Yifu Wu

He Cheng, Yifu Wu, Saksham Khatwani et al.

Knowledge graphs (KGs) are increasingly integrated with large language models (LLMs) to provide structured, verifiable reasoning. A core operation in this integration is multi-hop retrieval, yet existing systems struggle to balance efficiency, scalability, and interpretability. We introduce LogosKG, a novel, hardware-aligned framework that enables scalable and interpretable k-hop retrieval on large KGs by building on symbolic KG formulations and executing traversal as hardware-efficient operations over decomposed subject, object, and relation representations. To scale to billion-edge graphs, LogosKG integrates degree-aware partitioning, cross-graph routing, and on-demand caching. Experiments show substantial efficiency gains over CPU and GPU baselines without loss of retrieval fidelity. With proven performance in KG retrieval, a downstream two-round KG-LLM interaction demonstrates how LogosKG enables large-scale, evidence-grounded analysis of how KG topology, such as hop distribution and connectivity, shapes the alignment between structured biomedical knowledge and LLM diagnostic reasoning, thereby opening the door for next-generation KG-LLM integration. The source code is publicly available at https://github.com/LARK-NLP-Lab/LogosKG, and an online demo is available at https://lark-nlp-lab-logoskg.hf.space/.

Maya Kruse, Shiyue Hu, Nicholas Derby et al.

Recent advances in large language models (LLMs) have shown potential in clinical text summarization, but their ability to handle long patient trajectories with multi-modal data spread across time remains underexplored. This study systematically evaluates several state-of-the-art open-source LLMs, their Retrieval Augmented Generation (RAG) variants and chain-of-thought (CoT) prompting on long-context clinical summarization and prediction. We examine their ability to synthesize structured and unstructured Electronic Health Records (EHR) data while reasoning over temporal coherence, by re-engineering existing tasks, including discharge summarization and diagnosis prediction from two publicly available EHR datasets. Our results indicate that long context windows improve input integration but do not consistently enhance clinical reasoning, and LLMs are still struggling with temporal progression and rare disease prediction. While RAG shows improvements in hallucination in some cases, it does not fully address these limitations. Our work fills the gap in long clinical text summarization, establishing a foundation for evaluating LLMs with multi-modal data and temporal reasoning.

Yiwen Zhang, Yifu Wu, Wenyue Hua et al.

Large language models have been used to simulate human society using multi-agent systems. Most current social simulation research emphasizes interactive behaviors in fixed environments, ignoring information opacity, relationship variability, and diffusion diversity. In this paper, we first propose a general framework for exploring multi-agent information diffusion. We identified LLMs' deficiency in the perception and utilization of social relationships, as well as diverse actions. Then, we designed a dynamic attention mechanism to help agents allocate attention to different information, addressing the limitations of the LLM attention mechanism. Agents start by responding to external information stimuli within a five-agent group, increasing group size and forming information circles while developing relationships and sharing information. Additionally, we explore the information diffusion features in the asymmetric open environment by observing the evolution of information gaps, diffusion patterns, and the accumulation of social capital, which are closely linked to psychological, sociological, and communication theories.

Kai Zhou, Youbiao He, Chong Zhong et al.

Despite high reliability, modern power systems with growing renewable penetration face an increasing risk of cascading outages. Real-time cascade mitigation requires fast, complex operational decisions under uncertainty. In this work, we extend the influence graph into a Markov decision process model (MDP) for real-time mitigation of cascading outages in power transmission systems, accounting for uncertainties in generation, load, and initial contingencies. The MDP includes a do-nothing action to allow for conservative decision-making and is solved using reinforcement learning. We present a policy gradient learning algorithm initialized with a policy corresponding to the unmitigated case and designed to handle invalid actions. The proposed learning method converges faster than the conventional algorithm. Through careful reward design, we learn a policy that takes conservative actions without deteriorating system conditions. The model is validated on the IEEE 14-bus and IEEE 118-bus systems. The results show that proactive line disconnections can effectively reduce cascading risk, and certain lines consistently emerge as critical in mitigating cascade propagation.

Zhanhong Tan, Jiebo Song, Xiaolong Ma et al.

Weight pruning is a powerful technique to realize model compression. We propose PCNN, a fine-grained regular 1D pruning method. A novel index format called Sparsity Pattern Mask (SPM) is presented to encode the sparsity in PCNN. Leveraging SPM with limited pruning patterns and non-zero sequences with equal length, PCNN can be efficiently employed in hardware. Evaluated on VGG-16 and ResNet-18, our PCNN achieves the compression rate up to 8.4X with only 0.2% accuracy loss. We also implement a pattern-aware architecture in 55nm process, achieving up to 9.0X speedup and 28.39 TOPS/W efficiency with only 3.1% on-chip memory overhead of indices.

Yifu Wu, Jin Wei, Rigoberto Roche

Deep learning for supervised learning has achieved astonishing performance in various machine learning applications. However, annotated data is expensive and rare. In practice, only a small portion of data samples are annotated. Pseudo-ensembling-based approaches have achieved state-of-the-art results in computer vision related tasks. However, it still relies on the quality of an initial model built by labeled data. Less labeled data may degrade model performance a lot. Domain constraint is another way regularize the posterior but has some limitation. In this paper, we proposed a fuzzy domain-constraint-based framework which loses the requirement of traditional constraint learning and enhances the model quality for semi supervision. Simulations results show the effectiveness of our design.

Gihan J. Mendis, Yifu Wu, Jin Wei et al.

Thanks to the advances in machine learning, data-driven analysis tools have become valuable solutions for various applications. However, there still remain essential challenges to develop effective data-driven methods because of the need to acquire a large amount of data and to have sufficient computing power to handle the data. In many instances these challenges are addressed by relying on a dominant cloud computing vendor, but, although commercial cloud vendors provide valuable platforms for data analytics, they can suffer from a lack of transparency, security, and privacy-perservation. Furthermore, reliance on cloud servers prevents applying big data analytics in environments where the computing power is scattered. To address these challenges, a decentralize, secure, and privacy-preserving computing paradigm is proposed to enable an asynchronized cooperative computing process amongst scattered and untrustworthy computing nodes that may have limited computing power and computing intelligence. This paradigm is designed by exploring blockchain, decentralized learning, homomorphic encryption, and software defined networking(SDN) techniques. The performance of the proposed paradigm is evaluated via different scenarios in the simulation section.