Xiangyun Ding

Xiangyun Ding, Rohin Garg, Yan Gu et al.

The metric skip-list is a data structure designed for efficient nearest and $k$-nearest neighbor search in metric spaces. For many real-world datasets with reasonable distributions - specifically, those with a constant expansion rate - it supports $\tilde{O}(n)$ construction time and $O(k\log n)$ query time, where $n$ is the input size and $k$ is the number of nearest neighbors in queries. Notably, unlike alternative approaches, it does not require a bounded aspect ratio, making it more flexible for input data distributions. However, the inherently sequential nature of its original construction has, to our knowledge, precluded any existing parallel algorithm. In this paper, we present highly parallel and work-efficient algorithms for constructing metric skip lists. Under the assumption of a constant expansion rate, our approach achieves an expected work of $O(n \log n)$ and a polylogarithmic span with high probability. Our design is based on novel algorithmic insights that improves the sequential procedure, enabling a divide-and-conquer strategy that facilitates parallelism while maintaining efficiency. With our algorithms, we can also support improved bounds for relevant applications using nearest neighbor as building blocks, including bichromatic closest pair (BCP), density-based clustering, and $k$-NN graph construction, among others. To our knowledge, many of these results represent the first solutions to achieve both work efficiency and polylogarithmic span, relying solely on the assumption of a constant expansion rate.

Haohui Mai, Xiaoyan Guo, Xiangyun Ding et al.

LLM-based coding agents can generate functionally correct GPU kernels, yet their performance remains far below hand-optimized libraries on critical computations such as matrix multiplication, attention, and Mixture-of-Experts (MoE). Peak GPU performance requires coordinated reasoning over tightly coupled optimizations, including tiling, shared-memory staging, software pipelining, and instruction scheduling, while existing agents rely on sparse pass/fail feedback, leaving them unable to diagnose global constraint violations. We present Argus, an agentic framework that addresses this through data-flow invariants: compile-time specifications encoding how data must be choreographed throughout kernel execution. Argus introduces a tile-based, Pythonic DSL exposing hardware instructions and compiler policies while hiding low-level representations. The DSL provides tag functions to propagate symbolic annotations through data and control flow, and tag assertions to enforce relational constraints at use sites. When violations occur, the compiler returns concrete counterexamples identifying the thread, data element, and program point, enabling dense, structured feedback for targeted fixes. Invariants are verified at compile time via abstract interpretation over a layout algebra and SMT solving, with zero runtime overhead. An in-context reinforcement learning planner learns to select optimizations and synthesize effective invariants, supported by a curated knowledge base of GPU optimization techniques. We evaluate Argus on the AMD MI300X GPU across GEMM, flash attention, and MoE kernels accounting for over 90% of GPU time in LLM inference. Generated kernels achieve 99-104% of state-of-the-art hand-optimized assembly throughput and are 2-1543x faster than existing agentic systems. Argus further generalizes to 200 KernelBench tasks, solving 100% of Level 1 and 90% of Level 2 problems.

Xiangyun Ding, Wenjian Yu, Yuyang Xie et al.

A model-based collaborative filtering (CF) approach utilizing fast adaptive randomized singular value decomposition (SVD) is proposed for the matrix completion problem in recommender system. Firstly, a fast adaptive PCA frameworkis presented which combines the fixed-precision randomized matrix factorization algorithm [1] and accelerating skills for handling large sparse data. Then, a novel termination mechanism for the adaptive PCA is proposed to automatically determine a number of latent factors for achieving the near optimal prediction accuracy during the subsequent model-based CF. The resulted CF approach has good accuracy while inheriting high runtime efficiency. Experiments on real data show that, the proposed adaptive PCA is up to 2.7X and 6.7X faster than the original fixed-precision SVD approach [1] and svds in Matlab repsectively, while preserving accuracy. The proposed model-based CF approach is able to efficiently process the MovieLens data with 20M ratings and exhibits more than 10X speedup over the regularized matrix factorization based approach [2] and the fast singular value thresholding approach [3] with comparable or better accuracy. It also owns the advantage of parameter free. Compared with the deep-learning-based CF approach, the proposed approach is much more computationally efficient, with just marginal performance loss.

Yanyuet Man, Xiangyun Ding, Xingcheng Yao et al.

Throughout the world, breast cancer is one of the leading causes of female death. Recently, deep learning methods are developed to automatically grade breast cancer of histological slides. However, the performance of existing deep learning models is limited due to the lack of large annotated biomedical datasets. One promising way to relieve the annotating burden is to leverage the unannotated datasets to enhance the trained model. In this paper, we first apply active learning method in breast cancer grading, and propose a semi-supervised framework based on expectation maximization (EM) model. The proposed EM approach is based on the collaborative filtering among the annotated and unannotated datasets. The collaborative filtering method effectively extracts useful and credible datasets from the unannotated images. Results of pixel-wise prediction of whole-slide images (WSI) demonstrate that the proposed method not only outperforms state-of-art methods, but also significantly reduces the annotation cost by over 70%.