arXiv subjects

Yicheng Lu

Publications and source records attributed to Yicheng Lu.

3 recordsLinked to original sources

arXiv2026

CM-GAI: Continuum Mechanistic Generative Artificial Intelligence Theory for Data Dynamics

Generative artificial intelligence (GAI) plays a fundamental role in high-impact AI-based systems such as SORA and AlphaFold. Currently, GAI shows limited capability in the specialized domains due to data scarcity. In this paper, we develop a continuum mechanics-based theoretical framework to generalize the optimal transport theory from pure mathematics, which can be used to describe the dynamics of data, realizing the generative tasks with a small amount of data. The developed theory is used to solve three typical problem involved in many mechanical designs and engineering applications: at material level, how to generate the stress-strain response outside the range of experimental conditions based on experimentally measured stress-strain data; at structure level, how to generate the temperature-dependent stress fields under the thermal loading; at system level, how to generate the plastic strain fields under transient dynamic loading. Our results show the proposed theory can complete the generation successfully, showing its potential to solve many difficult problems involved in engineering applications, not limited to mechanics problems, such as image generation. The present work shows that mechanics can provide new tools for computer science. The limitation of the proposed theory is also discussed.

cs.CE↗

arXiv2025

AIBrix: Towards Scalable, Cost-Effective Large Language Model Inference Infrastructure

We introduce AIBrix, a cloud-native, open-source framework designed to optimize and simplify large-scale LLM deployment in cloud environments. Unlike traditional cloud-native stacks, AIBrix follows a co-design philosophy, ensuring every layer of the infrastructure is purpose-built for seamless integration with inference engines like vLLM. AIBrix introduces several key innovations to reduce inference costs and enhance performance including high-density LoRA management for dynamic adapter scheduling, LLM-specific autoscalers, and prefix-aware, load-aware routing. To further improve efficiency, AIBrix incorporates a distributed KV cache, boosting token reuse across nodes, leading to a 50% increase in throughput and a 70% reduction in inference latency. AIBrix also supports unified AI runtime which streamlines model management while maintaining vendor-agnostic engine compatibility. For large-scale multi-node inference, AIBrix employs hybrid orchestration -- leveraging Kubernetes for coarse-grained scheduling and Ray for fine-grained execution -- to balance efficiency and flexibility. Additionally, an SLO-driven GPU optimizer dynamically adjusts resource allocations, optimizing heterogeneous serving to maximize cost efficiency while maintaining service guarantees. Finally, AIBrix enhances system reliability with AI accelerator diagnostic tools, enabling automated failure detection and mock-up testing to improve fault resilience. AIBrix is available at https://github.com/vllm-project/aibrix.

cs.DC↗

arXiv2023

SplitGNN: Splitting GNN for Node Classification with Heterogeneous Attention

With the frequent happening of privacy leakage and the enactment of privacy laws across different countries, data owners are reluctant to directly share their raw data and labels with any other party. In reality, a lot of these raw data are stored in the graph database, especially for finance. For collaboratively building graph neural networks(GNNs), federated learning(FL) may not be an ideal choice for the vertically partitioned setting where privacy and efficiency are the main concerns. Moreover, almost all the existing federated GNNs are mainly designed for homogeneous graphs, which simplify various types of relations as the same type, thus largely limits their performance. We bridge this gap by proposing a split learning-based GNN(SplitGNN), where this model is divided into two sub-models: the local GNN model includes all the private data related computation to generate local node embeddings, whereas the global model calculates global embeddings by aggregating all the participants' local embeddings. Our SplitGNN allows the isolated heterogeneous neighborhood to be collaboratively utilized. To better capture representations, we propose a novel Heterogeneous Attention(HAT) algorithm and use both node-based and path-based attention mechanisms to learn various types of nodes and edges with multi-hop relation features. We demonstrate the effectiveness of our SplitGNN on node classification tasks for two standard public datasets and the real-world dataset. Extensive experimental results validate that our proposed SplitGNN significantly outperforms the state-of-the-art(SOTA) methods.

cs.LG↗