Text classification aims to effectively categorize documents into pre-defined categories. Traditional methods for text classification often rely on large amounts of manually annotated training data, making the process time-consuming and labor-intensive. To address this issue, recent studies have focused on weakly-supervised and extremely weakly-supervised settings, which require minimal or no human annotation, respectively. In previous methods of weakly supervised text classification, pseudo-training data is generated by assigning pseudo-labels to documents based on their alignment (e.g., keyword matching) with specific classes. However, these methods ignore the importance of incorporating the explanations of the generated pseudo-labels, or saliency of individual words, as additional guidance during the text classification training process. To address this limitation, we propose XAI-CLASS, a novel explanation-enhanced extremely weakly-supervised text classification method that incorporates word saliency prediction as an auxiliary task. XAI-CLASS begins by employing a multi-round question-answering process to generate pseudo-training data that promotes the mutual enhancement of class labels and corresponding explanation word generation. This pseudo-training data is then used to train a multi-task framework that simultaneously learns both text classification and word saliency prediction. Extensive experiments on several weakly-supervised text classification datasets show that XAI-CLASS outperforms other weakly-supervised text classification methods significantly. Moreover, experiments demonstrate that XAI-CLASS enhances both model performance and explainability.
Diagnosis of major depressive disorder (MDD) primarily relies on the patient's self-reported symptoms and a clinical evaluation. Effective connectivity (EC) from resting-state functional magnetic resonance imaging (rs-fMRI) analysis can reflect the directionality of connections between brain regions, making it a candidate method to classify MDD. This study used Granger causality analysis to extract EC features from a large multi-site MDD dataset. The ComBat algorithm and multivariate linear regression were used to harmonize site difference and to remove age and sex covariates, respectively. Two-sample t-tests and model-based feature selection methods were used to screen for highly discriminative EC features for MDD, and LightGBM was used to classify MDD. In this large-scale multi-site rs-fMRI dataset, 97 EC features deemed highly discriminative for MDD were screened. In the nested five-fold cross-validation, the best classification model with the 97 EC features achieved accuracy, sensitivity, and specificity of 94.35%, 93.52%, and 95.25%, respectively. In another independent large dataset, which tested the generalization performance of the 97 EC features, the best classification models achieved 94.74%, 90.59%, and 96.75% for accuracy, sensitivity, and specificity, respectively. This work demonstrated that EC had a reasonable discriminative ability and supported the notion for using EC to potentially assist clinical diagnosis of MDD.
Hierarchical classification (HC) plays a pivotal role in multi-class classification tasks, where objects are organized into a hierarchical structure. This study explores the performance of HC through a comprehensive analysis that encompasses both hierarchy generation and hierarchy exploitation. This analysis is particularly relevant in scenarios where a predefined hierarchy structure is not readily accessible. Notably, two novel hierarchy exploitation schemes, LCPN+ and LCPN+F, which extend the capabilities of LCPN and combine the strengths of global and local classification, have been introduced and evaluated alongside existing methods. The findings reveal the consistent superiority of LCPN+F, which outperforms other schemes across various datasets and scenarios. Moreover, this research emphasizes not only effectiveness but also efficiency, as LCPN+ and LCPN+F maintain runtime performance comparable to Flat Classification (FC). Additionally, this study underscores the importance of selecting the right hierarchy exploitation scheme to maximize classification performance. This work extends our understanding of HC and establishes a benchmark for future research, fostering advancements in multi-class classification methodologies.
Neural networks are susceptible to privacy attacks. To date, no verifier can reason about the privacy of individuals participating in the training set. We propose a new privacy property, called local differential classification privacy (LDCP), extending local robustness to a differential privacy setting suitable for black-box classifiers. Given a neighborhood of inputs, a classifier is LDCP if it classifies all inputs the same regardless of whether it is trained with the full dataset or whether any single entry is omitted. A naive algorithm is highly impractical because it involves training a very large number of networks and verifying local robustness of the given neighborhood separately for every network. We propose Sphynx, an algorithm that computes an abstraction of all networks, with a high probability, from a small set of networks, and verifies LDCP directly on the abstract network. The challenge is twofold: network parameters do not adhere to a known distribution probability, making it difficult to predict an abstraction, and predicting too large abstraction harms the verification. Our key idea is to transform the parameters into a distribution given by KDE, allowing to keep the over-approximation error small. To verify LDCP, we extend a MILP verifier to analyze an abstract network. Experimental results show that by training only 7% of the networks, Sphynx predicts an abstract network obtaining 93% verification accuracy and reducing the analysis time by $1.7\cdot10^4$x.
Cut-generating linear programs (CGLPs) play a key role as a separation oracle to produce valid inequalities for the feasible region of mixed-integer programs. When incorporated inside branch-and-bound, the cutting planes obtained from CGLPs help to tighten relaxations and improve dual bounds. However, running the CGLPs at the nodes of the branch-and-bound tree is computationally cumbersome due to the large number of node candidates and the lack of a priori knowledge on which nodes admit useful cutting planes. As a result, CGLPs are often avoided at default settings of branch-and-cut algorithms despite their potential impact on improving dual bounds. In this paper, we propose a novel framework based on machine learning to approximate the optimal value of a CGLP class that determines whether a cutting plane can be generated at a node of the branch-and-bound tree. Translating the CGLP as an indicator function of the objective function vector, we show that it can be approximated through conventional data classification techniques. We provide a systematic procedure to efficiently generate training data sets for the corresponding classification problem based on the CGLP structure. We conduct computational experiments on benchmark instances using classification methods such as logistic regression. These results suggest that the approximate CGLP obtained from classification can improve the solution time compared to that of conventional cutting plane methods. Our proposed framework can be efficiently applied to a large number of nodes in the branch-and-bound tree to identify the best candidates for adding a cut.
Transformer-based models, specifically BERT, have propelled research in various NLP tasks. However, these models are limited to a maximum token limit of 512 tokens. Consequently, this makes it non-trivial to apply it in a practical setting with long input. Various complex methods have claimed to overcome this limit, but recent research questions the efficacy of these models across different classification tasks. These complex architectures evaluated on carefully curated long datasets perform at par or worse than simple baselines. In this work, we propose a relatively simple extension to vanilla BERT architecture called ChunkBERT that allows finetuning of any pretrained models to perform inference on arbitrarily long text. The proposed method is based on chunking token representations and CNN layers, making it compatible with any pre-trained BERT. We evaluate chunkBERT exclusively on a benchmark for comparing long-text classification models across a variety of tasks (including binary classification, multi-class classification, and multi-label classification). A BERT model finetuned using the ChunkBERT method performs consistently across long samples in the benchmark while utilizing only a fraction (6.25\%) of the original memory footprint. These findings suggest that efficient finetuning and inference can be achieved through simple modifications to pre-trained BERT models.
The machine learning communities, such as those around computer vision or natural language processing, have developed numerous supportive tools and benchmark datasets to accelerate the development. In contrast, the network traffic classification field lacks standard benchmark datasets for most tasks, and the available supportive software is rather limited in scope. This paper aims to address the gap and introduces DataZoo, a toolset designed to streamline dataset management in network traffic classification and to reduce the space for potential mistakes in the evaluation setup. DataZoo provides a standardized API for accessing three extensive datasets -- CESNET-QUIC22, CESNET-TLS22, and CESNET-TLS-Year22. Moreover, it includes methods for feature scaling and realistic dataset partitioning, taking into consideration temporal and service-related factors. The DataZoo toolset simplifies the creation of realistic evaluation scenarios, making it easier to cross-compare classification methods and reproduce results.
Privacy is a complex, subjective and contextual concept that is difficult to define. Therefore, the annotation of images to train privacy classifiers is a challenging task. In this paper, we analyse privacy classification datasets and the properties of controversial images that are annotated with contrasting privacy labels by different assessors. We discuss suitable features for image privacy classification and propose eight privacy-specific and human-interpretable features. These features increase the performance of deep learning models and, on their own, improve the image representation for privacy classification compared with much higher dimensional deep features.
Named Entity Recognition (NER) and Relation Classification (RC) are important steps in extracting information from unstructured text and formatting it into a machine-readable format. We present a survey of recent deep learning models that address named entity recognition and relation classification, with focus on few-shot learning performance. Our survey is helpful for researchers in knowing the recent techniques in text mining and extracting structured information from raw text.
As a typical self-paced brain-computer interface (BCI) system, the motor imagery (MI) BCI has been widely applied in fields such as robot control, stroke rehabilitation, and assistance for patients with stroke or spinal cord injury. Many studies have focused on the traditional spatial filters obtained through the common spatial pattern (CSP) method. However, the CSP method can only obtain fixed spatial filters for specific input signals. Besides, CSP method only focuses on the variance difference of two types of electroencephalogram (EEG) signals, so the decoding ability of EEG signals is limited. To obtain more effective spatial filters for better extraction of spatial features that can improve classification to MI-EEG, this paper proposes an adaptive spatial filter solving method based on particle swarm optimization algorithm (PSO). A training and testing framework based on filter bank and spatial filters (FBCSP-ASP) is designed for MI EEG signal classification. Comparative experiments are conducted on two public datasets (2a and 2b) from BCI competition IV, which show the outstanding average recognition accuracy of FBCSP-ASP. The proposed method has achieved significant performance improvement on MI-BCI. The classification accuracy of the proposed method has reached 74.61% and 81.19% on datasets 2a and 2b, respectively. Compared with the baseline algorithm (FBCSP), the proposed algorithm improves 11.44% and 7.11% on two datasets respectively. Furthermore, the analysis based on mutual information, t-SNE and Shapley values further proves that ASP features have excellent decoding ability for MI-EEG signals, and explains the improvement of classification performance by the introduction of ASP features.
Due to the powerful ability in capturing the global information, Transformer has become an alternative architecture of CNNs for hyperspectral image classification. However, general Transformer mainly considers the global spectral information while ignores the multiscale spatial information of the hyperspectral image. In this paper, we propose a multiscale spectral-spatial convolutional Transformer (MultiscaleFormer) for hyperspectral image classification. First, the developed method utilizes multiscale spatial patches as tokens to formulate the spatial Transformer and generates multiscale spatial representation of each band in each pixel. Second, the spatial representation of all the bands in a given pixel are utilized as tokens to formulate the spectral Transformer and generate the multiscale spectral-spatial representation of each pixel. Besides, a modified spectral-spatial CAF module is constructed in the MultiFormer to fuse cross-layer spectral and spatial information. Therefore, the proposed MultiFormer can capture the multiscale spectral-spatial information and provide better performance than most of other architectures for hyperspectral image classification. Experiments are conducted over commonly used real-world datasets and the comparison results show the superiority of the proposed method.
While existing strategies to execute deep learning-based classification on low-power platforms assume the models are trained on all classes of interest, this paper posits that adopting context-awareness i.e. narrowing down a classification task to the current deployment context consisting of only recent inference queries can substantially enhance performance in resource-constrained environments. We propose a new paradigm, CACTUS, for scalable and efficient context-aware classification where a micro-classifier recognizes a small set of classes relevant to the current context and, when context change happens (e.g., a new class comes into the scene), rapidly switches to another suitable micro-classifier. CACTUS features several innovations, including optimizing the training cost of context-aware classifiers, enabling on-the-fly context-aware switching between classifiers, and balancing context switching costs and performance gains via simple yet effective switching policies. We show that CACTUS achieves significant benefits in accuracy, latency, and compute budget across a range of datasets and IoT platforms.
Small data learning problems are characterized by a significant discrepancy between the limited amount of response variable observations and the large feature space dimension. In this setting, the common learning tools struggle to identify the features important for the classification task from those that bear no relevant information, and cannot derive an appropriate learning rule which allows to discriminate between different classes. As a potential solution to this problem, here we exploit the idea of reducing and rotating the feature space in a lower-dimensional gauge and propose the Gauge-Optimal Approximate Learning (GOAL) algorithm, which provides an analytically tractable joint solution to the dimension reduction, feature segmentation and classification problems for small data learning problems. We prove that the optimal solution of the GOAL algorithm consists in piecewise-linear functions in the Euclidean space, and that it can be approximated through a monotonically convergent algorithm which presents -- under the assumption of a discrete segmentation of the feature space -- a closed-form solution for each optimization substep and an overall linear iteration cost scaling. The GOAL algorithm has been compared to other state-of-the-art machine learning (ML) tools on both synthetic data and challenging real-world applications from climate science and bioinformatics (i.e., prediction of the El Nino Southern Oscillation and inference of epigenetically-induced gene-activity networks from limited experimental data). The experimental results show that the proposed algorithm outperforms the reported best competitors for these problems both in learning performance and computational cost.
The bottom two layers of a neuromorphic architecture are designed and shown to be capable of online clustering and supervised classification. An active spiking dendrite model is used, and a single dendritic segment performs essentially the same function as a classic integrate-and-fire point neuron. A single dendrite is then composed of multiple segments and is capable of online clustering. Although this work focuses primarily on dendrite functionality, a multi-point neuron can be formed by combining multiple dendrites. To demonstrate its clustering capability, a dendrite is applied to spike sorting, an important component of brain-computer interface applications. Supervised online classification is implemented as a network composed of multiple dendrites and a simple voting mechanism. The dendrites operate independently and in parallel. The network learns in an online fashion and can adapt to macro-level changes in the input stream. Achieving brain-like capabilities, efficiencies, and adaptability will require a significantly different approach than conventional deep networks that learn via compute-intensive back propagation. The model described herein may serve as the foundation for such an approach.
Generative adversarial networks (GANs), modeled as a zero-sum game between a generator (G) and a discriminator (D), allow generating synthetic data with formal guarantees. Noting that D is a classifier, we begin by reformulating the GAN value function using class probability estimation (CPE) losses. We prove a two-way correspondence between CPE loss GANs and $f$-GANs which minimize $f$-divergences. We also show that all symmetric $f$-divergences are equivalent in convergence. In the finite sample and model capacity setting, we define and obtain bounds on estimation and generalization errors. We specialize these results to $\alpha$-GANs, defined using $\alpha$-loss, a tunable CPE loss family parametrized by $\alpha\in(0,\infty]$. We next introduce a class of dual-objective GANs to address training instabilities of GANs by modeling each player's objective using $\alpha$-loss to obtain $(\alpha_D,\alpha_G)$-GANs. We show that the resulting non-zero sum game simplifies to minimizing an $f$-divergence under appropriate conditions on $(\alpha_D,\alpha_G)$. Generalizing this dual-objective formulation using CPE losses, we define and obtain upper bounds on an appropriately defined estimation error. Finally, we highlight the value of tuning $(\alpha_D,\alpha_G)$ in alleviating training instabilities for the synthetic 2D Gaussian mixture ring as well as the large publicly available Celeb-A and LSUN Classroom image datasets.
We consider the problem of learning classification trees that are robust to distribution shifts between training and testing/deployment data. This problem arises frequently in high stakes settings such as public health and social work where data is often collected using self-reported surveys which are highly sensitive to e.g., the framing of the questions, the time when and place where the survey is conducted, and the level of comfort the interviewee has in sharing information with the interviewer. We propose a method for learning optimal robust classification trees based on mixed-integer robust optimization technology. In particular, we demonstrate that the problem of learning an optimal robust tree can be cast as a single-stage mixed-integer robust optimization problem with a highly nonlinear and discontinuous objective. We reformulate this problem equivalently as a two-stage linear robust optimization problem for which we devise a tailored solution procedure based on constraint generation. We evaluate the performance of our approach on numerous publicly available datasets, and compare the performance to a regularized, non-robust optimal tree. We show an increase of up to 12.48% in worst-case accuracy and of up to 4.85% in average-case accuracy across several datasets and distribution shifts from using our robust solution in comparison to the non-robust one.
A scenario-based testing approach can reduce the time required to obtain statistically significant evidence of the safety of Automated Driving Systems (ADS). Identifying these scenarios in an automated manner is a challenging task. Most methods on scenario classification do not work for complex scenarios with diverse environments (highways, urban) and interaction with other traffic agents. This is mirrored in their approaches which model an individual vehicle in relation to its environment, but neglect the interaction between multiple vehicles (e.g. cut-ins, stationary lead vehicle). Furthermore, existing datasets lack diversity and do not have per-frame annotations to accurately learn the start and end time of a scenario. We propose a method for complex traffic scenario classification that is able to model the interaction of a vehicle with the environment, as well as other agents. We use Graph Convolutional Networks to model spatial and temporal aspects of these scenarios. Expanding the nuScenes and Argoverse 2 driving datasets, we introduce a scenario-labeled dataset, which covers different driving environments and is annotated per frame. Training our method on this dataset, we present a promising baseline for future research on per-frame complex scenario classification.
Existing privacy-preserving speech representation learning methods target a single application domain. In this paper, we present a novel framework to anonymize utterance-level speech embeddings generated by pre-trained encoders and show its effectiveness for a range of speech classification tasks. Specifically, given the representations from a pre-trained encoder, we train a Transformer to estimate the representations for the same utterances spoken by other speakers. During inference, the extracted representations can be converted into different identities to preserve privacy. We compare the results with the voice anonymization baselines from the VoicePrivacy 2022 challenge. We evaluate our framework on speaker identification for privacy and emotion recognition, depression classification, and intent classification for utility. Our method outperforms the baselines on privacy and utility in paralinguistic tasks and achieves comparable performance for intent classification.
Graph neural networks (GNNs) have emerged as a powerful tool for tasks such as node classification and graph classification. However, much less work has been done on signal classification, where the data consists of many functions (referred to as signals) defined on the vertices of a single graph. These tasks require networks designed differently from those designed for traditional GNN tasks. Indeed, traditional GNNs rely on localized low-pass filters, and signals of interest may have intricate multi-frequency behavior and exhibit long range interactions. This motivates us to introduce the BLIS-Net (Bi-Lipschitz Scattering Net), a novel GNN that builds on the previously introduced geometric scattering transform. Our network is able to capture both local and global signal structure and is able to capture both low-frequency and high-frequency information. We make several crucial changes to the original geometric scattering architecture which we prove increase the ability of our network to capture information about the input signal and show that BLIS-Net achieves superior performance on both synthetic and real-world data sets based on traffic flow and fMRI data.
Diabetic retinopathy (DR) is caused by long-standing diabetes and is among the fifth leading cause for visual impairments. The process of early diagnosis and treatments could be helpful in curing the disease, however, the detection procedure is rather challenging and mostly tedious. Therefore, automated diabetic retinopathy classification using deep learning techniques has gained interest in the medical imaging community. Akin to several other real-world applications of deep learning, the typical assumption of i.i.d data is also violated in DR classification that relies on deep learning. Therefore, developing DR classification methods robust to unseen distributions is of great value. In this paper, we study the problem of generalizing a model to unseen distributions or domains (a.k.a domain generalization) in DR classification. To this end, we propose a simple and effective domain generalization (DG) approach that achieves self-distillation in vision transformers (ViT) via a novel prediction softening mechanism. This prediction softening is an adaptive convex combination one-hot labels with the model's own knowledge. We perform extensive experiments on challenging open-source DR classification datasets under both multi-source and single-source DG settings with three different ViT backbones to establish the efficacy and applicability of our approach against competing methods. For the first time, we report the performance of several state-of-the-art DG methods on open-source DR classification datasets after conducting thorough experiments. Finally, our method is also capable of delivering improved calibration performance than other methods, showing its suitability for safety-critical applications, including healthcare. We hope that our contributions would investigate more DG research across the medical imaging community.