Machine Learning (cs.LG)

Abstract: This paper proposes an efficient optimizer called AdaPlus which integrates Nesterov momentum and precise stepsize adjustment on AdamW basis. AdaPlus combines the advantages of AdamW, Nadam, and AdaBelief and, in particular, does not introduce any extra hyper-parameters. We perform extensive experimental evaluations on three machine learning tasks to validate the effectiveness of AdaPlus. The experiment results validate that AdaPlus (i) is the best adaptive method which performs most comparable with (even slightly better than) SGD with momentum on image classification tasks and (ii) outperforms other state-of-the-art optimizers on language modeling tasks and illustrates the highest stability when training GANs. The experiment code of AdaPlus is available at: https://github.com/guanleics/AdaPlus.

Abstract: In many learning applications, data are collected from multiple sources, each providing a \emph{batch} of samples that by itself is insufficient to learn its input-output relationship. A common approach assumes that the sources fall in one of several unknown subgroups, each with an unknown input distribution and input-output relationship. We consider one of this setup's most fundamental and important manifestations where the output is a noisy linear combination of the inputs, and there are $k$ subgroups, each with its own regression vector. Prior work~\cite{kong2020meta} showed that with abundant small-batches, the regression vectors can be learned with only few, $\tilde\Omega( k^{3/2})$, batches of medium-size with $\tilde\Omega(\sqrt k)$ samples each. However, the paper requires that the input distribution for all $k$ subgroups be isotropic Gaussian, and states that removing this assumption is an ``interesting and challenging problem". We propose a novel gradient-based algorithm that improves on the existing results in several ways. It extends the applicability of the algorithm by: (1) allowing the subgroups' underlying input distributions to be different, unknown, and heavy-tailed; (2) recovering all subgroups followed by a significant proportion of batches even for infinite $k$; (3) removing the separation requirement between the regression vectors; (4) reducing the number of batches and allowing smaller batch sizes.

Abstract: The recurrent neural network and its variants have shown great success in processing sequences in recent years. However, this deep neural network has not aroused much attention in anomaly detection through predictively process monitoring. Furthermore, the traditional statistic models work on assumptions and hypothesis tests, while neural network (NN) models do not need that many assumptions. This flexibility enables NN models to work efficiently on data with time-varying variability, a common inherent aspect of data in practice. This paper explores the ability of the recurrent neural network structure to monitor processes and proposes a control chart based on long short-term memory (LSTM) prediction intervals for data with time-varying variability. The simulation studies provide empirical evidence that the proposed model outperforms other NN-based predictive monitoring methods for mean shift detection. The proposed method is also applied to time series sensor data, which confirms that the proposed method is an effective technique for detecting abnormalities.

Abstract: Spatial time series imputation is critically important to many real applications such as intelligent transportation and air quality monitoring. Although recent transformer and diffusion model based approaches have achieved significant performance gains compared with conventional statistic based methods, spatial time series imputation still remains as a challenging issue due to the complex spatio-temporal dependencies and the noise uncertainty of the spatial time series data. Especially, recent diffusion process based models may introduce random noise to the imputations, and thus cause negative impact on the model performance. To this end, we propose a self-adaptive noise scaling diffusion model named SaSDim to more effectively perform spatial time series imputation. Specially, we propose a new loss function that can scale the noise to the similar intensity, and propose the across spatial-temporal global convolution module to more effectively capture the dynamic spatial-temporal dependencies. Extensive experiments conducted on three real world datasets verify the effectiveness of SaSDim by comparison with current state-of-the-art baselines.

Abstract: Many decision problems cannot be solved exactly and use several estimation algorithms that assign scores to the different available options. The estimation errors can have various correlations, from low (e.g. between two very different approaches) to high (e.g. when using a given algorithm with different hyperparameters). Most aggregation rules would suffer from this diversity of correlations. In this article, we propose different aggregation rules that take correlations into account, and we compare them to naive rules in various experiments based on synthetic data. Our results show that when sufficient information is known about the correlations between errors, a maximum likelihood aggregation should be preferred. Otherwise, typically with limited training data, we recommend a method that we call Embedded Voting (EV).

Abstract: Urban traffic emissions represent a significant concern due to their detrimental impacts on both public health and the environment. Consequently, decision-makers have flagged their reduction as a crucial goal. In this study, we first analyze the correlation between traffic flux and pollution in the city of Valencia, Spain. Our results demonstrate that traffic has a significant impact on the levels of certain pollutants (especially $\text{NO}_\text{x}$). Secondly, we develop an alarm system to predict if a street is likely to experience unusually high traffic in the next 30 minutes, using an independent three-tier level for each street. To make the predictions, we use traffic data updated every 10 minutes and Long Short-Term Memory (LSTM) neural networks. We trained the LSTM using traffic data from 2018, and tested it using traffic data from 2019.

Abstract: Self-Supervised Learning (SSL) is an important paradigm for learning representations from unlabelled data, and SSL with neural networks has been highly successful in practice. However current theoretical analysis of SSL is mostly restricted to generalisation error bounds. In contrast, learning dynamics often provide a precise characterisation of the behaviour of neural networks based models but, so far, are mainly known in supervised settings. In this paper, we study the learning dynamics of SSL models, specifically representations obtained by minimising contrastive and non-contrastive losses. We show that a naive extension of the dymanics of multivariate regression to SSL leads to learning trivial scalar representations that demonstrates dimension collapse in SSL. Consequently, we formulate SSL objectives with orthogonality constraints on the weights, and derive the exact (network width independent) learning dynamics of the SSL models trained using gradient descent on the Grassmannian manifold. We also argue that the infinite width approximation of SSL models significantly deviate from the neural tangent kernel approximations of supervised models. We numerically illustrate the validity of our theoretical findings, and discuss how the presented results provide a framework for further theoretical analysis of contrastive and non-contrastive SSL.

Abstract: Continuous-time dynamic graph modeling is a crucial task for many real-world applications, such as financial risk management and fraud detection. Though existing dynamic graph modeling methods have achieved satisfactory results, they still suffer from three key limitations, hindering their scalability and further applicability. i) Indiscriminate updating. For incoming edges, existing methods would indiscriminately deal with them, which may lead to more time consumption and unexpected noisy information. ii) Ineffective node-wise long-term modeling. They heavily rely on recurrent neural networks (RNNs) as a backbone, which has been demonstrated to be incapable of fully capturing node-wise long-term dependencies in event sequences. iii) Neglect of re-occurrence patterns. Dynamic graphs involve the repeated occurrence of neighbors that indicates their importance, which is disappointedly neglected by existing methods. In this paper, we present iLoRE, a novel dynamic graph modeling method with instant node-wise Long-term modeling and Re-occurrence preservation. To overcome the indiscriminate updating issue, we introduce the Adaptive Short-term Updater module that will automatically discard the useless or noisy edges, ensuring iLoRE's effectiveness and instant ability. We further propose the Long-term Updater to realize more effective node-wise long-term modeling, where we innovatively propose the Identity Attention mechanism to empower a Transformer-based updater, bypassing the limited effectiveness of typical RNN-dominated designs. Finally, the crucial re-occurrence patterns are also encoded into a graph module for informative representation learning, which will further improve the expressiveness of our method. Our experimental results on real-world datasets demonstrate the effectiveness of our iLoRE for dynamic graph modeling.

Abstract: Internet of Things (IoT) sensors are nowadays heavily utilized in various real-world applications ranging from wearables to smart buildings passing by agrotechnology and health monitoring. With the huge amounts of data generated by these tiny devices, Deep Learning (DL) models have been extensively used to enhance them with intelligent processing. However, with the urge for smaller and more accurate devices, DL models became too heavy to deploy. It is thus necessary to incorporate the hardware's limited resources in the design process. Therefore, inspired by the human brain known for its efficiency and low power consumption, we propose a shallow bidirectional network based on predictive coding theory and dynamic early exiting for halting further computations when a performance threshold is surpassed. We achieve comparable accuracy to VGG-16 in image classification on CIFAR-10 with fewer parameters and less computational complexity.

Abstract: Temporal Graph Networks (TGNs) have shown remarkable performance in learning representation for continuous-time dynamic graphs. However, real-world dynamic graphs typically contain diverse and intricate noise. Noise can significantly degrade the quality of representation generation, impeding the effectiveness of TGNs in downstream tasks. Though structure learning is widely applied to mitigate noise in static graphs, its adaptation to dynamic graph settings poses two significant challenges. i) Noise dynamics. Existing structure learning methods are ill-equipped to address the temporal aspect of noise, hampering their effectiveness in such dynamic and ever-changing noise patterns. ii) More severe noise. Noise may be introduced along with multiple interactions between two nodes, leading to the re-pollution of these nodes and consequently causing more severe noise compared to static graphs. In this paper, we present RDGSL, a representation learning method in continuous-time dynamic graphs. Meanwhile, we propose dynamic graph structure learning, a novel supervisory signal that empowers RDGSL with the ability to effectively combat noise in dynamic graphs. To address the noise dynamics issue, we introduce the Dynamic Graph Filter, where we innovatively propose a dynamic noise function that dynamically captures both current and historical noise, enabling us to assess the temporal aspect of noise and generate a denoised graph. We further propose the Temporal Embedding Learner to tackle the challenge of more severe noise, which utilizes an attention mechanism to selectively turn a blind eye to noisy edges and hence focus on normal edges, enhancing the expressiveness for representation generation that remains resilient to noise. Our method demonstrates robustness towards downstream tasks, resulting in up to 5.1% absolute AUC improvement in evolving classification versus the second-best baseline.

Abstract: Multivariate Hawkes processes (MHPs) are versatile probabilistic tools used to model various real-life phenomena: earthquakes, operations on stock markets, neuronal activity, virus propagation and many others. In this paper, we focus on MHPs with exponential decay kernels and estimate connectivity graphs, which represent the Granger causal relations between their components. We approach this inference problem by proposing an optimization criterion and model selection algorithm based on the minimum message length (MML) principle. MML compares Granger causal models using the Occam's razor principle in the following way: even when models have a comparable goodness-of-fit to the observed data, the one generating the most concise explanation of the data is preferred. While most of the state-of-art methods using lasso-type penalization tend to overfitting in scenarios with short time horizons, the proposed MML-based method achieves high F1 scores in these settings. We conduct a numerical study comparing the proposed algorithm to other related classical and state-of-art methods, where we achieve the highest F1 scores in specific sparse graph settings. We illustrate the proposed method also on G7 sovereign bond data and obtain causal connections, which are in agreement with the expert knowledge available in the literature.

Abstract: Unsupervised and self-supervised representation learning has become popular in recent years for learning useful features from unlabelled data. Representation learning has been mostly developed in the neural network literature, and other models for representation learning are surprisingly unexplored. In this work, we introduce and analyze several kernel-based representation learning approaches: Firstly, we define two kernel Self-Supervised Learning (SSL) models using contrastive loss functions and secondly, a Kernel Autoencoder (AE) model based on the idea of embedding and reconstructing data. We argue that the classical representer theorems for supervised kernel machines are not always applicable for (self-supervised) representation learning, and present new representer theorems, which show that the representations learned by our kernel models can be expressed in terms of kernel matrices. We further derive generalisation error bounds for representation learning with kernel SSL and AE, and empirically evaluate the performance of these methods in both small data regimes as well as in comparison with neural network based models.

Abstract: The immense evolution in Large Language Models (LLMs) has underscored the importance of massive, diverse, and high-quality data. Despite this, existing open-source tools for LLM data processing remain limited and mostly tailored to specific datasets, with an emphasis on the reproducibility of released data over adaptability and usability, inhibiting potential applications. In response, we propose a one-stop, powerful yet flexible and user-friendly LLM data processing system named Data-Juicer. Our system offers over 50 built-in versatile operators and pluggable tools, which synergize modularity, composability, and extensibility dedicated to diverse LLM data processing needs. By incorporating visualized and automatic evaluation capabilities, Data-Juicer enables a timely feedback loop to accelerate data processing and gain data insights. To enhance usability, Data-Juicer provides out-of-the-box components for users with various backgrounds, and fruitful data recipes for LLM pre-training and post-tuning usages. Further, we employ multi-facet system optimization and seamlessly integrate Data-Juicer with both LLM and distributed computing ecosystems, to enable efficient and scalable data processing. Empirical validation of the generated data recipes reveals considerable improvements in LLaMA performance for various pre-training and post-tuning cases, demonstrating up to 7.45% relative improvement of averaged score across 16 LLM benchmarks and 16.25% higher win rate using pair-wise GPT-4 evaluation. The system's efficiency and scalability are also validated, supported by up to 88.7% reduction in single-machine processing time, 77.1% and 73.1% less memory and CPU usage respectively, and 7.91x processing acceleration when utilizing distributed computing ecosystems. Our system, data recipes, and multiple tutorial demos are released, calling for broader research centered on LLM data.

Abstract: Inverse design refers to the problem of optimizing the input of an objective function in order to enact a target outcome. For many real-world engineering problems, the objective function takes the form of a simulator that predicts how the system state will evolve over time, and the design challenge is to optimize the initial conditions that lead to a target outcome. Recent developments in learned simulation have shown that graph neural networks (GNNs) can be used for accurate, efficient, differentiable estimation of simulator dynamics, and support high-quality design optimization with gradient- or sampling-based optimization procedures. However, optimizing designs from scratch requires many expensive model queries, and these procedures exhibit basic failures on either non-convex or high-dimensional problems.In this work, we show how denoising diffusion models (DDMs) can be used to solve inverse design problems efficiently and propose a particle sampling algorithm for further improving their efficiency. We perform experiments on a number of fluid dynamics design challenges, and find that our approach substantially reduces the number of calls to the simulator compared to standard techniques.

Abstract: In this paper, we propose a novel probabilistic self-supervised learning via Scoring Rule Minimization (ProSMIN), which leverages the power of probabilistic models to enhance representation quality and mitigate collapsing representations. Our proposed approach involves two neural networks; the online network and the target network, which collaborate and learn the diverse distribution of representations from each other through knowledge distillation. By presenting the input samples in two augmented formats, the online network is trained to predict the target network representation of the same sample under a different augmented view. The two networks are trained via our new loss function based on proper scoring rules. We provide a theoretical justification for ProSMIN's convergence, demonstrating the strict propriety of its modified scoring rule. This insight validates the method's optimization process and contributes to its robustness and effectiveness in improving representation quality. We evaluate our probabilistic model on various downstream tasks, such as in-distribution generalization, out-of-distribution detection, dataset corruption, low-shot learning, and transfer learning. Our method achieves superior accuracy and calibration, surpassing the self-supervised baseline in a wide range of experiments on large-scale datasets like ImageNet-O and ImageNet-C, ProSMIN demonstrates its scalability and real-world applicability.

Abstract: Online learning algorithms have been successfully used to design caching policies with regret guarantees. Existing algorithms assume that the cache knows the exact request sequence, but this may not be feasible in high load and/or memory-constrained scenarios, where the cache may have access only to sampled requests or to approximate requests' counters. In this paper, we propose the Noisy-Follow-the-Perturbed-Leader (NFPL) algorithm, a variant of the classic Follow-the-Perturbed-Leader (FPL) when request estimates are noisy, and we show that the proposed solution has sublinear regret under specific conditions on the requests estimator. The experimental evaluation compares the proposed solution against classic caching policies and validates the proposed approach under both synthetic and real request traces.

Abstract: Artificial Intelligence (AI) is currently spearheaded by machine learning (ML) methods such as deep learning (DL) which have accelerated progress on many tasks thought to be out of reach of AI. These ML methods can often be compute hungry, energy intensive, and result in significant carbon emissions, a known driver of anthropogenic climate change. Additionally, the platforms on which ML systems run are associated with environmental impacts including and beyond carbon emissions. The solution lionized by both industry and the ML community to improve the environmental sustainability of ML is to increase the efficiency with which ML systems operate in terms of both compute and energy consumption. In this perspective, we argue that efficiency alone is not enough to make ML as a technology environmentally sustainable. We do so by presenting three high level discrepancies between the effect of efficiency on the environmental sustainability of ML when considering the many variables which it interacts with. In doing so, we comprehensively demonstrate, at multiple levels of granularity both technical and non-technical reasons, why efficiency is not enough to fully remedy the environmental impacts of ML. Based on this, we present and argue for systems thinking as a viable path towards improving the environmental sustainability of ML holistically.

Abstract: This paper is concerned with deep generative models (DGMs) for unsupervised out-of-distribution (OOD) detection. In particular, we focus on vanilla Variational Autoencoders (VAE) that use a standard normal prior distribution for the latent variables. These models have a smaller model size, enabling faster training and inference, making them well-suited for resource-limited applications compared to more complex DGMs. We propose a novel OOD score called Error Reduction (ER) specifically designed for vanilla VAE. ER incorporate the idea of reconstructing image inputs from their lossy counterparts and takes into account the Kolmogorov complexity of the images. Experimental results on diverse datasets demonstrate the superiority of our approach over baseline methods. Our code is available at: https://github.com/ZJLAB-AMMI/VAE4OOD.

Abstract: Storing and streaming high dimensional data for foundation model training became a critical requirement with the rise of foundation models beyond natural language. In this paper we introduce TensorBank, a petabyte scale tensor lakehouse capable of streaming tensors from Cloud Object Store (COS) to GPU memory at wire speed based on complex relational queries. We use Hierarchical Statistical Indices (HSI) for query acceleration. Our architecture allows to directly address tensors on block level using HTTP range reads. Once in GPU memory, data can be transformed using PyTorch transforms. We provide a generic PyTorch dataset type with a corresponding dataset factory translating relational queries and requested transformations as an instance. By making use of the HSI, irrelevant blocks can be skipped without reading them as those indices contain statistics on their content at different hierarchical resolution levels. This is an opinionated architecture powered by open standards and making heavy use of open-source technology. Although, hardened for production use using geospatial-temporal data, this architecture generalizes to other use case like computer vision, computational neuroscience, biological sequence analysis and more.

Abstract: Sleep stage classification is crucial for detecting patients' health conditions. Existing models, which mainly use Convolutional Neural Networks (CNN) for modelling Euclidean data and Graph Convolution Networks (GNN) for modelling non-Euclidean data, are unable to consider the heterogeneity and interactivity of multimodal data as well as the spatial-temporal correlation simultaneously, which hinders a further improvement of classification performance. In this paper, we propose a dynamic learning framework STHL, which introduces hypergraph to encode spatial-temporal data for sleep stage classification. Hypergraphs can construct multi-modal/multi-type data instead of using simple pairwise between two subjects. STHL creates spatial and temporal hyperedges separately to build node correlations, then it conducts type-specific hypergraph learning process to encode the attributes into the embedding space. Extensive experiments show that our proposed STHL outperforms the state-of-the-art models in sleep stage classification tasks.

Abstract: We propose the simplest SGD enhanced method ever, Loss-Controlled Asymmetric Momentum(LCAM), aimed directly at the Saddle Point problem. Compared to the traditional SGD with Momentum, there's no increase in computational demand, yet it outperforms all current optimizers. We use the concepts of weight conjugation and traction effect to explain this phenomenon. We designed experiments to rapidly reduce the learning rate at specified epochs to trap parameters more easily at saddle points. We selected WRN28-10 as the test network and chose cifar10 and cifar100 as test datasets, an identical group to the original paper of WRN and Cosine Annealing Scheduling(CAS). We compared the ability to bypass saddle points of Asymmetric Momentum with different priorities. Finally, using WRN28-10 on Cifar100, we achieved a peak average test accuracy of 80.78\% around 120 epoch. For comparison, the original WRN paper reported 80.75\%, while CAS was at 80.42\%, all at 200 epoch. This means that while potentially increasing accuracy, we use nearly half convergence time. Our demonstration code is available at\\ https://github.com/hakumaicc/Asymmetric-Momentum-LCAM

Abstract: The aim of this work is to introduce Generalized Simplicial Attention Neural Networks (GSANs), i.e., novel neural architectures designed to process data defined on simplicial complexes using masked self-attentional layers. Hinging on topological signal processing principles, we devise a series of self-attention schemes capable of processing data components defined at different simplicial orders, such as nodes, edges, triangles, and beyond. These schemes learn how to weight the neighborhoods of the given topological domain in a task-oriented fashion, leveraging the interplay among simplices of different orders through the Dirac operator and its Dirac decomposition. We also theoretically establish that GSANs are permutation equivariant and simplicial-aware. Finally, we illustrate how our approach compares favorably with other methods when applied to several (inductive and transductive) tasks such as trajectory prediction, missing data imputation, graph classification, and simplex prediction.

Abstract: Model-based offline reinforcement learning (RL), which builds a supervised transition model with logging dataset to avoid costly interactions with the online environment, has been a promising approach for offline policy optimization. As the discrepancy between the logging data and online environment may result in a distributional shift problem, many prior works have studied how to build robust transition models conservatively and estimate the model uncertainty accurately. However, the over-conservatism can limit the exploration of the agent, and the uncertainty estimates may be unreliable. In this work, we propose a novel Model-based Offline policy optimization framework with Adversarial Network (MOAN). The key idea is to use adversarial learning to build a transition model with better generalization, where an adversary is introduced to distinguish between in-distribution and out-of-distribution samples. Moreover, the adversary can naturally provide a quantification of the model's uncertainty with theoretical guarantees. Extensive experiments showed that our approach outperforms existing state-of-the-art baselines on widely studied offline RL benchmarks. It can also generate diverse in-distribution samples, and quantify the uncertainty more accurately.

Abstract: We show that participating in federated learning can be detrimental to group fairness. In fact, the bias of a few parties against under-represented groups (identified by sensitive attributes such as gender or race) can propagate through the network to all the parties in the network. We analyze and explain bias propagation in federated learning on naturally partitioned real-world datasets. Our analysis reveals that biased parties unintentionally yet stealthily encode their bias in a small number of model parameters, and throughout the training, they steadily increase the dependence of the global model on sensitive attributes. What is important to highlight is that the experienced bias in federated learning is higher than what parties would otherwise encounter in centralized training with a model trained on the union of all their data. This indicates that the bias is due to the algorithm. Our work calls for auditing group fairness in federated learning and designing learning algorithms that are robust to bias propagation.

Abstract: Due to the complexity of modern computer systems, novel and unexpected behaviors frequently occur. Such deviations are either normal occurrences, such as software updates and new user activities, or abnormalities, such as misconfigurations, latency issues, intrusions, and software bugs. Regardless, novel behaviors are of great interest to developers, and there is a genuine need for efficient and effective methods to detect them. Nowadays, researchers consider system calls to be the most fine-grained and accurate source of information to investigate the behavior of computer systems. Accordingly, this paper introduces a novelty detection methodology that relies on a probability distribution over sequences of system calls, which can be seen as a language model. Language models estimate the likelihood of sequences, and since novelties deviate from previously observed behaviors by definition, they would be unlikely under the model. Following the success of neural networks for language models, three architectures are evaluated in this work: the widespread LSTM, the state-of-the-art Transformer, and the lower-complexity Longformer. However, large neural networks typically require an enormous amount of data to be trained effectively, and to the best of our knowledge, no massive modern datasets of kernel traces are publicly available. This paper addresses this limitation by introducing a new open-source dataset of kernel traces comprising over 2 million web requests with seven distinct behaviors. The proposed methodology requires minimal expert hand-crafting and achieves an F-score and AuROC greater than 95% on most novelties while being data- and task-agnostic. The source code and trained models are publicly available on GitHub while the datasets are available on Zenodo.

Abstract: Equilibrium propagation (EP) is a compelling alternative to the backpropagation of error algorithm (BP) for computing gradients of neural networks on biological or analog neuromorphic substrates. Still, the algorithm requires weight symmetry and infinitesimal equilibrium perturbations, i.e., nudges, to estimate unbiased gradients efficiently. Both requirements are challenging to implement in physical systems. Yet, whether and how weight asymmetry affects its applicability is unknown because, in practice, it may be masked by biases introduced through the finite nudge. To address this question, we study generalized EP, which can be formulated without weight symmetry, and analytically isolate the two sources of bias. For complex-differentiable non-symmetric networks, we show that the finite nudge does not pose a problem, as exact derivatives can still be estimated via a Cauchy integral. In contrast, weight asymmetry introduces bias resulting in low task performance due to poor alignment of EP's neuronal error vectors compared to BP. To mitigate this issue, we present a new homeostatic objective that directly penalizes functional asymmetries of the Jacobian at the network's fixed point. This homeostatic objective dramatically improves the network's ability to solve complex tasks such as ImageNet 32x32. Our results lay the theoretical groundwork for studying and mitigating the adverse effects of imperfections of physical networks on learning algorithms that rely on the substrate's relaxation dynamics.

Abstract: Three major challenges in reinforcement learning are the complex dynamical systems with large state spaces, the costly data acquisition processes, and the deviation of real-world dynamics from the training environment deployment. To overcome these issues, we study distributionally robust Markov decision processes with continuous state spaces under the widely used Kullback-Leibler, chi-square, and total variation uncertainty sets. We propose a model-based approach that utilizes Gaussian Processes and the maximum variance reduction algorithm to efficiently learn multi-output nominal transition dynamics, leveraging access to a generative model (i.e., simulator). We further demonstrate the statistical sample complexity of the proposed method for different uncertainty sets. These complexity bounds are independent of the number of states and extend beyond linear dynamics, ensuring the effectiveness of our approach in identifying near-optimal distributionally-robust policies. The proposed method can be further combined with other model-free distributionally robust reinforcement learning methods to obtain a near-optimal robust policy. Experimental results demonstrate the robustness of our algorithm to distributional shifts and its superior performance in terms of the number of samples needed.

Abstract: Sample size calculation is an essential step in most data-based disciplines. Large enough samples ensure representativeness of the population and determine the precision of estimates. This is true for most quantitative studies, including those that employ machine learning methods, such as natural language processing, where free-text is used to generate predictions and classify instances of text. Within the healthcare domain, the lack of sufficient corpora of previously collected data can be a limiting factor when determining sample sizes for new studies. This paper tries to address the issue by making recommendations on sample sizes for text classification tasks in the healthcare domain. Models trained on the MIMIC-III database of critical care records from Beth Israel Deaconess Medical Center were used to classify documents as having or not having Unspecified Essential Hypertension, the most common diagnosis code in the database. Simulations were performed using various classifiers on different sample sizes and class proportions. This was repeated for a comparatively less common diagnosis code within the database of diabetes mellitus without mention of complication. Smaller sample sizes resulted in better results when using a K-nearest neighbours classifier, whereas larger sample sizes provided better results with support vector machines and BERT models. Overall, a sample size larger than 1000 was sufficient to provide decent performance metrics. The simulations conducted within this study provide guidelines that can be used as recommendations for selecting appropriate sample sizes and class proportions, and for predicting expected performance, when building classifiers for textual healthcare data. The methodology used here can be modified for sample size estimates calculations with other datasets.

Abstract: Current time-series forecasting problems use short-term weather attributes as exogenous inputs. However, in specific time-series forecasting solutions (e.g., demand prediction in the supply chain), seasonal climate predictions are crucial to improve its resilience. Representing mid to long-term seasonal climate forecasts is challenging as seasonal climate predictions are uncertain, and encoding spatio-temporal relationship of climate forecasts with demand is complex. We propose a novel modeling framework that efficiently encodes seasonal climate predictions to provide robust and reliable time-series forecasting for supply chain functions. The encoding framework enables effective learning of latent representations -- be it uncertain seasonal climate prediction or other time-series data (e.g., buyer patterns) -- via a modular neural network architecture. Our extensive experiments indicate that learning such representations to model seasonal climate forecast results in an error reduction of approximately 13\% to 17\% across multiple real-world data sets compared to existing demand forecasting methods.

Abstract: In the domain of machine learning algorithms, the significance of the loss function is paramount, especially in supervised learning tasks. It serves as a fundamental pillar that profoundly influences the behavior and efficacy of supervised learning algorithms. Traditional loss functions, while widely used, often struggle to handle noisy and high-dimensional data, impede model interpretability, and lead to slow convergence during training. In this paper, we address the aforementioned constraints by proposing a novel robust, bounded, sparse, and smooth (RoBoSS) loss function for supervised learning. Further, we incorporate the RoBoSS loss function within the framework of support vector machine (SVM) and introduce a new robust algorithm named $\mathcal{L}_{rbss}$-SVM. For the theoretical analysis, the classification-calibrated property and generalization ability are also presented. These investigations are crucial for gaining deeper insights into the performance of the RoBoSS loss function in the classification tasks and its potential to generalize well to unseen data. To empirically demonstrate the effectiveness of the proposed $\mathcal{L}_{rbss}$-SVM, we evaluate it on $88$ real-world UCI and KEEL datasets from diverse domains. Additionally, to exemplify the effectiveness of the proposed $\mathcal{L}_{rbss}$-SVM within the biomedical realm, we evaluated it on two medical datasets: the electroencephalogram (EEG) signal dataset and the breast cancer (BreaKHis) dataset. The numerical results substantiate the superiority of the proposed $\mathcal{L}_{rbss}$-SVM model, both in terms of its remarkable generalization performance and its efficiency in training time.

Abstract: A clear need for automatic anomaly detection applied to automotive testing has emerged as more and more attention is paid to the data recorded and manual evaluation by humans reaches its capacity. Such real-world data is massive, diverse, multivariate and temporal in nature, therefore requiring modelling of the testee behaviour. We propose a variational autoencoder with multi-head attention (MA-VAE), which, when trained on unlabelled data, not only provides very few false positives but also manages to detect the majority of the anomalies presented. In addition to that, the approach offers a novel way to avoid the bypass phenomenon, an undesirable behaviour investigated in literature. Lastly, the approach also introduces a new method to remap individual windows to a continuous time series. The results are presented in the context of a real-world industrial data set and several experiments are undertaken to further investigate certain aspects of the proposed model. When configured properly, it is 9% of the time wrong when an anomaly is flagged and discovers 67% of the anomalies present. Also, MA-VAE has the potential to perform well with only a fraction of the training and validation subset, however, to extract it, a more sophisticated threshold estimation method is required.

Abstract: This paper introduces a novel graph-based filter method for automatic feature selection (abbreviated as GB-AFS) for multi-class classification tasks. The method determines the minimum combination of features required to sustain prediction performance while maintaining complementary discriminating abilities between different classes. It does not require any user-defined parameters such as the number of features to select. The methodology employs the Jeffries-Matusita (JM) distance in conjunction with t-distributed Stochastic Neighbor Embedding (t-SNE) to generate a low-dimensional space reflecting how effectively each feature can differentiate between each pair of classes. The minimum number of features is selected using our newly developed Mean Simplified Silhouette (abbreviated as MSS) index, designed to evaluate the clustering results for the feature selection task. Experimental results on public data sets demonstrate the superior performance of the proposed GB-AFS over other filter-based techniques and automatic feature selection approaches. Moreover, the proposed algorithm maintained the accuracy achieved when utilizing all features, while using only $7\%$ to $30\%$ of the features. Consequently, this resulted in a reduction of the time needed for classifications, from $15\%$ to $70\%$.

Abstract: Causal inference in a sub-population involves identifying the causal effect of an intervention on a specific subgroup within a larger population. However, ignoring the subtleties introduced by sub-populations can either lead to erroneous inference or limit the applicability of existing methods. We introduce and advocate for a causal inference problem in sub-populations (henceforth called s-ID), in which we merely have access to observational data of the targeted sub-population (as opposed to the entire population). Existing inference problems in sub-populations operate on the premise that the given data distributions originate from the entire population, thus, cannot tackle the s-ID problem. To address this gap, we provide necessary and sufficient conditions that must hold in the causal graph for a causal effect in a sub-population to be identifiable from the observational distribution of that sub-population. Given these conditions, we present a sound and complete algorithm for the s-ID problem.

Abstract: Graph contrastive learning (GCL) has recently emerged as a promising approach for graph representation learning. Some existing methods adopt the 1-vs-K scheme to construct one positive and K negative samples for each graph, but it is difficult to set K. For those methods that do not use negative samples, it is often necessary to add additional strategies to avoid model collapse, which could only alleviate the problem to some extent. All these drawbacks will undoubtedly have an adverse impact on the generalizability and efficiency of the model. In this paper, to address these issues, we propose a novel graph self-contrast framework GraphSC, which only uses one positive and one negative sample, and chooses triplet loss as the objective. Specifically, self-contrast has two implications. First, GraphSC generates both positive and negative views of a graph sample from the graph itself via graph augmentation functions of various intensities, and use them for self-contrast. Second, GraphSC uses Hilbert-Schmidt Independence Criterion (HSIC) to factorize the representations into multiple factors and proposes a masked self-contrast mechanism to better separate positive and negative samples. Further, Since the triplet loss only optimizes the relative distance between the anchor and its positive/negative samples, it is difficult to ensure the absolute distance between the anchor and positive sample. Therefore, we explicitly reduced the absolute distance between the anchor and positive sample to accelerate convergence. Finally, we conduct extensive experiments to evaluate the performance of GraphSC against 19 other state-of-the-art methods in both unsupervised and transfer learning settings.

Abstract: Field-programmable gate arrays (FPGAs) are widely used to implement deep learning inference. Standard deep neural network inference involves the computation of interleaved linear maps and nonlinear activation functions. Prior work for ultra-low latency implementations has hardcoded the combination of linear maps and nonlinear activations inside FPGA lookup tables (LUTs). Our work is motivated by the idea that the LUTs in an FPGA can be used to implement a much greater variety of functions than this. In this paper, we propose a novel approach to training neural networks for FPGA deployment using multivariate polynomials as the basic building block. Our method takes advantage of the flexibility offered by the soft logic, hiding the polynomial evaluation inside the LUTs with zero overhead. We show that by using polynomial building blocks, we can achieve the same accuracy using considerably fewer layers of soft logic than by using linear functions, leading to significant latency and area improvements. We demonstrate the effectiveness of this approach in three tasks: network intrusion detection, jet identification at the CERN Large Hadron Collider, and handwritten digit recognition using the MNIST dataset.

Abstract: Physical systems can often be described via a continuous-time dynamical system. In practice, the true system is often unknown and has to be learned from measurement data. Since data is typically collected in discrete time, e.g. by sensors, most methods in Gaussian process (GP) dynamics model learning are trained on one-step ahead predictions. This can become problematic in several scenarios, e.g. if measurements are provided at irregularly-sampled time steps or physical system properties have to be conserved. Thus, we aim for a GP model of the true continuous-time dynamics. Higher-order numerical integrators provide the necessary tools to address this problem by discretizing the dynamics function with arbitrary accuracy. Many higher-order integrators require dynamics evaluations at intermediate time steps making exact GP inference intractable. In previous work, this problem is often tackled by approximating the GP posterior with variational inference. However, exact GP inference is preferable in many scenarios, e.g. due to its mathematical guarantees. In order to make direct inference tractable, we propose to leverage multistep and Taylor integrators. We demonstrate how to derive flexible inference schemes for these types of integrators. Further, we derive tailored sampling schemes that allow to draw consistent dynamics functions from the learned posterior. This is crucial to sample consistent predictions from the dynamics model. We demonstrate empirically and theoretically that our approach yields an accurate representation of the continuous-time system.

Abstract: One of the most surprising puzzles in neural network generalisation is grokking: a network with perfect training accuracy but poor generalisation will, upon further training, transition to perfect generalisation. We propose that grokking occurs when the task admits a generalising solution and a memorising solution, where the generalising solution is slower to learn but more efficient, producing larger logits with the same parameter norm. We hypothesise that memorising circuits become more inefficient with larger training datasets while generalising circuits do not, suggesting there is a critical dataset size at which memorisation and generalisation are equally efficient. We make and confirm four novel predictions about grokking, providing significant evidence in favour of our explanation. Most strikingly, we demonstrate two novel and surprising behaviours: ungrokking, in which a network regresses from perfect to low test accuracy, and semi-grokking, in which a network shows delayed generalisation to partial rather than perfect test accuracy.

Abstract: In this paper, we present Delta-LoRA, which is a novel parameter-efficient approach to fine-tune large language models (LLMs). In contrast to LoRA and other low-rank adaptation methods such as AdaLoRA, Delta-LoRA not only updates the low-rank matrices $\bA$ and $\bB$, but also propagate the learning to the pre-trained weights $\bW$ via updates utilizing the delta of the product of two low-rank matrices ($\bA^{(t+1)}\bB^{(t+1)} - \bA^{(t)}\bB^{(t)}$). Such a strategy effectively addresses the limitation that the incremental update of low-rank matrices is inadequate for learning representations capable for downstream tasks. Moreover, as the update of $\bW$ does not need to compute the gradients of $\bW$ and store their momentums, Delta-LoRA shares comparable memory requirements and computational costs with LoRA. Extensive experiments show that Delta-LoRA significantly outperforms existing low-rank adaptation methods. We further support these results with comprehensive analyses that underscore the effectiveness of Delta-LoRA.

Abstract: In online ranking, a learning algorithm sequentially ranks a set of items and receives feedback on its ranking in the form of relevance scores. Since obtaining relevance scores typically involves human annotation, it is of great interest to consider a partial feedback setting where feedback is restricted to the top-$k$ items in the rankings. Chaudhuri and Tewari [2017] developed a framework to analyze online ranking algorithms with top $k$ feedback. A key element in their work was the use of techniques from partial monitoring. In this paper, we further investigate online ranking with top $k$ feedback and solve some open problems posed by Chaudhuri and Tewari [2017]. We provide a full characterization of minimax regret rates with the top $k$ feedback model for all $k$ and for the following ranking performance measures: Pairwise Loss, Discounted Cumulative Gain, and Precision@n. In addition, we give an efficient algorithm that achieves the minimax regret rate for Precision@n.

Abstract: As the size and complexity of data continue to increase, the need for efficient and effective analysis methods becomes ever more crucial. Tensorization, the process of converting 2-dimensional datasets into multidimensional structures, has emerged as a promising approach for multiway analysis methods. This paper explores the steps involved in tensorization, multidimensional data sources, various multiway analysis methods employed, and the benefits of these approaches. A small example of Blind Source Separation (BSS) is presented comparing 2-dimensional algorithms and a multiway algorithm in Python. Results indicate that multiway analysis is more expressive. Additionally, tensorization techniques aid in compressing deep learning models by reducing the number of required parameters while enhancing the expression of relationships across dimensions. A survey of the multi-away analysis methods and integration with various Deep Neural Networks models is presented using case studies in different domains.

Abstract: Agents that are aware of the separation between themselves and their environments can leverage this understanding to form effective representations of visual input. We propose an approach for learning such structured representations for RL algorithms, using visual knowledge of the agent, such as its shape or mask, which is often inexpensive to obtain. This is incorporated into the RL objective using a simple auxiliary loss. We show that our method, Structured Environment-Agent Representations, outperforms state-of-the-art model-free approaches over 18 different challenging visual simulation environments spanning 5 different robots. Website at https://sear-rl.github.io/