arXiv subjects

Henry Hu

Publications and source records attributed to Henry Hu.

4 recordsLinked to original sources

arXiv2026

Optimal Spectral Algorithms for Correlated Two-view Models in High Dimensions

We study high-dimensional inference in correlated two-view models, focusing on spectral methods for strong detection and weak recovery. We introduce a general framework, motivated by a TAP type heuristic from statistical physics, that provides a unified treatment of three canonical models: high-dimensional canonical correlation analysis, and the correlated spiked Wigner and Wishart models. Our main contribution is to construct explicit spectral algorithms in all three settings, that achieve strong detection and weak recovery down to the corresponding thresholds, where we prove matching information-theoretic lower bounds. Furthermore, our spectral procedures operate without knowledge of the model parameters, relying solely on the observed data. This demonstrates the optimality of spectral methods in these models and the broad statistical applicability of the framework.

math.ST↗

arXiv2026

CCCL: Node-Spanning GPU Collectives with CXL Memory Pooling

Large language models (LLMs) training or inference across multiple nodes introduces significant pressure on GPU memory and interconnect bandwidth. The Compute Express Link (CXL) shared memory pool offers a scalable solution by enabling memory sharing across nodes, reducing over-provisioning and improving resource utilization. We propose \name, a collective communication library, leveraging the CXL shared memory pool to support cross-node GPU operations without relying on traditional RDMA-based networking. Our design addresses the challenges on synchronization, data interleaving, and communication parallelization faced by using the CXL shared memory pool for collective communications. Evaluating on multiple nodes with a TITAN-II CXL switch and six Micron CZ120 memory cards, we show that \name achieves highly efficient collective operations across hosts, demonstrating CXL's potential for scalable, memory-centric GPU communication. Our evaluation demonstrates that \name achieves average performance improvements of 1.34$\times$ for AllGather, 1.84$\times$ for Broadcast, 1.94$\times$ for Gather, and 1.04$\times$ for Scatter, compared to the original RDMA-based implementation over 200 Gbps InfiniBand. \textcolor{dong}{In addition, the evaluation with a case of LLM training shows 1.11$\times$ speedup compared with the InfiniBand while saving production cost by $2.75\times$ in hardware.}

cs.DC↗

arXiv2025

Efficient Vision-Language Reasoning via Adaptive Token Pruning

Real-world deployment of Vision-Language Models (VLMs) is hindered by high computational demands, as existing architectures inefficiently process all tokens uniformly. We introduce Adaptive Token Pruning (ATP), a dynamic inference mechanism that retains only the most informative tokens based on contextual relevance. ATP operates at the vision-language interface, assigning a hybrid importance score combining ViT CLS attention (intra-modal saliency) and CLIP text-image similarity (inter-modal relevance) to keep top-K tokens for the LLM. Unlike static compression, ATP adapts to each input without modifying the backbone. Proposed as a lightweight gating module, ATP is compatible with popular backbones like BLIP-2, LLaVA, and Flamingo. Preliminary evaluations across VQAv2, GQA, and COCO indicate that ATP reduces inference FLOPs by around 40% and achieves roughly 1.5x speedups in end-to-end latency with negligible accuracy loss (less than 1%). Qualitative analyses suggest ATP preserves visual grounding and enhances interpretability. Beyond efficiency, we investigate robustness under corruptions; observations suggest adaptive pruning suppresses spurious correlations, improving stability. These findings imply that resource-constrained inference and model reliability are not competing objectives. Finally, we discuss ATP's role in efficient multimodal edge computing pipelines.

cs.CV↗

arXiv2023

On the regularity conditions in the CLT for the LUE

We consider the Laguerre Unitary Ensemble (LUE), the set of $n\times n$ sample covariance matrices $M = \frac{1}{n}X^*X$ where the $m\times n$ ($n \le m$) matrix $X$ has i.i.d. standard complex Gaussian entries. In particular we are concerned with the case where $\alpha := m - n$ is fixed, in which case the limiting eigenvalue density has a hard edge at $0$. We study the minimal regularity conditions required for a central limit theorem (CLT) type result to hold for the linear spectral statistics of the LUE. As long as the expression for the limiting variance is finite (and a slightly stonger condition holds near the soft edge) we show that the variance of the linear spectral statistic converges and consequently the CLT holds. Our methods are based on analyzing the explicit kernel for the LUE using asymptotics of the Laguerre polynomials. The CLT follows from approximating the test function of the statistic by Chebyshev polynomials.

math.PR↗