Search for: All records

Accurate chemical reaction prediction is essential for drug discovery and synthetic planning. However, this task becomes particularly challenging in low-data scenarios, where novel reaction types lack sufficient training examples. To address this challenge, we propose FewRxn, a novel model-agnostic few-shot reaction prediction framework that enables rapid adaptation to unseen reaction types using only a few training samples. FewRxn integrates several key innovations, including segmentation masks for enhanced reactant representation, fingerprint embeddings for richer molecular context, and task-aware meta-learning for effective knowledge transfer. Through extensive evaluations, FewRxn achieves state-of-the-art accuracy in few-shot settings, significantly outperforming traditional fine-tuning methods. Additionally, our work provides insights into the impact of molecular representations on reaction knowledge transfer, demonstrating that knowledge captured under molecular graph-based formulation consistently outperforms those learned in forms of SMILES generation in few-shot learning. 

Graph Neural Networks (GNNs) resurge as a trending research subject owing to their impressive ability to capture representations from graph-structured data. However, the black-box nature of GNNs presents a significant challenge in terms of comprehending and trusting these models, thereby limiting their practical applications in mission-critical scenarios. Although there has been substantial progress in the field of explaining GNNs in recent years, the majority of these studies are centered on static graphs, leaving the explanation of dynamic GNNs less explored. Dynamic GNNs, with their ever-evolving graph structures, pose a unique challenge and require additional efforts to effectively capture temporal dependencies and structural relationships. To address this challenge, we present DyExplainer, a novel approach to explaining dynamic GNNs on the fly. DyExplainer trains a dynamic GNN backbone to extract representations of the graph at each snapshot, while simultaneously exploring structural relationships and temporal dependencies through a sparse attention technique. To preserve the desired properties of the explanation, such as structural consistency and temporal continuity, we augment our approach with contrastive learning techniques to providea priori-guided regularization. To model longer-term temporal dependencies, we develop a buffer-based live-updating scheme for training. The results of our extensive experiments on various datasets demonstrate the superiority of DyExplainer, not only providing faithful explainability of the model predictions but also significantly improving the model prediction accuracy, as evidenced in the link prediction task. 

Recent research has developed a number of eXplainable AI (XAI) techniques, such as gradient-based approaches, input perturbation-base methods, and black-box explanation methods. While these XAI techniques can extract meaningful insights from deep learning models, how to properly evaluate them remains an open problem. The most widely used approach is to perturb or even remove what the XAI method considers to be the most important features in an input and observe the changes in the output prediction. This approach, although straightforward, suffers the Out-of-Distribution (OOD) problem as the perturbed samples may no longer follow the original data distribution. A recent method RemOve And Retrain (ROAR) solves the OOD issue by retraining the model with perturbed samples guided by explanations. However, using the model retrained based on XAI methods to evaluate these explainers may cause information leakage and thus lead to unfair comparisons. We propose Fine-tuned Fidelity (F-Fidelity), a robust evaluation framework for XAI, which utilizes i) an explanation-agnostic fine-tuning strategy, thus mitigating the information leakage issue, and ii) a random masking operation that ensures that the removal step does not generate an OOD input. We also design controlled experiments with state-of-the-art (SOTA) explainers and their degraded version to verify the correctness of our framework. We conduct experiments on multiple data modalities, such as images, time series, and natural language. The results demonstrate that F-Fidelity significantly improves upon prior evaluation metrics in recovering the ground-truth ranking of the explainers. Furthermore, we show both theoretically and empirically that, given a faithful explainer, the F-Fidelity metric can be used to compute the sparsity of influential input components, i.e., to extract the true explanation size. 

Time series analysis has evolved from traditional autoregressive models to deep learning, Transformers, and Large Language Models (LLMs). While vision models have also been explored along the way, their contributions are less recognized due to the predominance of sequence modeling. However, challenges such as the mismatch between continuous time series and LLMs’ discrete token space, and the difficulty in capturing multivariate correlations, have led to growing interest in Large Vision Models (LVMs) and Vision-Language Models (VLMs). This survey highlights the advantages of vision models over LLMs in time series analysis, offering a comprehensive dual-view taxonomy that answers key research questions like how to encode time series as images and how to model imaged time series. Additionally, we address pre- and post-processing challenges in this framework and outline future directions for advancing the field. 

The 2019 report of ferroelectricity in (Al,Sc)N [Fichtner et al., J. Appl. Phys. 125, 114103 (2019)] broke a long-standing tradition of considering AlN the textbook example of a polar but non-ferroelectric material. Combined with the recent emergence of ferroelectricity in HfO2-based fluorites [Böscke et al., Appl. Phys. Lett. 99, 102903 (2011)], these unexpected discoveries have reinvigorated studies of integrated ferroelectrics, with teams racing to understand the fundamentals and/or deploy these new materials—or, more correctly, attractive new capabilities of old materials—in commercial devices. The five years since the seminal report of ferroelectric (Al,Sc)N [Fichtner et al., J. Appl. Phys. 125, 114103 (2019)] have been particularly exciting, and several aspects of recent advances have already been covered in recent review articles [Jena et al., Jpn. J. Appl. Phys. 58, SC0801 (2019); Wang et al., Appl. Phys. Lett. 124, 150501 (2024); Kim et al., Nat. Nanotechnol. 18, 422–441 (2023); and F. Yang, Adv. Electron. Mater. 11, 2400279 (2024)]. We focus here on how the ferroelectric wurtzites have made the field rethink domain walls and the polarization reversal process—including the very character of spontaneous polarization itself—beyond the classic understanding that was based primarily around perovskite oxides and extended to other chemistries with various caveats. The tetrahedral and highly covalent bonding of AlN along with the correspondingly large bandgap lead to fundamental differences in doping/alloying, defect compensation, and charge distribution when compared to the classic ferroelectric systems; combined with the unipolar symmetry of the wurtzite structure, the result is a class of ferroelectrics that are both familiar and puzzling, with characteristics that seem to be perfectly enabling and simultaneously nonstarters for modern integrated devices. The goal of this review is to (relatively) quickly bring the reader up to speed on the current—at least as of early 2025—understanding of domains and defects in wurtzite ferroelectrics, covering the most relevant work on the fundamental science of these materials as well as some of the most exciting work in early demonstrations of device structures.