Graph Neural Networks (GNNs), with their ability to effectively handle non-Euclidean data structures, have demonstrated state-of-the-art performance in learning node and graph-level representations. However, GNNs face significant computational overhead due to their message-passing mechanisms, making them impractical for real-time large-scale applications. Recently, Graph-to-MLP (G2M) knowledge distillation has emerged as a promising solution, utilizing MLPs to reduce inference latency. However, existing methods often lack structural awareness (SA), limiting their ability to capture essential graph-specific information. Moreover, some methods require access to large-scale graphs, undermining their scalability. To address these issues, we propose SALE-MLP (Structure-Aware Latent Embeddings for GNN-to-Graph-Free MLP Distillation), a novel graph-free and structure-aware approach that leverages unsupervised structural losses to align the MLP feature space with the underlying graph structure. SALE-MLP does not rely on precomputed GNN embeddings nor require graph during inference, making it efficient for real-world applications. Extensive experiments demonstrate that SALE-MLP outperforms existing G2M methods across tasks and datasets, achieving 3–4% improvement in node classification for inductive settings while maintaining strong transductive performance.

Graph Neural Networks (Gnns) have achieved remarkable success in various downstream tasks, such as node classification and link prediction. Yet, efficiently deploying Gnns remains a challenge due to their computational complexity. Graph knowledge distillation aims to address this by transferring task-specific structural knowledge from teacher Gnns to lightweight student Gnns or Multi-Layer Perceptrons (MLPs). Despite its promise, existing distillation approaches suffer from several limitations: (i) they require extensive task-specific supervision(ii) they must be retrained separately for each downstream task, and (iii) they often struggle in heterophilous settings. To overcome these challenges, we propose TAG2M, a Task-Agnostic Gnn-to-MLP distillation framework designed for efficient and accurate few-shot inference. TAG2M introduces several novel strategies, including a self-supervised contrastive loss that captures topological information solely from node attributes. Additionally, it leverages Lipschitz embeddings to encode positional information with provable distortion bounds, ensuring robust representation learning. To further enhance adaptability for few-shot inference, TAG2M incorporates a learnable prompt head, which facilitates rapid task adaptation even in label-scarce settings. Unlike prior methods, TAG2M generalizes well across both homophilous and heterophilous datasets while delivering a significant computational advantage, achieving up to a 20X -200X speed-up. Extensive evaluations on 11 public datasets demonstrate its superior accuracy across diverse tasks, including node classification, link prediction, and node regression, outperforming state-of-the-art approaches.

Partial label learning (PLL) aims to train classification models under conditions where each training sample is associated with a candidate set of labels. This set contains multiple labels, among which only one is correct. This work addresses instance-dependent noise in PLL by leveraging hierarchical structures within the label space. We introduce a method to derive label hierarchies from instance-dependent partial labels. Subsequently, we propose Structural Hierarchies for Instance-dependent Partial label (SHIP). SHIP is a modular component that integrates into deep learning architectures with applications that have intrinsic hierarchies. SHIP harnesses label hierarchy to enhance instance-dependent PLL performance across various deep-learning algorithms with hierarchy in the dataset. We conduct experiments on five publicly available benchmark datasets with four recent PLL algorithms. Experimental results show that incorporating SHIP into state-of-the-art architectures yields up to a 2.6% improvement in accuracy when hierarchies are present in the data. Moreover, when the number of classes is high, SHIP achieves up to a 2.5% reduction in mean mistake severity, highlighting its effectiveness in mitigating error severity.

Many existing Graph Neural Networks (GNN) methods assume that labels are reliable and sufficient, which may not be the case in real-world scenarios. This paper addresses one such problem of Partial Label Learning (PLL) on graph-structured data. In the PLL for graphs, each node is represented by a candidate set of labels, where only one is true while the others are inaccurate. Despite advancements with PLL in tabular and vision domains, the graph-structured data still needs to be explored. In this work, we first define PLL for graphs. Subsequently, we propose a new PLD-Graph algorithm for PLL in homogeneous graphs with scarce labels. We utilize graph augmentation to reduce the effects of inexact labels and provide additional supervision from unlabeled nodes. Progressive label disambiguation is performed based on the model's ability to predict correct classes. Furthermore, an additional loss estimates the label corruption matrix to capture associations between correct and incorrect labels. We show the effectiveness of the proposed algorithm on multiple graph datasets, with two types of noise and varying levels of ambiguous labels. Overall, the proposed PLD-Graph algorithm outperforms state-of-the-art PLL methods.

Latent fingerprint recognition involves acquisition and comparison of latent fingerprints with an exemplar gallery of fingerprints. The diversity in the type of surface leads to different procedures to recover the latent fingerprint. The appearance of latent fingerprints vary significantly due to the development techniques, leading to large intra-class variation. Due to lack of large datasets acquired using multiple mechanisms and surfaces, existing algorithms for latent fingerprint enhancement and comparison may perform poorly. In this study, we propose a Multi-Surface Multi-Technique (MUST) Latent Fingerprint Database. The database consists of more than 16,000 latent fingerprint impressions from 120 unique classes (120 fingers from 12 participants). Including corresponding exemplar fingerprints (livescan and rolled) and extended gallery, the dataset has nearly 21,000 impressions. It has latent fingerprints acquired under 35 different scenarios and additional four subsets of exemplar prints captured using live scan sensor and inked-rolled prints. With 39 different subsets, the database illustrates intra-class variations in latent fingerprints. The database has a potential usage towards building robust algorithms for latent fingerprint enhancement, segmentation, comparison, and multi-task learning. We also provide annotations for manually marked minutiae, acquisition Pixel Per Inch (PPI), and semantic segmentation masks. We also present the experimental protocol and the baseline results for the proposed dataset. The availability of the proposed database can encourage research in handling intra-class variation in latent fingerprint recognition.

Temporal Point Processes (TPP) are probabilistic generative frameworks. They model discrete event sequences localized in continuous time. Generally, real-life events reveal descriptive information, known as marks. Marked TPPs model time and marks of the event together for practical relevance. Conditioned on past events, marked TPPs aim to learn the joint distribution of the time and the mark of the next event. For simplicity, conditionally independent TPP models assume time and marks are independent given event history. They factorize the conditional joint distribution of time and mark into the product of individual conditional distributions. This structural limitation in the design of TPP models hurt the predictive performance on entangled time and mark interactions. In this work, we model the conditional inter-dependence of time and mark to overcome the limitations of conditionally independent models. We construct a multivariate TPP conditioning the time distribution on the current event mark in addition to past events. Besides the conventional intensity-based models for conditional joint distribution, we also draw on flexible intensity-free TPP models from the literature. The proposed TPP models outperform conditionally independent and dependent models in standard prediction tasks. Our experimentation on various datasets with multiple evaluation metrics highlights the merit of the proposed approach.

In many surveillance applications, the cameras are placed at overhead heights for human identification. In such real-world scenarios, the person of interest might be walking away from the camera and the only information available is "image of the person's head". In this research, we investigate the usage of head images for person recognition and propose it as a soft-biometric modality. With its viability for human recognition, application of head images can also be extended with other face recognition algorithms for surveillance. We propose a head image database pertaining to 103 subjects with more than 600 images. In addition to the database, we propose a framework for head image-based person verification. As a pre-processing stage, the framework includes evaluation of two segmentation algorithms. We also perform benchmarking evaluations of various texture, key-point, and learning-based representation algorithms and establish the baseline results. The experiments suggest that head images can be effectively used to ascertain human identity and the availability of this database could pave further research in this field.

In many applications such as law enforcement, attendance systems, and medical services, biometrics is utilized for identifying individuals. However, current systems, in general, do not enroll all possible age groups, particularly, toddlers and pre-school children. This research is the first of its kind attempt to prepare a multimodal biometric database for such potential users of biometric systems. In the proposed database, face, fingerprint, and iris modalities of over 100 children (age range of 18 months to 4 years) are captured in two different sessions, months apart. We also perform benchmarking evaluation of existing tools and algorithms to establish the baseline results for different unimodal and multimodal scenarios. Our experience and results suggest that while iris is highly accurate, it requires constant adult supervision to attain cooperation from children. On the other hand, face is the most easy-to-capture modality but yields very low verification performance. We assert that the availability of this database can instigate research in this important research problem.