to new architectures such as the Quaternion attention Model and Quaternion Transformer.Extensive experiments on a battery of NLP tasks demonstrates the utility of proposed Quaternion-inspired models, enabling up to 75% reduction in parameter size without significant loss in performance.

Electroencephalography (EEG) signals reflect and measure activities in certain brain areas. Its zero clinical risk and easy-to-use features make it a good choice of providing insights into the cognitive process. However, effective analysis of time-varying EEG signals remains challenging. First, EEG signal processing and feature engineering are time-consuming and highly rely on expert knowledge, and most existing studies focus on domain-specific classification algorithms, which may not apply to other domains. Second, EEG signals usually have low signal-to-noise ratios and are more chaotic than other sensor signals. In this regard, we propose a generic EEG-based cognitive activity recognition framework that can adaptively support a wide range of cognitive applications to address the above issues. The framework uses a reinforced selective attention model to choose the characteristic information among raw EEG signals automatically. It employs a convolutional mapping operation to dynamically transform the selected information into a feature space to uncover the implicit spatial dependency of EEG sample distribution. We demonstrate the effectiveness of the framework under three representative scenarios: intention recognition with motor imagery EEG, person identification, and neurological diagnosis, and further evaluate it on three widely used public datasets. The experimental results show our framework outperforms multiple state-of-the-art baselines and achieves competitive accuracy on all the datasets while achieving low latency and high resilience in handling complex EEG signals across various domains. The results confirm the suitability of the proposed generic approach for a range of problems in the realm of brain-computer Interface applications.

Deep learning based recommender systems have been extensively explored in recent years. However, the large number of models proposed each year poses a big challenge for both researchers and practitioners in reproducing the results for further comparisons. Although a portion of papers provides source code, they adopted different programming languages or different deep learning packages, which also raises the bar in grasping the ideas. To alleviate this problem, we released the open source project: \textbf{DeepRec}. In this toolkit, we have implemented a number of deep learning based recommendation algorithms using Python and the widely used deep learning package - Tensorflow. Three major recommendation scenarios: rating prediction, top-N recommendation (item ranking) and sequential recommendation, were considered. Meanwhile, DeepRec maintains good modularity and extensibility to easily incorporate new models into the framework. It is distributed under the terms of the GNU General Public License. The source code is available at github: https://github.com/cheungdaven/DeepRec

This paper proposes Quaternion Collaborative Filtering (QCF), a novel representation learning method for recommendation. Our proposed QCF relies on and exploits computation with Quaternion algebra, benefiting from the expressiveness and rich representation learning capability of Hamilton products. Quaternion representations, based on hypercomplex numbers, enable rich inter-latent dependencies between imaginary components. This encourages intricate relations to be captured when learning user-item interactions, serving as a strong inductive bias as compared with the real-space inner product. All in all, we conduct extensive experiments on six real-world datasets, demonstrating the effectiveness of Quaternion algebra in recommender systems. The results exhibit that QCF outperforms a wide spectrum of strong neural baselines on all datasets. Ablative experiments confirm the effectiveness of Hamilton-based composition over multi-embedding composition in real space.

In this work, we move beyond the traditional complex-valued representations, introducing more expressive hypercomplex representations to model entities and relations for knowledge graph embeddings. More specifically, quaternion embeddings, hypercomplex-valued embeddings with three imaginary components, are utilized to represent entities. Relations are modelled as rotations in the quaternion space. The advantages of the proposed approach are: (1) Latent inter-dependencies (between all components) are aptly captured with Hamilton product, encouraging a more compact interaction between entities and relations; (2) Quaternions enable expressive rotation in four-dimensional space and have more degree of freedom than rotation in complex plane; (3) The proposed framework is a generalization of ComplEx on hypercomplex space while offering better geometrical interpretations, concurrently satisfying the key desiderata of relational representation learning (i.e., modeling symmetry, anti-symmetry and inversion). Experimental results demonstrate that our method achieves state-of-the-art performance on four well-established knowledge graph completion benchmarks.

A brain-computer interface (BCI) acquires brain signals, analyzes, and translates them into commands that are relayed to actuation devices for carrying out desired actions. With the widespread connectivity of everyday devices realized by the advent of the Internet of Things (IoT), BCI can empower individuals to directly control objects such as smart home appliances or assistive robots, directly via their thoughts. However, realization of this vision is faced with a number of challenges, most importantly being the issue of accurately interpreting the intent of the individual from the raw brain signals that are often of low fidelity and subject to noise. Moreover, preprocessing brain signals and the subsequent feature engineering are both time-consuming and highly reliant on human domain expertise. To address the aforementioned issues, in this paper, we propose a unified deep learning-based framework that enables effective human-thing cognitive interactivity in order to bridge individuals and IoT objects. We design a reinforcement learning-based selective attention mechanism (SAM) to discover the distinctive features from the input brain signals. In addition, we propose a modified long short-term memory to distinguish the interdimensional information forwarded from the SAM. To evaluate the efficiency of the proposed framework, we conduct extensive real-world experiments and demonstrate that our model outperforms a number of competitive state-of-the-art baselines. Two practical real-time human-thing cognitive interaction applications are presented to validate the feasibility of our approach.

Modeling user-item interaction patterns is an important task for personalized recommendations. Many recommender systems are based on the assumption that there exists a linear relationship between users and items while neglecting the intricacy and non-linearity of real-life historical interactions. In this paper, we propose a neural network based recommendation model (NeuRec) that untangles the complexity of user-item interactions and establish an integrated network to combine non-linear transformation with latent factors. We further design two variants of NeuRec: user-based NeuRec and item-based NeuRec, by focusing on different aspects of the interaction matrix. Extensive experiments on four real-world datasets demonstrated their superior performances on personalized ranking task.

Collaborative filtering (CF) has been successfully used to provide users with personalized products and services. However, dealing with the increasing sparseness of user-item matrix still remains a challenge. To tackle such issue, hybrid CF such as combining with content based filtering and leveraging side information of users and items has been extensively studied to enhance performance. However, most of these approaches depend on hand-crafted feature engineering, which is usually noise-prone and biased by different feature extraction and selection schemes. In this paper, we propose a new hybrid model by generalizing contractive auto-encoder paradigm into matrix factorization framework with good scalability and computational efficiency, which jointly models content information as representations of effectiveness and compactness, and leverage implicit user feedback to make accurate recommendations. Extensive experiments conducted over three large-scale real datasets indicate the proposed approach outperforms the compared methods for item recommendation.