Multimodal Fusion Learning ( MFL ), leveraging disparate data from various imaging modalities (e.g., MRI, CT, SPECT ), has shown great potential for addressing medical problems such as skin cancer and brain tumor prediction. However, existing MFL methods face three key limitations: a) they often specialize in specific modalities, and overlooks effective shared complementary information across diverse modalities, hence limiting their generalizability for multi-disease analysis; b) they rely on computationally expensive models, restricting their applicability in resource-limited settings; and c) they lack robustness against adversarial attacks, compromising reliability in medical AI applications. To address these limitations, we propose a novel Multi-Attention Integration Learning ( MAIL ) network, incorporating two key components: a) an efficient residual learning attention block for capturing refined modality-specific multi-scale patterns and b) an efficient multimodal cross-attention module for learning enriched complementary shared representations across diverse modalities. Furthermore, to ensure adversarial robustness, we extend MAIL network to design Robust-MAIL by incorporating random projection filters and modulated attention noise. Extensive evaluations on 20 public datasets show that both MAIL and Robust-MAIL outperform existing methods, achieving performance gains of up to 9.34% while reducing computational costs by up to 78.3%. These results highlight the superiority of our approaches, ensuring more reliable predictions than top competitors.

Multimodal fusion frameworks, which integrate diverse medical imaging modalities (e.g., MRI, CT), have shown great potential in applications such as skin cancer detection, dementia diagnosis, and brain tumor prediction. However, existing multimodal fusion methods face significant challenges. First, they often rely on computationally expensive models, limiting their applicability in low-resource environments. Second, they often employ cascaded attention modules, which potentially increase risk of information loss during inter-module transitions and hinder their capacity to effectively capture robust shared representations across modalities. This restricts their generalization in multi-disease analysis tasks. To address these limitations, we propose a Hybrid Parallel-Fusion Cascaded Attention Network (HyPCA-Net), composed of two core novel blocks: (a) a computationally efficient residual adap tive learning attention block for capturing refined modality-specific representations, and (b) a dual-view cascaded attention block aimed at learning robust shared representations across diverse modalities. Extensive experiments on ten publicly available datasets exhibit that HyPCA-Netsignificantly outperforms existing leading methods, with improvements of up to 5.2% in performance and reductions of up to 73.1% in computational cost. Code: https://github.com/misti1203/HyPCA-Net.

Multimodal fusion learning has shown significant promise in classifying various diseases such as skin cancer and brain tumors. However, existing methods face three key limitations. First, they often lack generalizability to other diagnosis tasks due to their focus on a particular disease. Second, they do not fully leverage multiple health records from diverse modalities to learn robust complementary information. And finally, they typically rely on a single attention mechanism, missing the benefits of multiple attention strategies within and across various modalities. To address these issues, this paper proposes a dual robust information fusion attention mechanism (DRIFA) that leverages two attention modules - i.e., multi-branch fusion attention module and the multimodal information fusion attention module. DRIFA can be integrated with any deep neural network, forming a multimodal fusion learning framework denoted as DRIFA-Net. We show that the multi-branch fusion attention of DRIFA learns enhanced representations for each modality, such as dermoscopy, pap smear, MRI, and CT-scan, whereas multimodal information fusion attention module learns more refined multimodal shared representations - improving the network's generalization across multiple tasks and enhancing overall performance. Additionally, to estimate the uncertainty of DRIFA-Net predictions, we have employed an ensemble Monte Carlo dropout strategy. Extensive experiments on five publicly available datasets with diverse modalities demonstrate that our approach consistently outperforms state-of-the-art methods. The code is available at https://github.com/misti1203/DRIFA-Net.

Recent advances in vision-language models have resulted in improved accuracy in visual question answering (VQA) tasks. However, their robustness remains limited when faced with out-of-distribution data containing unanswerable questions. In this study, we first construct a simple randomised VQA dataset, incorporating unanswerable questions from the VQA v2 dataset, to evaluate the robustness of a state-of-the-art VQA model. Our findings reveal that the model struggles to predict the “unknown” answer or provides inaccurate responses with high confidence scores for irrelevant questions. To address this issue without retraining the large backbone models, we propose Cross Modal Augmentation (CMA), a model-agnostic, test-time-only, multi-modal semantic augmentation technique. CMA generates multiple semantically-consistent but heterogeneous instances from the visual and textual inputs, which are then fed to the model, and the predictions are combined to achieve a more robust output. We demonstrate that implementing CMA enables the VQA model to provide more reliable answers in scenarios involving unanswerable questions, and show that the approach is generalisable across different categories of pre-trained vision language models.

Abstract Artificial neural networks ( $${{{\texttt {ANN}}}}$$ ANN ) are widely used machine learning models. Their widespread use has attracted a lot of interest in their robustness. Many studies show that ’s performance can be highly vulnerable to input manipulation such as adversarial attacks and covariate drift. Therefore, various techniques that focus on improving $${{{\texttt {ANN}}}}$$ ANN ’s robustness have been proposed in the last few years. However, most of these works have mostly focused on image data. In this paper, we investigate the role of discretization in improving $${{{\texttt {ANN}}}}$$ ANN ’s robustness on tabular datasets. Two custom $${{{\texttt {ANN}}}}$$ ANN layers– and (collectively called ) are proposed. The two layers integrate discretization during the training phase to improve $${{{\texttt {ANN}}}}$$ ANN ’s ability to defend against adversarial attacks. Additionally, integrates dynamic discretization during testing phase as well, to provide a unified strategy to handle adversarial attacks and covariate drift. The experimental results on 24 publicly available datasets show that our proposed add much-needed robustness to $${{{\texttt {ANN}}}}$$ ANN for tabular datasets.

Abstract Large scale categorical datasets are ubiquitous in machine learning and the success of most deployed machine learning models rely on how effectively the features are engineered. For large-scale datasets, parametric methods are generally used, among which three strategies for feature engineering are quite common. The first strategy focuses on managing the breadth (or width) of a network, e.g., generalized linear models (aka. ). The second strategy focuses on the depth of a network, e.g., Artificial Neural networks or (aka. ). The third strategy relies on factorizing the interaction terms, e.g., Factorization Machines (aka. ). Each of these strategies brings its own advantages and disadvantages. Recently, it has been shown that for categorical data, combination of various strategies leads to excellent results. For example, -Learning, , etc., leads to state-of-the-art results. Following the trend, in this work, we have proposed another learning framework—-Learning, based on the combination of , , , and a newly introduced component named (). is in the form of a Bayesian network classifier whose structure is learned apriori, and parameters are learned by optimizing a joint objective function along with , and parts. We denote the learning of parameters as . Additionally, the parameters of are constrained to be actual probabilities—therefore, it is extremely interpretable. Furthermore, one can sample or generate data from , which can facilitate learning and provides a framework for knowledge-guided machine learning. We demonstrate that our proposed framework possesses the resilience to maintain excellent classification performance when confronted with biased datasets. We evaluate the efficacy of our framework in terms of classification performance on various benchmark large-scale categorical datasets and compare against state-of-the-art methods. It is shown that, framework (a) exhibits superior performance on classification tasks, (b) boasts outstanding interpretability and (c) demonstrates exceptional resilience and effectiveness in scenarios involving skewed distributions.

Abstract Graph representation learning (graph embedding) has led to breakthrough results in various machine learning graph-based applications such as node classification, link prediction and recommendation. Many real-world graphs can be characterized as the property graphs, because besides the structure information, there exists rich property information related to each node in the graphs. Many existing graph representation learning methods—e.g. random walk-based methods like and , focus only on the structure of graph for learning the node embedding. Although graph representation learning based on neural networks (e.g. typical methods such as ) uses the property of nodes as the initial features of nodes and then aggregates feature information of the neighbours, their limitation is that the neighbourhood of a node is considered to be uniform—i.e. there is no way to differentiate among neighbours of a node when learning a node embedding. Additionally, their definition of neighbourhood is local, i.e. only nodes connected to the current node are considered as neighbours. Hence, those methods fail to capture implicit/latent relationships among nodes, which are implicit in the given structure. In this study, our aim is to improve the performance of graph representation learning methods on property graphs. We present a new framework called ()—a graph representation learning framework to address above-mentioned limitations. Our proposed framework relies on the notion of latent neighbourhood, as well as systematic sampling of neighbouring nodes to obtain better representation of the nodes. The experimental results on five publicly available graph datasets demonstrate that outperforms state-of-the-art baselines for the task of node classification. We further evaluate the superiority of our proposed formulation by defining a novel quantitative metric to measure the usefulness of the sampled neighbourhood in the graph.

Tabular data generation has seen renewed interest with the advent of Generative Adversarial Networks (GAN). Recently, it has been shown that one can use a Bayesian network as either a generator or a discriminator in the GAN framework, resulting in an algorithm known as GANBLR. It has been shown that GANBLR gives state of the art results for tabular data generation. However, the model has one limitation. It uses class attributes during model training. For example, a supervised Bayesian network is needed as a generator at training time. This makes GANBLR inapplicable for cases where we do not have access to class information. Addressing this shortcoming of GANBLR has been the main motivation of this work. In this work, we have proposed a new model of tabular data generation – Masked Ensemble Tabular Generator (MEG), which does not require class labels to generate tabular data. The proposed models rely on a novel strategy of using a collection of Bayesian networks as part of the generator, and relies on masking operations to train the generator efficiently. It also uses a group-based similarity measure to adjust the number of samples generated from each Bayesian network in the collection. We perform extensive experiments on a variety of datasets and demonstrate that MEG not only outperforms baselines that do not have class information during training, such as CTGAN and TVAE, but also outperforms baselines that provide access to class information during training, such as TableGAN and CtabGANmethods. It has almost similar performance in terms of machine learning utility to GANBLR, and of course is greatly advantaged by being truly unsupervised in nature. We highlight this by demonstrating its applicability to a clustering task. We also investigate the privacy preserving capabilities of MEG and demonstrate its superior performance compared to other baselines.

Abstract Generative adversarial network () models have been successfully utilized in a wide range of machine learning applications, and tabular data generation domain is not an exception. Notably, some state-of-the-art models of tabular data generation, such as , , , etc. are based on models. Even though these models have resulted in superior performance in generating artificial data when trained on a range of datasets, there is a lot of room (and desire) for improvement. Not to mention that existing methods do have some weaknesses other than performance. For example, the current methods focus only on the performance of the model, and limited emphasis is given on the interpretation of the model. Secondly, the current models operate on raw features only, and hence they fail to exploit any prior knowledge on explicit feature interactions that can be utilized during data generation process. To alleviate the two above-mentioned limitations, in this work, we propose a novel tabular data generation model— G enerative A dversarial Network modelling inspired from N aive B ayes and L ogistic R egression’s relationship ( $${ { \texttt {GANBLR} } }$$ GANBLR ), which not only address the interpretation limitation of existing tabular -based models but provides capability to handle explicit feature interactions as well. Through extensive evaluations on wide range of datasets, we demonstrate $${ { \texttt {GANBLR} } }$$ GANBLR ’s superior performance as well as better interpretable capability (explanation of feature importance in the synthetic generation process) as compared to existing state-of-the-art tabular data generation models.

Evaluating human dialogue is a complex task, as our conversation are never structured. There are, however, cases where there is some structure in our conversation, e.g., in a typical call center, dialogue between a call center agent and customer revolves around certain topics of conversation. These dialogues can be evaluated based on some pre-specified criteria as well as sub-criteria. This evaluation is typically done manually, which can be be time-consuming, motivating the need for an automated system that employs Artificial Intelligence (AI) algorithms to evaluate dialogues efficiently. In this paper, we have proposed a novel dialogue-evaluation framework that leverages recent advancements in deep learning research. The contributions of this work are two fold. Firstly, we introduce a straightforward end-to-end framework – CallAI, for evaluating dialogues in any domain, based on some predefined hierarchical criteria. Secondly, we present a novel algorithm – TAABLM, utilizing a novel combination of aspect-based learning along with traditional TF-IDF features for text. We show in this paper, that TAABLM outperforms conventional baselines such as BERT, LSTM, etc, delivering improved performance in automated dialogue evaluation, whereas CallAI offers a simple yet elegant framework for an AI-based solution to hierarchical dialogue evaluation. We demonstrate the efficacy of our proposed framework and proposed algorithm on three datasets, where we quantify performance in terms of an aggregated dialogue-score as well as in terms of either accuracy or AuROC metrics.

Generative Adversarial Networks (GAN) models have led to a major breakthrough in data generation of various sorts. Over the years, we have seen several applications of GAN-based learning for tabular data generation as well. Very recently, GAN-based learning by incorporating Bayesian Networks (BN) as generator and discriminator – GANBLR, has shown to lead to state-of-the-art (SOTA) results for tabular data generation. Despite the impressive performance, GANBLR has an inherent weakness that it can only generate data with categorical attributes. Additionally, the model is trained and tested only with a restricted Bayesian Network. In this work, we have proposed an extension over GANBLR framework – GANBLR++, that has the capacity to generate numeric attributes, by leveraging Dirichlet Mixture Model. We also leverage unrestricted BN in GANBLR framework, and discuss how the use of unrestricted BN can lead to better quality data, as well as more interpretable model. We evaluate the effectiveness of GANBLR++ on wide range of datasets by demonstrating that it generates data of better quality as compared to existing SOTA models for tabular (numeric and categorical) data generation such as CTGAN, MedGAN and TableGAN.

Generative Adversarial Network (GAN) models have shown to be effective in a wide range of machine learning applications, and tabular data generation process has not been an exception. Notably, some state-of-the-art models of tabular data generation, such as CTGAN, TableGan, MedGAN, etc. are based on GAN models. Even though these models have resulted in superiour performance in generating artificial data when trained on a range of datasets, there is a lot of room (and desire) for improvement. Not to mention that existing methods do have some weaknesses other than performance. E.g., the current methods focus only on the performance of the model, and limited emphasis is given to the interpretation of the model. Secondly, the current models operate on raw features only, and hence they fail to exploit any prior knowledge on explicit feature interactions that can be utilized during data generation process. To alleviate the two above-mentioned limitations, in this work, we propose a novel tabular data generation model– Generative Adversarial Network modelling inspired from Naive Bayes and Logistic Regression’s relationship (GANBLR), which can not only address the interpretation limitation in existing tabular GAN-based models but can provide capability to handle explicit feature interactions. By extensively evaluating on wide range of datasets, we demonstrate GANBLR’S superiour performance as well as better interpretable capability (explanation of feature importance in the synthetic generation process) as compared to existing state-of-the-art tabular data generation models.

Deep neural networks tend to underestimate uncertainty and produce overly confident predictions. Recently proposed solutions, such as MC Dropout and SDENet, require complex training and/or auxiliary out-of-distribution data. We propose a simple solution by extending the time-tested iterative reweighted least square (IRLS) in generalised linear regression. We use two sub-networks to parametrise the prediction and uncertainty estimation, enabling easy handling of complex inputs and nonlinear response. The two sub-networks have shared representations and are trained via two complementary loss functions for the prediction and the uncertainty estimates, with interleaving steps as in a cooperative game. Compared with more complex models such as MC-Dropout or SDE-Net, our proposed network is simpler to implement and more robust (insensitive to varying aleatoric and epistemic uncertainty).

Machine learning researchers are facing a data deluge quantities of training data have been increasing at a rapid rate. However, most of machine learning algorithms were proposed in the context of learning from relatively smaller quantities of data. We argue that a big data classifier should have superior feature engineering capability, minimal tuning parameters and should be able to learn decision boundaries in fewer passes through the data. In this paper, we have proposed an (computationally) efficient yet (classification-wise) effective family of learning algorithms that fulfils these properties. The proposed family of learning algorithms is based on recently proposed accelerated higher-order logistic regression algorithm: ALRn. The contributions of this work are three-fold. First, we have added the functionality of out-of-core learning in ALRn, resulting in a limited pass learning algorithm. Second, superior feature engineering capabilities are built and third, a far more efficient (memory-wise) implementation has been proposed. We demonstrate the competitiveness of our proposed algorithm by comparing its performance not only with state-of-the-art classifier in out-of-core learning such as Selective KDB but also with state-of-the-art in in-core learning such as Random Forest.

With the emergence of big data, there has been a growing interest in optimization routines that lead to faster convergence of Logistic Regression (LR). Among many optimization methods such as Gradient Descent, Quasi-Newton, Conjugate Gradient, etc., the Trust-region based truncated Newton method (TRON) algorithm has been shown to converge the fastest. The TRON algorithm also forms an important component of the highly efficient and widely used liblinear package. It has been shown that the WANBIA-C trick of scaling with the log of the naive Bayes conditional probabilities can greatly accelerate the convergence of LR trained using (first-order) Gradient Descent and (approximate second-order) Quasi-Newton optimization. In this work we study the applicability of the WANBIA-C trick to TRON. We first devise a TRON algorithm optimizing the softmax objective function and then demonstrate that WANBIA-C style preconditioning can be beneficial for TRON, leading to an extremely fast (batch) LR algorithm. Second, we present a comparative analysis of one-vs-all LR and softmax LR in terms of the 0–1 Loss, Bias, Variance, RMSE, Log-Loss, Training and Classification time, and show that softmax LR leads to significantly better RMSE and Log-Loss. We evaluate our proposed approach on 51 benchmark datasets.

Recent advances have demonstrated substantial benefits from learning with both generative and discriminative parameters. On the one hand, generative approaches address the estimation of the parameters of the joint distribution— $$\mathrm{P}(y,\mathbf{x})$$ , which for most network types is very computationally efficient (a notable exception to this are Markov networks) and on the other hand, discriminative approaches address the estimation of the parameters of the posterior distribution—and, are more effective for classification, since they fit $$\mathrm{P}(y|\mathbf{x})$$ directly. However, discriminative approaches are less computationally efficient as the normalization factor in the conditional log-likelihood precludes the derivation of closed-form estimation of parameters. This paper introduces a new discriminative parameter learning method for Bayesian network classifiers that combines in an elegant fashion parameters learned using both generative and discriminative methods. The proposed method is discriminative in nature, but uses estimates of generative probabilities to speed-up the optimization process. A second contribution is to propose a simple framework to characterize the parameter learning task for Bayesian network classifiers. We conduct an extensive set of experiments on 72 standard datasets and demonstrate that our proposed discriminative parameterization provides an efficient alternative to other state-of-the-art parameterizations.

Ever increasing data quantity makes ever more urgent the need for highly scalable learners that have good classification performance. Therefore, an out-of-core learner with excellent time and space complexity, along with high expressivity (that is, capacity to learn very complex multivariate probability distributions) is extremely desirable. This paper presents such a learner. We propose an extension to the $k$-dependence Bayesian classifier (KDB) that discriminatively selects a sub- model of a full KDB classifier. It requires only one additional pass through the training data, making it a three-pass learner. Our extensive experimental evaluation on $16$ large data sets reveals that this out-of-core algorithm achieves competitive classification performance, and substantially better training and classification time than state-of-the-art in-core learners such as random forest and linear and non-linear logistic regression.

This paper introduces Accelerated Logistic Regression : a hybrid generative-discriminative approach to training Logistic Regression with high-order features. We present two main results: (1) that our combined generative-discriminative approach significantly improves the efficiency of Logistic Regression and (2) that incorporating higher order features (i.e. features that are the Cartesian products of the original features) reduces the bias of Logistic Regression, which in turn significantly reduces its error on large datasets. We assess the efficacy of Accelerated Logistic Regression by conducting an extensive set of experiments on 75 standard datasets. We demonstrate its competitiveness, particularly on large datasets, by comparing against state-of-the-art classifiers including Random Forest and Averaged n -Dependence Estimators.

We propose an alternative parameterization of Logistic Regression (LR) for the categorical data, multi-class setting. LR optimizes the conditional log-likelihood over the training data and is based on an iterative optimization procedure to tune this objective function. The optimization procedure employed may be sensitive to scale and hence an effective pre-conditioning method is recommended. Many problems in machine learning involve arbitrary scales or categorical data (where simple standardization of features is not applicable). The problem can be alleviated by using optimization routines that are invariant to scale such as (second-order) Newton methods. However, computing and inverting the Hessian is a costly procedure and not feasible for big data. Thus one must often rely on first-order methods such as gradient descent (GD), stochastic gradient descent (SGD) or approximate second-order such as quasi-Newton (QN) routines, which are not invariant to scale. This paper proposes a simple yet effective pre-conditioner for speeding-up LR based on naive Bayes conditional probability estimates. The idea is to scale each attribute by the log of the conditional probability of that attribute given the class. This formulation substantially speeds-up LR's convergence. It also provides a weighted naive Bayes formulation which yields an effective framework for hybrid generative-discriminative classification.

In this paper we propose a confidence rated boosting algorithm based on Ada-boost for generic object detection. Confidence rated Ada-boost algorithm has not been applied to generic object detection problem; in that sense our work is novel. We represent images as bag of words, where the words are SIFT descriptors extracted over some interest points. We compare our boosting algorithm to another version of boosting algorithm called Gentle-boost. Our approach generalizes well and performs equal or better than Gentle-boost. We show our results on four categories from the Caltech data sets, in terms of ROC curves.