Designing molecules with desired properties, aka the oRiented molEcule Design (RED), is a fundamental task in chemistry and materials science. While graph diffusion models (GDMs) and reinforcement learning techniques (RL) show promise in molecule structure generation and property optimization stages individually, their integration in the unified RED task often suffers from poor compatibility. The large variance among candidate molecular structures generated by GDMs can be amplified in the iterative optimization process of RL, leading to slow and unstable convergence. In this work, motivated by the adaptive and divide-and-conquer characteristics of Mixture of Experts (MoE) architecture, we propose a novel framework called MoE-Guided Graph Diffusion Model (MEGD) that incorporates the MoE architecture to guide the orchestration of GDM and RL, promoting faster and more stable convergence in the design process. MEGD is evaluated on benchmark datasets optimizing the physical and chemical properties of AI-generated molecular structures. On all three datasets, our method outperforms the best of 9 alternative models by 7.73% on the target structural properties, while not penalizing other important application-level quality metrics of the generated molecules. A real-world case study on an emerging class of material, i.e., metal-organic framework, is also conducted, which further demonstrates the effectiveness of our method in accomplishing the RED task.

Model visualization (ModelVis) has emerged as a major research direction, yet existing taxonomies are largely organized by data or tasks, making it difficult to treat models as first-class analysis objects. We present a model-centric two-stage framework that employs abstract listeners to capture spatial and temporal model behaviors, and then connects the translated model behavior data to the classical InfoVis pipeline. To apply the framework at scale, we build a retrieval-augmented human–large language model (LLM) extraction workflow and curate a corpus of 128 VIS/VAST ModelVis papers with 331 coded figures. Our analysis shows a dominant result-centric priority on visualizing model outcomes, quantitative/nominal data type, statistical charts, and performance evaluation. Citation-weighted trends further indicate that less frequent model-mechanism-oriented studies have disproportionately high impact while are less investigated recently. Overall, the framework is a general approach for comparing existing ModelVis systems and guiding possible future designs.

Understanding the evolution of scholarly research is essential for many real-life decision-making processes in academia, such as research planning, frontier exploration, and award selection. Popular platforms like Google Scholar and Web of Science rely on numerical indicators that are too abstract to convey the context and content of scientific research, while most existing visualization approaches on mapping science do not consider the presentation of individual scholars’ research evolution using curated self-citation data. This paper builds on our previous work and proposes an integrated pipeline to visualize a scholar’s research evolution from multiple topic facets. A novel 3D prism-shaped visual metaphor is introduced as the overview of a scholar’s research profile, whilst their scientific evolution on each topic is displayed in a more structured manner. Additional designs by topic chord diagram, streamgraph visualization, and inter-topic flow map, optimized by an elaborate layout algorithm, assist in perceiving the scholar’s scientific evolution across topics. A new six-degree-impact glyph metaphor highlights key interdisciplinary works driving the evolution. The proposed visualization methods are evaluated through case studies analyzing the careers of prestigious Turing award laureates, one major visualization venue, and a focused user study.

Message-passing Graph Neural Networks (GNNs) are often criticized for their limited expressiveness, issues like over-smoothing and over-squashing, and challenges in capturing long-range dependencies. Conversely, Graph Transformers (GTs) are regarded as superior due to their employment of global attention mechanisms, which potentially mitigate these challenges. Literature frequently suggests that GTs outperform GNNs in graph-level tasks, especially for graph classification and regression on small molecular graphs. In this study, we explore the untapped potential of GNNs through an enhanced framework, GNN+, which integrates six widely used techniques: edge feature integration, normalization, dropout, residual connections, feed-forward networks, and positional encoding, to effectively tackle graph-level tasks. We conduct a systematic re-evaluation of three classic GNNs (GCN, GIN, and GatedGCN) enhanced by the GNN+ framework across 14 well-known graph-level datasets. Our results reveal that, contrary to prevailing beliefs, these classic GNNs consistently match or surpass the performance of GTs, securing top-three rankings across all datasets and achieving first place in eight. Furthermore, they demonstrate greater efficiency, running several times faster than GTs on many datasets. This highlights the potential of simple GNN architectures, challenging the notion that complex mechanisms in GTs are essential for superior graph-level performance. Our source code is available at https://github.com/LUOyk1999/GNNPlus.

We present a novel end-to-end framework that generates highly compact (typically 6-15 dimensions), discrete (int4 type), and interpretable node representations, termed node identifiers (node IDs), to tackle inference challenges on large-scale graphs. By employing vector quantization, we compress continuous node embeddings from multiple layers of a Graph Neural Network (GNN) into discrete codes, applicable under both self-supervised and supervised learning paradigms. These node IDs capture high-level abstractions of graph data and offer interpretability that traditional GNN embeddings lack. Extensive experiments on 34 datasets, encompassing node classification, graph classification, link prediction, and attributed graph clustering tasks, demonstrate that the generated node IDs significantly enhance speed and memory efficiency while achieving competitive performance compared to current state-of-the-art methods.

Many contemporary data-driven research efforts in the natural sciences, such as chemistry and materials science, require large-scale, high-performance entity recognition from scientific datasets. Large language models (LLMs) have increasingly been adopted to solve the entity recognition task, with the same trend being observed on all-spectrum NLP tasks. The prevailing entity recognition LLMs rely on fine-tuned technology, yet the fine-tuning process often incurs significant cost. To achieve a best performance-cost trade-off, we propose ALLabel, a three-stage framework designed to select the most informative and representative samples in preparing the demonstrations for LLM modeling. The annotated examples are used to construct a ground-truth retrieval corpus for LLM in-context learning. By sequentially employing three distinct active learning strategies, ALLabel consistently outperforms all baselines under the same annotation budget across three specialized domain datasets. Experimental results also demonstrate that selectively annotating only 5%-10% of the dataset with ALLabel can achieve performance comparable to the method annotating the entire dataset. Further analyses and ablation studies verify the effectiveness and generalizability of our proposal.

Recent advances in retrieval-based in-context learning (ICL) train the retriever using a classification objective, which categorizes in-context examples (ICEs) into the most useful and the rest based on absolute scores. However, during inference, ICEs are retrieved by score ranking rather than classification — The classification training objective deviates from this test scenario. Hence, in this paper, we propose a novel algorithm that trains a retrieval model by ranking formulation, where the preference rankings between ICEs are given by comparing the likelihood of the LLM generating the correct answer conditioned on each exemplar. By learning to rank, we motivate the retriever to automatically learn diverse rationales why specific examples are more useful for ICL decisions. This addresses the issue that classification models poorly capture broader utility. Experimental results demonstrate the top-1 performance of our proposal across 9 NLP tasks, with ablation studies and case studies further validating the effectiveness of our design. *The code can be found in: https://github.com/2022neo/SeDPO_NIPS25*

The extraction of the metal-organic framework synthesis route from the literature has been crucial for the rational MOFs design with desirable functionality. The recent advent of large language models (LLMs) provides a disruptive new solution to this long-standing problem. While the latest research on chemical data extraction mostly adopts zero-shot LLMs lacking specialized material knowledge or fine-tuned LLMs inducing high cost and inflexibility in our scenario, we introduce in this work the MaterialBrain pipeline that optimizes the few-shot LLMs in-context learning technique to accurately extract synthesis routes and design high-performance materials. First, a batch-epoch-iteration-based human-AI data curation approach is proposed to optimize both the quantity and quality of annotation database for the synthesis extraction task, which are pivotal to MaterialBrain's performance. Second, an information retrieval algorithm is applied to pick and quantify a few-shot demonstrations from the annotation database for each extraction. Over three data sets randomly sampled from nearly 90,000 well-defined MOFs, we conduct triple evaluations to validate our pipeline. The synthesis extraction, structure inference, and material design performance of MaterialBrain significantly outplay zero-shot LLMs and baseline methods. The specific surface area of the lab-synthesized material guided by LLMs surpasses that of 99.2% of MOFs of the same class reported in the literature.

This paper focuses on the visual analytics support for extracting domain-specific entities from extensive scientific literature, a task with inherent limitations using traditional named entity resolution methods. With the advent of large language models (LLMs) such as GPT-4, significant improvements over conventional machine learning approaches have been achieved due to LLM’s capability on entity resolution integrate abilities such as understanding multiple types of text. This research introduces a new visual analysis pipeline that integrates these advanced LLMs with versatile visualization and interaction designs to support batch entity resolution. Specifically, we focus on a specific material science field of Metal-Organic Frameworks (MOFs) and a large data collection namely CSD-MOFs. Through collaboration with domain experts in material science, we obtain well-labeled synthesis paragraphs. We propose human-in-the-loop refinement over the entity resolution process using visual analytics techniques, which allows domain experts to interactively integrate insights into LLM intelligence, including error analysis and interpretation of the retrieval-augmented generation (RAG) algorithm. Our evaluation through the case study of example selection for RAG demonstrates that this visual analysis approach effectively improves the accuracy of single-document entity resolution.

Link prediction (LP) is crucial for Knowledge Graphs (KG) completion but commonly suffers from interpretability issues. While several methods have been proposed to explain embedding-based LP models, they are generally limited to local explanations on KG and are deficient in providing human interpretable semantics. Based on real-world observations of the characteristics of KGs from multiple domains, we propose to explain LP models in KG with path-based explanations. An integrated framework, namely eXpath, is introduced which incorporates the concept of relation path with ontological closed path rules to enhance both the efficiency and effectiveness of LP interpretation. Notably, the eXpath explanations can be fused with other single-link explanation approaches to achieve a better overall solution. Extensive experiments across benchmark datasets and LP models demonstrate that introducing eXpath can boost the quality of resulting explanations by about 20% on two key metrics and reduce the required explanation time by 61.4%, in comparison to the best existing method. Case studies further highlight eXpath's ability to provide more semantically meaningful explanations through path-based evidence.

Geo-entity resolution involves linking records that refer to the same entities across different spatial datasets, which underpins location-based services. Given the varying quality of geo-data, this task is known to be challenging, as directly comparing the semantic-centric representations of two entities is no longer reliable. To robustify geo-entity resolution in this context, the main research question is how to effectively extend the current semantics-centric representations of geo-entity with geographical context from its spatial neighbors. Existing methods consider names from neighbors, but they struggle to fully utilize the unaligned neighbor attributes. In this paper, we study the representation of geo-context for robust geo-entity resolution and propose two adaptations that efficiently leverage unaligned geo-entity attributes across spatial neighbors: (1) A plugin module, namely Unaligned Message-Passing (UMP), that propagates unaligned neighbor features to integrate geo-context into the token embeddings output by language model; (2) a contextualized pretraining framework (CP) that allows the former to leverage unlabelled geo-entity data. Experiments show that our method surpasses the baselines, achieving higher F1 scores on 8 real-world geo-datasets in terms of robustness, with an improvement of up to 7.9%. The ablation study further justifies our proposal.

Graph Transformers (GTs) have recently emerged as popular alternatives to traditional message-passing Graph Neural Networks (GNNs), due to their theoretically superior expressiveness and impressive performance reported on standard node classification benchmarks, often significantly outperforming GNNs. In this paper, we conduct a thorough empirical analysis to reevaluate the performance of three classic GNN models (GCN, GAT, and GraphSAGE) against GTs. Our findings suggest that the previously reported superiority of GTs may have been overstated due to suboptimal hyperparameter configurations in GNNs. Remarkably, with slight hyperparameter tuning, these classic GNN models achieve state-of-the-art performance, matching or even exceeding that of recent GTs across 17 out of the 18 diverse datasets examined. Additionally, we conduct detailed ablation studies to investigate the influence of various GNN configurations, such as normalization, dropout, residual connections, and network depth, on node classification performance. Our study aims to promote a higher standard of empirical rigor in the field of graph machine learning, encouraging more accurate comparisons and evaluations of model capabilities.

This article deeply studies the application of intelligent control system for high-voltage transmission converter valve based on computer technology in hybrid multi-terminal DC transmission system and its coordinated control strategy for DC voltage. With the help of advanced computer technology, the system achieves precise control and intelligent management of converter valve, significantly improving the stability and controllability of power system. In steady state conditions, this study specifically considers the impact of equivalent resistance of DC line on the control of DC voltage station, and designs targeted control strategy to effectively avoid long-term overvoltage operation problem of DC side voltage in multi-terminal DC transmission system. In transient conditions, this paper proposes a DC voltage coordination control scheme dominated by master-slave control, supplemented by voltage deviation control as a backup strategy. At the same time, the study also designs a smooth switching control strategy for slave station to take over DC voltage control authority and a logic simplified voltage deviation control strategy to further enhance the robustness and stability of the system. Finally, an electromagnetic transient model of hybrid multi-terminal DC transmission is built through computer software Matlab/Simulink environment, and the proposed control strategy is verified by simulation. The results prove its correctness and effectiveness. This study provides strong theoretical support and practical guidance for intelligent control and optimization of high-voltage DC transmission system.

Accurate traffic forecasting is crucial for effective urban planning and transportation management, enabling efficient resource allocation and enhanced travel experiences. However, existing models often face limitations in generalization, struggling with zero-shot prediction on unseen regions and cities, as well as diminished long-term accuracy. This is primarily due to the inherent challenges in handling the spatial and temporal heterogeneity of traffic data, coupled with the significant distribution shift across time and space. In this work, we aim to unlock new possibilities for building versatile, resilient and adaptive spatio-temporal foundation models for traffic prediction. To achieve this goal, we introduce a novel foundation model, named OpenCity, that can effectively capture and normalize the underlying spatio-temporal patterns from diverse data characteristics, facilitating zero-shot generalization across diverse urban environments. OpenCity integrates the Transformer architecture with graph neural networks to model the complex spatio-temporal dependencies in traffic data. By pre-training OpenCity on large-scale, heterogeneous traffic datasets, we enable the model to learn rich, generalizable representations that can be seamlessly applied to a wide range of traffic forecasting scenarios. Experimental results demonstrate that OpenCity exhibits exceptional zero-shot predictive performance. Moreover, OpenCity showcases promising scaling laws, suggesting the potential for developing a truly one-for-all traffic prediction solution that can adapt to new urban contexts with minimal overhead. We made our proposed OpenCity model open-source and it is available at the following link: https://github.com/HKUDS/OpenCity.

Graph Neural Networks (GNNs) have evolved to understand graph structures through recursive exchanges and aggregations among nodes. To enhance robustness, self-supervised learning (SSL) has become a vital tool for data augmentation. Traditional methods often depend on fine-tuning with task-specific labels, limiting their effectiveness when labeled data is scarce. Our research tackles this by advancing graph model generalization in zero-shot learning environments. Inspired by the success of large language models (LLMs), we aim to create a graph-oriented LLM capable of exceptional generalization across various datasets and tasks without relying on downstream graph data. We introduce the GraphGPT framework, which integrates LLMs with graph structural knowledge through graph instruction tuning. This framework includes a text-graph grounding component to link textual and graph structures and a dual-stage instruction tuning approach with a lightweight graph-text alignment projector. These innovations allow LLMs to comprehend complex graph structures and enhance adaptability across diverse datasets and tasks. Our framework demonstrates superior generalization in both supervised and zero-shot graph learning tasks, surpassing existing benchmarks. The open-sourced model implementation of our GraphGPT is available at https://github.com/HKUDS/GraphGPT.

Metal-Organic Frameworks (MOFs), renowned for their structural diversity and design flexibility, have demonstrated considerable potential in catalysis. However, the traditional trial-and-error approach is inefficient for optimizing catalytic performance in synthetic space. In this work, we propose MOFsyn Agent, a framework leveraging Large Language Models (LLMs) to guide the efficient synthesis and performance optimization of MOFs. MOFsyn Agent is an integrated system comprising two core components: data automatic analyzer and material mechanism analyzer. The system can be driven to perform data analysis operations such as task disassembly, code writing and execution by simple natural language commands only. Furthermore, the incorporation of the Retrieval-Augmented Generation (RAG) technology enables MOFsyn Agent to access external knowledge bases in real time, significantly enhancing its domain applicability in MOFs synthesis. Taking Ni@UiO-66(Ce) as an illustrative example, MOFsyn Agent conducted a comprehensive analysis of the relationships between synthetic conditions, structural characteristics, and catalytic performance. Through machine learning (ML), key features such as nickel content and metallic nickel were identified. Notably, MOFsyn Agent not only realizes the numerical prediction based on ML, but also optimizes the synthesis scheme from the textual level, and creatively proposes a step-by-step reduction strategy, which replaces the traditional one-pot direct reduction method. Experimental results demonstrated that the optimized Ni@UiO-66(Ce) can achieve saturated hydrogenation of dicyclopentadiene under more mild conditions (80 °C, 2 MPa) in only 1 h, validating the accuracy and reliability of the proposed method. This research provides an efficient tool for material intelligent synthesis to researchers without a programming background, accelerating the development process of new materials.

Self-Supervised Learning (SSL) methods harness the concept of semantic invariance by utilizing data augmentation strategies to produce similar representations for different deformations of the same input. Essentially, the model captures the shared information among multiple augmented views of samples, while disregarding the non-shared information that may be beneficial for downstream tasks. To address this issue, we introduce a module called CompMod with Meta Comprehensive Regularization (MCR), embedded into existing self-supervised frameworks, to make the learned representations more comprehensive. Specifically, we update our proposed model through a bi-level optimization mechanism, enabling it to capture comprehensive features. Additionally, guided by the constrained extraction of features using maximum entropy coding, the self-supervised learning model learns more comprehensive features on top of learning consistent features. In addition, we provide theoretical support for our proposed method from information theory and causal counterfactual perspective. Experimental results show that our method achieves significant improvement in classification, object detection and instance segmentation tasks on multiple benchmark datasets.

Understanding the evolution of scholarly impact is essential for many real-life decision-making processes in academia, such as research planning, frontier exploration, and award selection. Popular platforms like Google Scholar and Web of Science rely on numerical indicators that are too abstract to convey the context and content of scientific impact, while most existing visualization approaches on mapping science do not consider the presentation of individual scholars' impact evolution using curated self-citation data. This paper builds on our previous work and proposes an integrated pipeline to visualize a scholar's impact evolution from multiple topic facets. A novel 3D prism-shaped visual metaphor is introduced as the overview of a scholar's impact, whilst their scientific evolution on each topic is displayed in a more structured manner. Additional designs by topic chord diagram, streamgraph visualization, and inter-topic flow map, optimized by an elaborate layout algorithm, assist in perceiving the scholar's scientific evolution across topics. A new six-degree-impact glyph metaphor highlights key interdisciplinary works driving the evolution. The proposed visualization methods are evaluated through case studies analyzing the careers of prestigious Turing award laureates and a major visualization venue.

Plutonium-238 has always been considered as the one of the promising radioisotopes for space nuclear power supply, which has long half-life, low radiation protection level, high power density, and stable fuel form at high temperatures. The industrial-scale production of 238Pu mainly depends on irradiating solid 237NpO2 target in high flux reactors, however the production process faces problems such as large fission loss and high requirements for product quality control. In this paper, a conceptual design study of producing 238Pu in a multi-purpose high flux reactor was evaluated and analyzed, which includes a sensitivity analysis on 238Pu production and a further study on the irradiation scheme. It demonstrated that the target structure and its location in the reactor, as well as the operation scheme has an impact on 238Pu amount and product quality. Furthermore, the production efficiency could be improved by optimizing target material concentration, target locations in the core and reflector. This work provides technical support for irradiation production of 238Pu in high flux reactors.