We consider the nonparametric regression and the classification problems for $\psi$-weakly dependent processes. This weak dependence structure is more general than conditions such as, mixing, association, $\ldots$. A penalized estimation method for sparse deep neural networks is performed. In both nonparametric regression and binary classification problems, we establish oracle inequalities for the excess risk of the sparse-penalized deep neural networks estimators. Convergence rates of the excess risk of these estimators are also derived. The simulation results displayed show that, the proposed estimators overall work well than the non penalized estimators.
The image deepfake detection task has been greatly addressed by the scientific community to discriminate real images from those generated by Artificial Intelligence (AI) models: a binary classification task. In this work, the deepfake detection and recognition task was investigated by collecting a dedicated dataset of pristine images and fake ones generated by 9 different Generative Adversarial Network (GAN) architectures and by 4 additional Diffusion Models (DM). A hierarchical multi-level approach was then introduced to solve three different deepfake detection and recognition tasks: (i) Real Vs AI generated; (ii) GANs Vs DMs; (iii) AI specific architecture recognition. Experimental results demonstrated, in each case, more than 97% classification accuracy, outperforming state-of-the-art methods.
Click-based interactive segmentation (IS) aims to extract the target objects under user interaction. For this task, most of the current deep learning (DL)-based methods mainly follow the general pipelines of semantic segmentation. Albeit achieving promising performance, they do not fully and explicitly utilize and propagate the click information, inevitably leading to unsatisfactory segmentation results, even at clicked points. Against this issue, in this paper, we propose to formulate the IS task as a Gaussian process (GP)-based pixel-wise binary classification model on each image. To solve this model, we utilize amortized variational inference to approximate the intractable GP posterior in a data-driven manner and then decouple the approximated GP posterior into double space forms for efficient sampling with linear complexity. Then, we correspondingly construct a GP classification framework, named GPCIS, which is integrated with the deep kernel learning mechanism for more flexibility. The main specificities of the proposed GPCIS lie in: 1) Under the explicit guidance of the derived GP posterior, the information contained in clicks can be finely propagated to the entire image and then boost the segmentation; 2) The accuracy of predictions at clicks has good theoretical support. These merits of GPCIS as well as its good generality and high efficiency are substantiated by comprehensive experiments on several benchmarks, as compared with representative methods both quantitatively and qualitatively.
\textit{Complementary label learning} (CLL) requires annotators to give \emph{irrelevant} labels instead of relevant labels for instances. Currently, CLL has shown its promising performance on multi-class data by estimating a transition matrix. However, current multi-class CLL techniques cannot work well on multi-labeled data since they assume each instance is associated with one label while each multi-labeled instance is relevant to multiple labels. Here, we show theoretically how the estimated transition matrix in multi-class CLL could be distorted in multi-labeled cases as they ignore co-existing relevant labels. Moreover, theoretical findings reveal that calculating a transition matrix from label correlations in \textit{multi-labeled CLL} (ML-CLL) needs multi-labeled data, while this is unavailable for ML-CLL. To solve this issue, we propose a two-step method to estimate the transition matrix from candidate labels. Specifically, we first estimate an initial transition matrix by decomposing the multi-label problem into a series of binary classification problems, then the initial transition matrix is corrected by label correlations to enforce the addition of relationships among labels. We further show that the proposal is classifier-consistent, and additionally introduce an MSE-based regularizer to alleviate the tendency of BCE loss overfitting to noises. Experimental results have demonstrated the effectiveness of the proposed method.
We study the problem of online binary classification where strategic agents can manipulate their observable features in predefined ways, modeled by a manipulation graph, in order to receive a positive classification. We show this setting differs in fundamental ways from non-strategic online classification. For instance, whereas in the non-strategic case, a mistake bound of $\ln|H|$ is achievable via the halving algorithm when the target function belongs to a known class $H$, we show that no deterministic algorithm can achieve a mistake bound $o(\Delta)$ in the strategic setting, where $\Delta$ is the maximum degree of the manipulation graph (even when $|H|=O(\Delta)$). We obtain an algorithm achieving mistake bound $O(\Delta\ln|H|)$. We also extend this to the agnostic setting and obtain an algorithm with a $\Delta$ multiplicative regret, and we show no deterministic algorithm can achieve $o(\Delta)$ multiplicative regret. Next, we study two randomized models based on whether the random choices are made before or after agents respond, and show they exhibit fundamental differences. In the first model, at each round the learner deterministically chooses a probability distribution over classifiers inducing expected values on each vertex (probabilities of being classified as positive), which the strategic agents respond to. We show that any learner in this model has to suffer linear regret. On the other hand, in the second model, while the adversary who selects the next agent must respond to the learner's probability distribution over classifiers, the agent then responds to the actual hypothesis classifier drawn from this distribution. Surprisingly, we show this model is more advantageous to the learner, and we design randomized algorithms that achieve sublinear regret bounds against both oblivious and adaptive adversaries.
Recent studies demonstrated that the adversarially robust learning under $\ell_\infty$ attack is harder to generalize to different domains than standard domain adaptation. How to transfer robustness across different domains has been a key question in domain adaptation field. To investigate the fundamental difficulty behind adversarially robust domain adaptation (or robustness transfer), we propose to analyze a key complexity measure that controls the cross-domain generalization: the adversarial Rademacher complexity over {\em symmetric difference hypothesis space} $\mathcal{H} \Delta \mathcal{H}$. For linear models, we show that adversarial version of this complexity is always greater than the non-adversarial one, which reveals the intrinsic hardness of adversarially robust domain adaptation. We also establish upper bounds on this complexity measure. Then we extend them to the ReLU neural network class by upper bounding the adversarial Rademacher complexity in the binary classification setting. Finally, even though the robust domain adaptation is provably harder, we do find positive relation between robust learning and standard domain adaptation. We explain \emph{how adversarial training helps domain adaptation in terms of standard risk}. We believe our results initiate the study of the generalization theory of adversarially robust domain adaptation, and could shed lights on distributed adversarially robust learning from heterogeneous sources, e.g., federated learning scenario.
This paper compares the performance of various data processing methods in terms of predictive performance for structured data. This paper also seeks to identify and recommend preprocessing methodologies for tree-based binary classification models, with a focus on eXtreme Gradient Boosting (XGBoost) models. Three data sets of various structures, interactions, and complexity were constructed, which were supplemented by a real-world data set from the Lending Club. We compare several methods for feature selection, categorical handling, and null imputation. Performance is assessed using relative comparisons among the chosen methodologies, including model prediction variability. This paper is presented by the three groups of preprocessing methodologies, with each section consisting of generalized observations. Each observation is accompanied by a recommendation of one or more preferred methodologies. Among feature selection methods, permutation-based feature importance, regularization, and XGBoost's feature importance by weight are not recommended. The correlation coefficient reduction also shows inferior performance. Instead, XGBoost importance by gain shows the most consistency and highest caliber of performance. Categorical featuring encoding methods show greater discrimination in performance among data set structures. While there was no universal "best" method, frequency encoding showed the greatest performance for the most complex data sets (Lending Club), but had the poorest performance for all synthetic (i.e., simpler) data sets. Finally, missing indicator imputation dominated in terms of performance among imputation methods, whereas tree imputation showed extremely poor and highly variable model performance.
How can one meaningfully make a measurement, if the meter does not conform to any standard and its scale expands or shrinks depending on what is measured? In the present work it is argued that current evaluation practices for machine-learning classifiers are affected by this kind of problem, leading to negative consequences when classifiers are put to real use; consequences that could have been avoided. It is proposed that evaluation be grounded on Decision Theory, and the implications of such foundation are explored. The main result is that every evaluation metric must be a linear combination of confusion-matrix elements, with coefficients - "utilities" - that depend on the specific classification problem. For binary classification, the space of such possible metrics is effectively two-dimensional. It is shown that popular metrics such as precision, balanced accuracy, Matthews Correlation Coefficient, Fowlkes-Mallows index, F1-measure, and Area Under the Curve are never optimal: they always give rise to an in-principle avoidable fraction of incorrect evaluations. This fraction is even larger than would be caused by the use of a decision-theoretic metric with moderately wrong coefficients.
Usually standard algorithms employ a loss where each error is the mere absolute difference between the true value and the prediction, in case of a regression task. In the present, we introduce several error weighting schemes that are a generalization of the consolidated routine. We study both a binary classification model for Support Vector Classification and a regression net for Multi-layer Perceptron. Results proves that the error is never worse than the standard procedure and several times it is better.
Discovering interpretable patterns for classification of sequential data is of key importance for a variety of fields, ranging from genomics to fraud detection or more generally interpretable decision-making. In this paper, we propose a novel differentiable fully interpretable method to discover both local and global patterns (i.e. catching a relative or absolute temporal dependency) for rule-based binary classification. It consists of a convolutional binary neural network with an interpretable neural filter and a training strategy based on dynamically-enforced sparsity. We demonstrate the validity and usefulness of the approach on synthetic datasets and on an open-source peptides dataset. Key to this end-to-end differentiable method is that the expressive patterns used in the rules are learned alongside the rules themselves.
Receiver Operating Characteristic (ROC) curves are plots of true positive rate versus false positive rate which are used to evaluate binary classification algorithms. Because the Area Under the Curve (AUC) is a constant function of the predicted values, learning algorithms instead optimize convex relaxations which involve a sum over all pairs of labeled positive and negative examples. Naive learning algorithms compute the gradient in quadratic time, which is too slow for learning using large batch sizes. We propose a new functional representation of the square loss and squared hinge loss, which results in algorithms that compute the gradient in either linear or log-linear time, and makes it possible to use gradient descent learning with large batch sizes. In our empirical study of supervised binary classification problems, we show that our new algorithm can achieve higher test AUC values on imbalanced data sets than previous algorithms, and make use of larger batch sizes than were previously feasible.
Deep learning provides an excellent avenue for optimizing diagnosis and patient monitoring for clinical-based applications, which can critically enhance the response time to the onset of various conditions. For cardiovascular disease, one such condition where the rising number of patients increasingly outweighs the availability of medical resources in different parts of the world, a core challenge is the automated classification of various cardiac abnormalities. Existing deep learning approaches have largely been limited to detecting the existence of an irregularity, as in binary classification, which has been achieved using networks such as CNNs and RNN/LSTMs. The next step is to accurately perform multi-class classification and determine the specific condition(s) from the inherently noisy multi-variate waveform, which is a difficult task that could benefit from (1) a more powerful sequential network, and (2) the integration of clinical notes, which provide valuable semantic and clinical context from human doctors. Recently, Transformers have emerged as the state-of-the-art architecture for forecasting and prediction using time-series data, with their multi-headed attention mechanism, and ability to process whole sequences and learn both long and short-range dependencies. The proposed novel multi-modal Transformer architecture would be able to accurately perform this task while demonstrating the cross-domain effectiveness of Transformers, establishing a method for incorporating multiple data modalities within a Transformer for classification tasks, and laying the groundwork for automating real-time patient condition monitoring in clinical and ER settings.
Despite being highly over-parametrized, and having the ability to fully interpolate the training data, deep networks are known to generalize well to unseen data. It is now understood that part of the reason for this is that the training algorithms used have certain implicit regularization properties that ensure interpolating solutions with "good" properties are found. This is best understood in linear over-parametrized models where it has been shown that the celebrated stochastic gradient descent (SGD) algorithm finds an interpolating solution that is closest in Euclidean distance to the initial weight vector. Different regularizers, replacing Euclidean distance with Bregman divergence, can be obtained if we replace SGD with stochastic mirror descent (SMD). Empirical observations have shown that in the deep network setting, SMD achieves a generalization performance that is different from that of SGD (and which depends on the choice of SMD's potential function. In an attempt to begin to understand this behavior, we obtain the generalization error of SMD for over-parametrized linear models for a binary classification problem where the two classes are drawn from a Gaussian mixture model. We present simulation results that validate the theory and, in particular, introduce two data models, one for which SMD with an $\ell_2$ regularizer (i.e., SGD) outperforms SMD with an $\ell_1$ regularizer, and one for which the reverse happens.
The face anti-spoofing (FAS) method performs well under intra-domain setups. However, its cross-domain performance is unsatisfactory. As a result, the domain generalization (DG) method has gained more attention in FAS. Existing methods treat FAS as a simple binary classification task and propose a heuristic training objective to learn domain-invariant features. However, there is no theoretical explanation of what a domain-invariant feature is. Additionally, the lack of theoretical support makes domain generalization techniques such as adversarial training lack training stability. To address these issues, this paper proposes the EnfoMax framework, which uses information theory to analyze cross-domain FAS tasks. This framework provides theoretical guarantees and optimization objectives for domain-generalized FAS tasks. EnfoMax maximizes the domain entropy and mutual information of live samples in source domains without using adversarial learning. Experimental results demonstrate that our approach performs well on extensive public datasets and outperforms state-of-the-art methods.
Federated learning aims to collaboratively train models without accessing their client's local private data. The data may be Non-IID for different clients and thus resulting in poor performance. Recently, personalized federated learning (PFL) has achieved great success in handling Non-IID data by enforcing regularization in local optimization or improving the model aggregation scheme on the server. However, most of the PFL approaches do not take into account the unfair competition issue caused by the imbalanced data distribution and lack of positive samples for some classes in each client. To address this issue, we propose a novel and generic PFL framework termed Federated Averaging via Binary Classification, dubbed FedABC. In particular, we adopt the ``one-vs-all'' training strategy in each client to alleviate the unfair competition between classes by constructing a personalized binary classification problem for each class. This may aggravate the class imbalance challenge and thus a novel personalized binary classification loss that incorporates both the under-sampling and hard sample mining strategies is designed. Extensive experiments are conducted on two popular datasets under different settings, and the results demonstrate that our FedABC can significantly outperform the existing counterparts.
Arunachalam and de Wolf (2018) showed that the sample complexity of quantum batch learning of boolean functions, in the realizable and agnostic settings, has the same form and order as the corresponding classical sample complexities. In this paper, we extend this, ostensibly surprising, message to batch multiclass learning, online boolean learning, and online multiclass learning. For our online learning results, we first consider an adaptive adversary variant of the classical model of Dawid and Tewari (2022). Then, we introduce the first (to the best of our knowledge) model of online learning with quantum examples.
A recent line of works apply machine learning techniques to assist or rebuild cost-based query optimizers in DBMS. While exhibiting superiority in some benchmarks, their deficiencies, e.g., unstable performance, high training cost, and slow model updating, stem from the inherent hardness of predicting the cost or latency of execution plans using machine learning models. In this paper, we introduce a learning-to-rank query optimizer, called Lero, which builds on top of a native query optimizer and continuously learns to improve the optimization performance. The key observation is that the relative order or rank of plans, rather than the exact cost or latency, is sufficient for query optimization. Lero employs a pairwise approach to train a classifier to compare any two plans and tell which one is better. Such a binary classification task is much easier than the regression task to predict the cost or latency, in terms of model efficiency and accuracy. Rather than building a learned optimizer from scratch, Lero is designed to leverage decades of wisdom of databases and improve the native query optimizer. With its non-intrusive design, Lero can be implemented on top of any existing DBMS with minimal integration efforts. We implement Lero and demonstrate its outstanding performance using PostgreSQL. In our experiments, Lero achieves near optimal performance on several benchmarks. It reduces the plan execution time of the native optimizer in PostgreSQL by up to 70% and other learned query optimizers by up to 37%. Meanwhile, Lero continuously learns and automatically adapts to query workloads and changes in data.
Error-correcting codes (ECC) are used to reduce multiclass classification tasks to multiple binary classification subproblems. In ECC, classes are represented by the rows of a binary matrix, corresponding to codewords in a codebook. Codebooks are commonly either predefined or problem dependent. Given predefined codebooks, codeword-to-class assignments are traditionally overlooked, and codewords are implicitly assigned to classes arbitrarily. Our paper shows that these assignments play a major role in the performance of ECC. Specifically, we examine similarity-preserving assignments, where similar codewords are assigned to similar classes. Addressing a controversy in existing literature, our extensive experiments confirm that similarity-preserving assignments induce easier subproblems and are superior to other assignment policies in terms of their generalization performance. We find that similarity-preserving assignments make predefined codebooks become problem-dependent, without altering other favorable codebook properties. Finally, we show that our findings can improve predefined codebooks dedicated to extreme classification.
Tool wear monitoring is crucial for quality control and cost reduction in manufacturing processes, of which drilling applications are one example. In this paper, we present a U-Net based semantic image segmentation pipeline, deployed on microscopy images of cutting inserts, for the purpose of wear detection. The wear area is differentiated in two different types, resulting in a multiclass classification problem. Joining the two wear types in one general wear class, on the other hand, allows the problem to be formulated as a binary classification task. Apart from the comparison of the binary and multiclass problem, also different loss functions, i. e., Cross Entropy, Focal Cross Entropy, and a loss based on the Intersection over Union (IoU), are investigated. Furthermore, models are trained on image tiles of different sizes, and augmentation techniques of varying intensities are deployed. We find, that the best performing models are binary models, trained on data with moderate augmentation and an IoU-based loss function.
Kernel methods are an important class of techniques in machine learning. To be effective, good feature maps are crucial for mapping non-linearly separable input data into a higher dimensional (feature) space, thus allowing the data to be linearly separable in feature space. Previous work has shown that quantum feature map design can be automated for a given dataset using NSGA-II, a genetic algorithm, while both minimizing circuit size and maximizing classification accuracy. However, the evaluation of the accuracy achieved by a candidate feature map is costly. In this work, we demonstrate the suitability of kernel-target alignment as a substitute for accuracy in genetic algorithm-based quantum feature map design. Kernel-target alignment is faster to evaluate than accuracy and doesn't require some data points to be reserved for its evaluation. To further accelerate the evaluation of genetic fitness, we provide a method to approximate kernel-target alignment. To improve kernel-target alignment and root mean squared error, the final trainable parameters of the generated circuits are further trained using COBYLA to determine whether a hybrid approach applying conventional circuit parameter training can easily complement the genetic structure optimization approach. A total of eight new approaches are compared to the original across nine varied binary classification problems from the UCI machine learning repository, showing that kernel-target alignment and its approximation produce feature map circuits enabling comparable accuracy to the previous work but with larger margins on training data (in excess of 20\% larger) that improve further with circuit parameter training.