Explore the vast range of titles available.

Water molecules play a significant role in maintaining protein structural stability and facilitating molecular interactions. Accurate prediction of water molecule positions around protein structures is essential for understanding their biological roles and has significant implications for protein engineering and drug discovery. Here, we introduce SuperWater, a novel generative AI framework that integrates a score-based diffusion model with equivariant graph neural networks to predict water molecule placements around proteins with high accuracy. SuperWater surpasses existing methods, delivering state-of-the-art performance in both crystal water coverage and prediction precision, achieving water localization within 0.3  ± 0.06 Å of experimentally validated positions. We demonstrate the capabilities of SuperWater through case studies involving protein hydration, protein-ligand binding, and protein-protein binding sites. This framework can be adapted for various applications, including structural biology, binding site prediction, multi-body docking, and water-mediated drug design. Water molecules are crucial for protein stability and interactions, yet predicting their precise positions around protein structures remain challenging. Here, the authors present SuperWater, a generative AI framework combining a score-based diffusion model and equivariant graph neural networks, to achieve unprecedented accuracy in water placement, demonstrating the capabilities involving protein hydration, protein-ligand binding and protein-protein interactions.

In today’s world, users typically take to online social media to express their opinions, which can inherently be both positive and negative. In fact, online social networks can be best modeled as signed networks, where opinions in the form of positive and negative links can exist between users, such as our friends and foes (e.g., “unfriended” users), respectively. Furthermore, users can also express their opinions to content generated by others through online social interactions, such as commenting or rating. Intuitively, these two types of opinions in the form of links and interactions should be related. For example, users’ interactions are likely to be positive (or negative) to those they have positively (or negatively) established links with. Similarly, we tend to establish positive (or negative) links with those whose generated content we frequently positively (or negatively) interact with online. Hence, in this paper, we first verify these assumptions by understanding the correlation between these two types of opinions from both a local and global perspective. Then, we propose a framework that jointly solves the link and interaction polarity prediction problem based on our newly found understanding of how these two problems are correlated. We ultimately perform experiments on a real-world signed network to demonstrate the effectiveness of our proposed approach to help mitigate both the data sparsity and cold-start problems found in the two tasks of link and interaction polarity prediction.

As we rapidly continue into the information age, the rate at which data is produced has created an unprecedented demand for novel methods to effectively extract insightful patterns. We can then seek to understand the past, make predictions about the future, and ultimately take actionable steps towards improving our society. Thus, due to the fact that much of today’s big data can be represented as graphs, emphasis is being taken to harness the natural structure of data through network analysis. Traditionally, network analysis has focused on networks having only positive links, or unsigned networks. However, in many real-world systems, relations between nodes in a graph can be both positive and negative, or signed networks. For example, in online social media, users not only have positive links such as friends, followers, and those they trust, but also can establish negative links to those they distrust, towards their foes, or block and unfriend users.Thus, although signed networks are ubiquitous due to their ability to represent negative links in addition to positive links, they have been significantly under explored. In addition, due to the rise in popularity of today's social media and increased polarization online, this has led to both an increased attention and demand for advanced methods to perform the typical network analysis tasks when also taking into consideration negative links. More specifically, there is a need for methods that can measure, model, mine, and apply signed networks that harness both these positive and negative relations. However, this raises novel challenges, as the properties and principles of negative links are not necessarily the same as positive links, and furthermore the social theories that have been used in unsigned networks might not apply with the inclusion of negative links.The chief objective of this dissertation is to first analyze the distinct properties negative links have as compared to positive links and towards improving network analysis with negative links by researching the utility and how to harness social theories that have been established in a holistic view of networks containing both positive and negative links. We discover that simply extending unsigned network analysis is typically not sufficient and that although the existence of negative links introduces numerous challenges, they also provide unprecedented opportunities for advancing the frontier of the network analysis domain. In particular, we develop advanced methods in signed networks for measuring node relevance and centrality (i.e., signed network measuring), present the first generative signed network model and extend/analyze balance theory to signed bipartite networks (i.e., signed network modeling), construct the first signed graph convolutional network which learns node representations that can achieve state-of-the-art prediction performance and then furthermore introduce the novel idea of transformation-based network embedding (i.e., signed network mining), and apply signed networks by creating a framework that can infer both link and interaction polarity levels in online social media and constructing an advanced comprehensive congressional vote prediction framework built around harnessing signed networks.

Traditional multimodal learning approaches require expensive alignment pre-training to bridge vision and language modalities, typically projecting visual features into discrete text token spaces. We challenge both fundamental assumptions underlying this paradigm by proposing Inverse-LLaVA, a novel approach that eliminates alignment pre-training entirely while inverting the conventional mapping direction. Rather than projecting visual features to text space, our method maps text embeddings into continuous visual representation space and performs fusion within transformer intermediate layers. Through selective additive components in attention mechanisms, we enable dynamic integration of visual and textual representations without requiring massive image-text alignment datasets. Comprehensive experiments across nine multimodal benchmarks demonstrate nuanced performance trade-offs: Inverse-LLaVA achieves notable improvements on reasoning-intensive and cognitive tasks (MM-VET: +0.2%, VizWiz: +1.8%, ScienceQA: +0.2%, cognitive reasoning: +27.2%), while showing expected decreases in perception tasks requiring memorized visual-text associations (celebrity recognition: -49.5%, OCR: -21.3%). These results provide the first empirical evidence that alignment pre-training is not necessary for effective multimodal learning, particularly for complex reasoning tasks. Our work establishes the feasibility of a new paradigm that reduces computational requirements by 45%, challenges conventional wisdom about modality fusion, and opens new research directions for efficient multimodal architectures that preserve modality-specific characteristics. Our project website with code and additional resources is available at https://inverse-llava.github.io.

Machine learning provides a valuable tool for analyzing high-dimensional functional neuroimaging data, and is proving effective in predicting various neurological conditions, psychiatric disorders, and cognitive patterns. In functional magnetic resonance imaging (MRI) research, interactions between brain regions are commonly modeled using graph-based representations. The potency of graph machine learning methods has been established across myriad domains, marking a transformative step in data interpretation and predictive modeling. Yet, despite their promise, the transposition of these techniques to the neuroimaging domain has been challenging due to the expansive number of potential preprocessing pipelines and the large parameter search space for graph-based dataset construction. In this paper, we introduce NeuroGraph, a collection of graph-based neuroimaging datasets, and demonstrated its utility for predicting multiple categories of behavioral and cognitive traits. We delve deeply into the dataset generation search space by crafting 35 datasets that encompass static and dynamic brain connectivity, running in excess of 15 baseline methods for benchmarking. Additionally, we provide generic frameworks for learning on both static and dynamic graphs. Our extensive experiments lead to several key observations. Notably, using correlation vectors as node features, incorporating larger number of regions of interest, and employing sparser graphs lead to improved performance. To foster further advancements in graph-based data driven neuroimaging analysis, we offer a comprehensive open-source Python package that includes the benchmark datasets, baseline implementations, model training, and standard evaluation.

Graph Neural Networks (GNNs) have been successfully adopted in recommender systems by virtue of the message-passing that implicitly captures collaborative effect. Nevertheless, most of the existing message-passing mechanisms for recommendation are directly inherited from GNNs without scrutinizing whether the captured collaborative effect would benefit the prediction of user preferences. In this paper, we first analyze how message-passing captures the collaborative effect and propose a recommendation-oriented topological metric, Common Interacted Ratio (CIR), which measures the level of interaction between a specific neighbor of a node with the rest of its neighbors. After demonstrating the benefits of leveraging collaborations from neighbors with higher CIR, we propose a recommendation-tailored GNN, Collaboration-Aware Graph Convolutional Network (CAGCN), that goes beyond 1-Weisfeiler-Lehman(1-WL) test in distinguishing non-bipartite-subgraph-isomorphic graphs. Experiments on six benchmark datasets show that the best CAGCN variant outperforms the most representative GNN-based recommendation model, LightGCN, by nearly 10% in Recall@20 and also achieves around 80% speedup. Our code is publicly available at https://github.com/YuWVandy/CAGCN.

Graph Neural Networks (GNNs) have achieved significant success in learning better representations by performing feature propagation and transformation iteratively to leverage neighborhood information. Nevertheless, iterative propagation restricts the information of higher-layer neighborhoods to be transported through and fused with the lower-layer neighborhoods', which unavoidably results in feature smoothing between neighborhoods in different layers and can thus compromise the performance, especially on heterophily networks. Furthermore, most deep GNNs only recognize the importance of higher-layer neighborhoods while yet to fully explore the importance of multi-hop dependency within the context of different layer neighborhoods in learning better representations. In this work, we first theoretically analyze the feature smoothing between neighborhoods in different layers and empirically demonstrate the variance of the homophily level across neighborhoods at different layers. Motivated by these analyses, we further propose a tree decomposition method to disentangle neighborhoods in different layers to alleviate feature smoothing among these layers. Moreover, we characterize the multi-hop dependency via graph diffusion within our tree decomposition formulation to construct Tree Decomposed Graph Neural Network (TDGNN), which can flexibly incorporate information from large receptive fields and aggregate this information utilizing the multi-hop dependency. Comprehensive experiments demonstrate the superior performance of TDGNN on both homophily and heterophily networks under a variety of node classification settings. Extensive parameter analysis highlights the ability of TDGNN to prevent over-smoothing and incorporate features from shallow layers with deeper multi-hop dependencies, which provides new insights towards deeper graph neural networks. Code of TDGNN: http://github.com/YuWVandy/TDGNN

Capability distillation applies knowledge distillation to selected model capabilities, aiming to compress a large language model (LLM) into a smaller one while preserving the abilities needed for a downstream task. However, most existing methods treat capabilities as independent training targets and overlook how improving one capability can reshape the student's broader capability profile, especially when multiple abilities jointly determine task success. We study capability distillation under a fixed token budget and identify two consistent patterns: distillation induces systematic, budget-dependent cross-capability transfer, and additional budget often brings limited task-relevant gains while sometimes degrading other useful abilities. Building on these insights, we propose ReAD, a Reinforcement-guided cApability Distillation framework that explicitly accounts for capability interdependence. ReAD first infers task-essential capabilities, then generates capability-targeted supervision on the fly, and finally uses an uncertainty-aware contextual bandit to adaptively allocate the distillation budget based on expected utility gains. Extensive experiments show that ReAD improves downstream utility under the same token budget while reducing harmful spillover and wasted distillation effort compared to strong baselines. Our code is publicly available at https://github.com/LabRAI/ReAD.

In this study, we present the first comprehensive evaluation of modern LLMs - including GPT-4, GPT-4o, GPT-3.5-Turbo, Gemini 1.5 Pro, DeepSeek-V3, Llama 3.2, and BERT - across three core social media analytics tasks on a Twitter (X) dataset: (I) Social Media Authorship Verification, (II) Social Media Post Generation, and (III) User Attribute Inference. For the authorship verification, we introduce a systematic sampling framework over diverse user and post selection strategies and evaluate generalization on newly collected tweets from January 2024 onward to mitigate \"seen-data\" bias. For post generation, we assess the ability of LLMs to produce authentic, user-like content using comprehensive evaluation metrics. Bridging Tasks I and II, we conduct a user study to measure real users' perceptions of LLM-generated posts conditioned on their own writing. For attribute inference, we annotate occupations and interests using two standardized taxonomies (IAB Tech Lab 2023 and 2018 U.S. SOC) and benchmark LLMs against existing baselines. Overall, our unified evaluation provides new insights and establishes reproducible benchmarks for LLM-driven social media analytics. The code and data are provided in the supplementary material and will also be made publicly available upon publication.

Ligand-based virtual screening (VS) is an essential step in drug discovery that evaluates large chemical libraries to identify compounds that potentially bind to a therapeutic target. However, VS faces three major challenges: class imbalance due to the low active rate, structural imbalance among active molecules where certain scaffolds dominate, and the need to identify structurally diverse active compounds for novel drug development. We introduce ScaffAug, a scaffold-aware VS framework that addresses these challenges through three modules. The augmentation module first generates synthetic data conditioned on scaffolds of actual hits using generative models, specifically a graph diffusion model. This helps mitigate the class imbalance and furthermore the structural imbalance, due to our proposed scaffold-aware sampling algorithm, designed to produce more samples for active molecules with underrepresented scaffolds. A model-agnostic self-training module is then used to safely integrate the generated synthetic data from our augmentation module with the original labeled data. Lastly, we introduce a reranking module that improves VS by enhancing scaffold diversity in the top recommended set of molecules, while still maintaining and even enhancing the overall general performance of identifying novel, active compounds. We conduct comprehensive computational experiments across five target classes, comparing ScaffAug against existing baseline methods by reporting the performance of multiple evaluation metrics and performing ablation studies on ScaffAug. Overall, this work introduces novel perspectives on effectively enhancing VS by leveraging generative augmentations, reranking, and general scaffold-awareness.